chris/Cog
2
1
Fork 0
Code Issues 65 Pull requests Projects Releases Packages Wiki Activity Actions

Updated libopus to version 1.1.1 and opusfile to 0.6

Browse source
This commit is contained in:
Chris Moeller 2015-12-22 00:38:41 -08:00
parent d700174fa6
commit 0f13825a04
375 changed files with 39185 additions and 15635 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

130
Frameworks/Opus/Opus.xcodeproj/project.pbxproj
View file

@ -204,6 +204,33 @@
8375B36817FFEFFF0092A79F /* config.h in Headers */ = {isa = PBXBuildFile; fileRef = 8375B36617FFEFFF0092A79F /* config.h */; };
8375B36917FFEFFF0092A79F /* version.h in Headers */ = {isa = PBXBuildFile; fileRef = 8375B36717FFEFFF0092A79F /* version.h */; };
8375B37617FFF8FE0092A79F /* opus_encoder.c in Sources */ = {isa = PBXBuildFile; fileRef = 8375B37517FFF8FE0092A79F /* opus_encoder.c */; };
837928B31C293C6F005DBCEA /* cpu_support.h in Headers */ = {isa = PBXBuildFile; fileRef = 837928AF1C293C6F005DBCEA /* cpu_support.h */; };
837928B41C293C6F005DBCEA /* static_modes_fixed_arm_ne10.h in Headers */ = {isa = PBXBuildFile; fileRef = 837928B01C293C6F005DBCEA /* static_modes_fixed_arm_ne10.h */; };
837928B51C293C6F005DBCEA /* static_modes_float_arm_ne10.h in Headers */ = {isa = PBXBuildFile; fileRef = 837928B11C293C6F005DBCEA /* static_modes_float_arm_ne10.h */; };
837928B61C293C6F005DBCEA /* celt_decoder.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928B21C293C6F005DBCEA /* celt_decoder.c */; };
837928C21C293C78005DBCEA /* celt_lpc_sse.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928B91C293C78005DBCEA /* celt_lpc_sse.c */; settings = {COMPILER_FLAGS = "-msse4.1"; }; };
837928C31C293C78005DBCEA /* celt_lpc_sse.h in Headers */ = {isa = PBXBuildFile; fileRef = 837928BA1C293C78005DBCEA /* celt_lpc_sse.h */; };
837928C41C293C78005DBCEA /* pitch_sse.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928BB1C293C78005DBCEA /* pitch_sse.c */; };
837928C51C293C78005DBCEA /* pitch_sse.h in Headers */ = {isa = PBXBuildFile; fileRef = 837928BC1C293C78005DBCEA /* pitch_sse.h */; };
837928C61C293C78005DBCEA /* pitch_sse2.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928BD1C293C78005DBCEA /* pitch_sse2.c */; };
837928C71C293C78005DBCEA /* pitch_sse4_1.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928BE1C293C78005DBCEA /* pitch_sse4_1.c */; settings = {COMPILER_FLAGS = "-msse4.1"; }; };
837928C81C293C78005DBCEA /* x86_celt_map.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928BF1C293C78005DBCEA /* x86_celt_map.c */; };
837928C91C293C78005DBCEA /* x86cpu.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928C01C293C78005DBCEA /* x86cpu.c */; };
837928CA1C293C78005DBCEA /* x86cpu.h in Headers */ = {isa = PBXBuildFile; fileRef = 837928C11C293C78005DBCEA /* x86cpu.h */; };
837928D31C293CA1005DBCEA /* main_sse.h in Headers */ = {isa = PBXBuildFile; fileRef = 837928CC1C293CA1005DBCEA /* main_sse.h */; };
837928D41C293CA1005DBCEA /* NSQ_del_dec_sse.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928CD1C293CA1005DBCEA /* NSQ_del_dec_sse.c */; settings = {COMPILER_FLAGS = "-msse4.1"; }; };
837928D51C293CA1005DBCEA /* NSQ_sse.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928CE1C293CA1005DBCEA /* NSQ_sse.c */; settings = {COMPILER_FLAGS = "-msse4.1"; }; };
837928D61C293CA1005DBCEA /* SigProc_FIX_sse.h in Headers */ = {isa = PBXBuildFile; fileRef = 837928CF1C293CA1005DBCEA /* SigProc_FIX_sse.h */; };
837928D71C293CA1005DBCEA /* VAD_sse.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928D01C293CA1005DBCEA /* VAD_sse.c */; settings = {COMPILER_FLAGS = "-msse4.1"; }; };
837928D81C293CA1005DBCEA /* VQ_WMat_EC_sse.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928D11C293CA1005DBCEA /* VQ_WMat_EC_sse.c */; settings = {COMPILER_FLAGS = "-msse4.1"; }; };
837928D91C293CA1005DBCEA /* x86_silk_map.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928D21C293CA1005DBCEA /* x86_silk_map.c */; };
837928DB1C29406E005DBCEA /* celt_encoder.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928DA1C29406E005DBCEA /* celt_encoder.c */; };
837928DF1C2940A9005DBCEA /* analysis.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928DC1C2940A9005DBCEA /* analysis.c */; };
837928E01C2940A9005DBCEA /* analysis.h in Headers */ = {isa = PBXBuildFile; fileRef = 837928DD1C2940A9005DBCEA /* analysis.h */; };
837928E11C2940A9005DBCEA /* opus_multistream_decoder.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928DE1C2940A9005DBCEA /* opus_multistream_decoder.c */; };
837928E51C2940C0005DBCEA /* mlp.h in Headers */ = {isa = PBXBuildFile; fileRef = 837928E21C2940C0005DBCEA /* mlp.h */; };
837928E61C2940C0005DBCEA /* mlp_data.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928E31C2940C0005DBCEA /* mlp_data.c */; };
837928E71C2940C0005DBCEA /* mlp.c in Sources */ = {isa = PBXBuildFile; fileRef = 837928E41C2940C0005DBCEA /* mlp.c */; };
/* End PBXBuildFile section */
/* Begin PBXContainerItemProxy section */
@ -442,6 +469,33 @@
8375B36617FFEFFF0092A79F /* config.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = config.h; sourceTree = "<group>"; };
8375B36717FFEFFF0092A79F /* version.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = version.h; sourceTree = "<group>"; };
8375B37517FFF8FE0092A79F /* opus_encoder.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = opus_encoder.c; sourceTree = "<group>"; };
837928AF1C293C6F005DBCEA /* cpu_support.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = cpu_support.h; sourceTree = "<group>"; };
837928B01C293C6F005DBCEA /* static_modes_fixed_arm_ne10.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = static_modes_fixed_arm_ne10.h; sourceTree = "<group>"; };
837928B11C293C6F005DBCEA /* static_modes_float_arm_ne10.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = static_modes_float_arm_ne10.h; sourceTree = "<group>"; };
837928B21C293C6F005DBCEA /* celt_decoder.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = celt_decoder.c; sourceTree = "<group>"; };
837928B91C293C78005DBCEA /* celt_lpc_sse.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = celt_lpc_sse.c; sourceTree = "<group>"; };
837928BA1C293C78005DBCEA /* celt_lpc_sse.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = celt_lpc_sse.h; sourceTree = "<group>"; };
837928BB1C293C78005DBCEA /* pitch_sse.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = pitch_sse.c; sourceTree = "<group>"; };
837928BC1C293C78005DBCEA /* pitch_sse.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = pitch_sse.h; sourceTree = "<group>"; };
837928BD1C293C78005DBCEA /* pitch_sse2.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = pitch_sse2.c; sourceTree = "<group>"; };
837928BE1C293C78005DBCEA /* pitch_sse4_1.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = pitch_sse4_1.c; sourceTree = "<group>"; };
837928BF1C293C78005DBCEA /* x86_celt_map.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = x86_celt_map.c; sourceTree = "<group>"; };
837928C01C293C78005DBCEA /* x86cpu.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = x86cpu.c; sourceTree = "<group>"; };
837928C11C293C78005DBCEA /* x86cpu.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = x86cpu.h; sourceTree = "<group>"; };
837928CC1C293CA1005DBCEA /* main_sse.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = main_sse.h; sourceTree = "<group>"; };
837928CD1C293CA1005DBCEA /* NSQ_del_dec_sse.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = NSQ_del_dec_sse.c; sourceTree = "<group>"; };
837928CE1C293CA1005DBCEA /* NSQ_sse.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = NSQ_sse.c; sourceTree = "<group>"; };
837928CF1C293CA1005DBCEA /* SigProc_FIX_sse.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = SigProc_FIX_sse.h; sourceTree = "<group>"; };
837928D01C293CA1005DBCEA /* VAD_sse.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = VAD_sse.c; sourceTree = "<group>"; };
837928D11C293CA1005DBCEA /* VQ_WMat_EC_sse.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = VQ_WMat_EC_sse.c; sourceTree = "<group>"; };
837928D21C293CA1005DBCEA /* x86_silk_map.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = x86_silk_map.c; sourceTree = "<group>"; };
837928DA1C29406E005DBCEA /* celt_encoder.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = celt_encoder.c; sourceTree = "<group>"; };
837928DC1C2940A9005DBCEA /* analysis.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = analysis.c; sourceTree = "<group>"; };
837928DD1C2940A9005DBCEA /* analysis.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = analysis.h; sourceTree = "<group>"; };
837928DE1C2940A9005DBCEA /* opus_multistream_decoder.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = opus_multistream_decoder.c; sourceTree = "<group>"; };
837928E21C2940C0005DBCEA /* mlp.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = mlp.h; sourceTree = "<group>"; };
837928E31C2940C0005DBCEA /* mlp_data.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = mlp_data.c; sourceTree = "<group>"; };
837928E41C2940C0005DBCEA /* mlp.c */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.c; path = mlp.c; sourceTree = "<group>"; };
/* End PBXFileReference section */
/* Begin PBXFrameworksBuildPhase section */
@ -533,6 +587,12 @@
8375B0A217FFEB2C0092A79F /* celt */ = {
isa = PBXGroup;
children = (
837928B81C293C78005DBCEA /* x86 */,
837928AF1C293C6F005DBCEA /* cpu_support.h */,
837928DA1C29406E005DBCEA /* celt_encoder.c */,
837928B01C293C6F005DBCEA /* static_modes_fixed_arm_ne10.h */,
837928B11C293C6F005DBCEA /* static_modes_float_arm_ne10.h */,
837928B21C293C6F005DBCEA /* celt_decoder.c */,
8375B0A317FFEB2C0092A79F /* _kiss_fft_guts.h */,
8375B0A417FFEB2C0092A79F /* arch.h */,
8375B0A517FFEB2C0092A79F /* bands.c */,
@ -595,6 +655,7 @@
8375B10917FFEB2C0092A79F /* silk */ = {
isa = PBXGroup;
children = (
837928CB1C293CA1005DBCEA /* x86 */,
8375B10A17FFEB2C0092A79F /* A2NLSF.c */,
8375B10B17FFEB2C0092A79F /* ana_filt_bank_1.c */,
8375B10C17FFEB2C0092A79F /* API.h */,
@ -740,6 +801,12 @@
8375B1B117FFEB2C0092A79F /* src */ = {
isa = PBXGroup;
children = (
837928E21C2940C0005DBCEA /* mlp.h */,
837928E31C2940C0005DBCEA /* mlp_data.c */,
837928E41C2940C0005DBCEA /* mlp.c */,
837928DC1C2940A9005DBCEA /* analysis.c */,
837928DD1C2940A9005DBCEA /* analysis.h */,
837928DE1C2940A9005DBCEA /* opus_multistream_decoder.c */,
8375B37517FFF8FE0092A79F /* opus_encoder.c */,
8375B1B217FFEB2C0092A79F /* opus.c */,
8375B1B617FFEB2C0092A79F /* opus_decoder.c */,
@ -781,6 +848,36 @@
path = src;
sourceTree = "<group>";
};
837928B81C293C78005DBCEA /* x86 */ = {
isa = PBXGroup;
children = (
837928B91C293C78005DBCEA /* celt_lpc_sse.c */,
837928BA1C293C78005DBCEA /* celt_lpc_sse.h */,
837928BB1C293C78005DBCEA /* pitch_sse.c */,
837928BC1C293C78005DBCEA /* pitch_sse.h */,
837928BD1C293C78005DBCEA /* pitch_sse2.c */,
837928BE1C293C78005DBCEA /* pitch_sse4_1.c */,
837928BF1C293C78005DBCEA /* x86_celt_map.c */,
837928C01C293C78005DBCEA /* x86cpu.c */,
837928C11C293C78005DBCEA /* x86cpu.h */,
);
path = x86;
sourceTree = "<group>";
};
837928CB1C293CA1005DBCEA /* x86 */ = {
isa = PBXGroup;
children = (
837928CC1C293CA1005DBCEA /* main_sse.h */,
837928CD1C293CA1005DBCEA /* NSQ_del_dec_sse.c */,
837928CE1C293CA1005DBCEA /* NSQ_sse.c */,
837928CF1C293CA1005DBCEA /* SigProc_FIX_sse.h */,
837928D01C293CA1005DBCEA /* VAD_sse.c */,
837928D11C293CA1005DBCEA /* VQ_WMat_EC_sse.c */,
837928D21C293CA1005DBCEA /* x86_silk_map.c */,
);
path = x86;
sourceTree = "<group>";
};
/* End PBXGroup section */
/* Begin PBXHeadersBuildPhase section */
@ -798,9 +895,12 @@
8375B24017FFEB2C0092A79F /* vq.h in Headers */,
8375B21E17FFEB2C0092A79F /* fixed_debug.h in Headers */,
8375B21017FFEB2C0092A79F /* celt.h in Headers */,
837928CA1C293C78005DBCEA /* x86cpu.h in Headers */,
8375B30417FFEB2C0092A79F /* tables.h in Headers */,
8375B2B817FFEB2C0092A79F /* main_FLP.h in Headers */,
8375B2D817FFEB2C0092A79F /* macros.h in Headers */,
837928D61C293CA1005DBCEA /* SigProc_FIX_sse.h in Headers */,
837928C51C293C78005DBCEA /* pitch_sse.h in Headers */,
8375B27617FFEB2C0092A79F /* control.h in Headers */,
8375B2E417FFEB2C0092A79F /* pitch_est_defines.h in Headers */,
8375B2D917FFEB2C0092A79F /* main.h in Headers */,
@ -808,6 +908,7 @@
8375B23117FFEB2C0092A79F /* quant_bands.h in Headers */,
8375B22F17FFEB2C0092A79F /* pitch.h in Headers */,
8375B28817FFEB2C0092A79F /* errors.h in Headers */,
837928C31C293C78005DBCEA /* celt_lpc_sse.h in Headers */,
8375B2C217FFEB2C0092A79F /* SigProc_FLP.h in Headers */,
8375B21917FFEB2C0092A79F /* entcode.h in Headers */,
8375B22B17FFEB2C0092A79F /* modes.h in Headers */,
@ -828,23 +929,29 @@
8375B23417FFEB2C0092A79F /* stack_alloc.h in Headers */,
8375B2F417FFEB2C0092A79F /* resampler_structs.h in Headers */,
8375B2CE17FFEB2C0092A79F /* Inlines.h in Headers */,
837928B31C293C6F005DBCEA /* cpu_support.h in Headers */,
8375B23317FFEB2C0092A79F /* rate.h in Headers */,
8375B22017FFEB2C0092A79F /* float_cast.h in Headers */,
8375B20B17FFEB2C0092A79F /* _kiss_fft_guts.h in Headers */,
8375B21717FFEB2C0092A79F /* ecintrin.h in Headers */,
837928B41C293C6F005DBCEA /* static_modes_fixed_arm_ne10.h in Headers */,
8375B30D17FFEB2C0092A79F /* typedef.h in Headers */,
8375B22D17FFEB2C0092A79F /* os_support.h in Headers */,
8375B22217FFEB2C0092A79F /* kiss_fft.h in Headers */,
8375B23517FFEB2C0092A79F /* static_modes_fixed.h in Headers */,
8375B22917FFEB2C0092A79F /* mfrngcod.h in Headers */,
8375B22617FFEB2C0092A79F /* mathops.h in Headers */,
837928B51C293C6F005DBCEA /* static_modes_float_arm_ne10.h in Headers */,
8375B21D17FFEB2C0092A79F /* entenc.h in Headers */,
8375B36817FFEFFF0092A79F /* config.h in Headers */,
8375B21F17FFEB2C0092A79F /* fixed_generic.h in Headers */,
8375B2D617FFEB2C0092A79F /* MacroCount.h in Headers */,
8375B21617FFEB2C0092A79F /* cwrs.h in Headers */,
837928D31C293CA1005DBCEA /* main_sse.h in Headers */,
8375B2D717FFEB2C0092A79F /* MacroDebug.h in Headers */,
837928E01C2940A9005DBCEA /* analysis.h in Headers */,
8375B35217FFEB2C0092A79F /* internal.h in Headers */,
837928E51C2940C0005DBCEA /* mlp.h in Headers */,
8375B2C717FFEB2C0092A79F /* structs_FLP.h in Headers */,
8375B28417FFEB2C0092A79F /* define.h in Headers */,
8375B21B17FFEB2C0092A79F /* entdec.h in Headers */,
@ -948,12 +1055,15 @@
8375B30B17FFEB2C0092A79F /* tables_pulses_per_block.c in Sources */,
8375B22A17FFEB2C0092A79F /* modes.c in Sources */,
8375B2A617FFEB2C0092A79F /* apply_sine_window_FLP.c in Sources */,
837928C81C293C78005DBCEA /* x86_celt_map.c in Sources */,
8375B2FB17FFEB2C0092A79F /* stereo_decode_pred.c in Sources */,
8375B27917FFEB2C0092A79F /* control_SNR.c in Sources */,
8375B2C917FFEB2C0092A79F /* wrappers_FLP.c in Sources */,
8375B2DC17FFEB2C0092A79F /* NLSF_del_dec_quant.c in Sources */,
837928D91C293CA1005DBCEA /* x86_silk_map.c in Sources */,
8375B2AD17FFEB2C0092A79F /* find_LPC_FLP.c in Sources */,
8375B2FF17FFEB2C0092A79F /* stereo_MS_to_LR.c in Sources */,
837928C21C293C78005DBCEA /* celt_lpc_sse.c in Sources */,
8375B2CD17FFEB2C0092A79F /* init_encoder.c in Sources */,
8375B2FD17FFEB2C0092A79F /* stereo_find_predictor.c in Sources */,
8375B30317FFEB2C0092A79F /* table_LSF_cos.c in Sources */,
@ -961,8 +1071,10 @@
8375B2E317FFEB2C0092A79F /* NSQ_del_dec.c in Sources */,
8375B2B317FFEB2C0092A79F /* levinsondurbin_FLP.c in Sources */,
8375B2CB17FFEB2C0092A79F /* HP_variable_cutoff.c in Sources */,
837928D51C293CA1005DBCEA /* NSQ_sse.c in Sources */,
8375B27D17FFEB2C0092A79F /* decode_core.c in Sources */,
8375B27117FFEB2C0092A79F /* bwexpander.c in Sources */,
837928C61C293C78005DBCEA /* pitch_sse2.c in Sources */,
8375B2C117FFEB2C0092A79F /* schur_FLP.c in Sources */,
8375B21817FFEB2C0092A79F /* entcode.c in Sources */,
8375B2BB17FFEB2C0092A79F /* prefilter_FLP.c in Sources */,
@ -974,6 +1086,8 @@
8375B2BA17FFEB2C0092A79F /* pitch_analysis_core_FLP.c in Sources */,
8375B27C17FFEB2C0092A79F /* dec_API.c in Sources */,
8375B2EA17FFEB2C0092A79F /* resampler.c in Sources */,
837928D81C293CA1005DBCEA /* VQ_WMat_EC_sse.c in Sources */,
837928C41C293C78005DBCEA /* pitch_sse.c in Sources */,
8375B2F517FFEB2C0092A79F /* shell_coder.c in Sources */,
8375B2CC17FFEB2C0092A79F /* init_decoder.c in Sources */,
8375B28717FFEB2C0092A79F /* encode_pulses.c in Sources */,
@ -989,21 +1103,27 @@
8375B2B917FFEB2C0092A79F /* noise_shape_analysis_FLP.c in Sources */,
8375B23217FFEB2C0092A79F /* rate.c in Sources */,
8375B2B217FFEB2C0092A79F /* k2a_FLP.c in Sources */,
837928E11C2940A9005DBCEA /* opus_multistream_decoder.c in Sources */,
8375B28617FFEB2C0092A79F /* encode_indices.c in Sources */,
8375B31617FFEB2C0092A79F /* opus_decoder.c in Sources */,
8375B27A17FFEB2C0092A79F /* debug.c in Sources */,
837928E61C2940C0005DBCEA /* mlp_data.c in Sources */,
8375B28517FFEB2C0092A79F /* enc_API.c in Sources */,
8375B2C617FFEB2C0092A79F /* sort_FLP.c in Sources */,
8375B27017FFEB2C0092A79F /* biquad_alt.c in Sources */,
8375B30217FFEB2C0092A79F /* sum_sqr_shift.c in Sources */,
8375B37617FFF8FE0092A79F /* opus_encoder.c in Sources */,
837928D71C293CA1005DBCEA /* VAD_sse.c in Sources */,
8375B2C817FFEB2C0092A79F /* warped_autocorrelation_FLP.c in Sources */,
8375B2E517FFEB2C0092A79F /* pitch_est_tables.c in Sources */,
8375B22717FFEB2C0092A79F /* mdct.c in Sources */,
8375B30517FFEB2C0092A79F /* tables_gain.c in Sources */,
837928C91C293C78005DBCEA /* x86cpu.c in Sources */,
8375B2BE17FFEB2C0092A79F /* residual_energy_FLP.c in Sources */,
8375B2D417FFEB2C0092A79F /* LPC_analysis_filter.c in Sources */,
8375B20D17FFEB2C0092A79F /* bands.c in Sources */,
837928C71C293C78005DBCEA /* pitch_sse4_1.c in Sources */,
837928DF1C2940A9005DBCEA /* analysis.c in Sources */,
8375B23F17FFEB2C0092A79F /* vq.c in Sources */,
8375B2CF17FFEB2C0092A79F /* inner_prod_aligned.c in Sources */,
8375B2C517FFEB2C0092A79F /* solve_LS_FLP.c in Sources */,
@ -1019,6 +1139,7 @@
8375B27817FFEB2C0092A79F /* control_codec.c in Sources */,
8375B27417FFEB2C0092A79F /* CNG.c in Sources */,
8375B2A717FFEB2C0092A79F /* autocorrelation_FLP.c in Sources */,
837928B61C293C6F005DBCEA /* celt_decoder.c in Sources */,
8375B2B117FFEB2C0092A79F /* inner_product_FLP.c in Sources */,
8375B2D117FFEB2C0092A79F /* lin2log.c in Sources */,
8375B2D317FFEB2C0092A79F /* LP_variable_cutoff.c in Sources */,
@ -1037,6 +1158,7 @@
8375B30A17FFEB2C0092A79F /* tables_pitch_lag.c in Sources */,
8375B35117FFEB2C0092A79F /* internal.c in Sources */,
8375B2F117FFEB2C0092A79F /* resampler_private_up2_HQ.c in Sources */,
837928D41C293CA1005DBCEA /* NSQ_del_dec_sse.c in Sources */,
8375B31B17FFEB2C0092A79F /* opus_multistream.c in Sources */,
8375B2F017FFEB2C0092A79F /* resampler_private_IIR_FIR.c in Sources */,
8375B2FC17FFEB2C0092A79F /* stereo_encode_pred.c in Sources */,
@ -1059,6 +1181,7 @@
8375B30617FFEB2C0092A79F /* tables_LTP.c in Sources */,
8375B2B717FFEB2C0092A79F /* LTP_scale_ctrl_FLP.c in Sources */,
8375B2AC17FFEB2C0092A79F /* energy_FLP.c in Sources */,
837928E71C2940C0005DBCEA /* mlp.c in Sources */,
8375B2AF17FFEB2C0092A79F /* find_pitch_lags_FLP.c in Sources */,
8375B2B617FFEB2C0092A79F /* LTP_analysis_filter_FLP.c in Sources */,
8375B27317FFEB2C0092A79F /* check_control_input.c in Sources */,
@ -1067,6 +1190,7 @@
8375B2B517FFEB2C0092A79F /* LPC_inv_pred_gain_FLP.c in Sources */,
8375B30917FFEB2C0092A79F /* tables_other.c in Sources */,
8375B2AB17FFEB2C0092A79F /* encode_frame_FLP.c in Sources */,
837928DB1C29406E005DBCEA /* celt_encoder.c in Sources */,
8375B2DD17FFEB2C0092A79F /* NLSF_encode.c in Sources */,
8375B2FA17FFEB2C0092A79F /* sort.c in Sources */,
8375B2BD17FFEB2C0092A79F /* regularize_correlations_FLP.c in Sources */,
@ -1188,6 +1312,9 @@
"$(inherited)",
/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/include,
../../Frameworks/Ogg/include,
Opus/opus/silk,
Opus/opus/celt,
Opus/opus,
);
INFOPLIST_FILE = "Opus/Opus-Info.plist";
INSTALL_PATH = "@loader_path/../Frameworks";
@ -1222,6 +1349,9 @@
"$(inherited)",
/Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/include,
../../Frameworks/Ogg/include,
Opus/opus/silk,
Opus/opus/celt,
Opus/opus,
);
INFOPLIST_FILE = "Opus/Opus-Info.plist";
INSTALL_PATH = "@loader_path/../Frameworks";

371
Frameworks/Opus/Opus/opus/INSTALL
View file

@ -1 +1,370 @@
See README file for details
Installation Instructions
*************************
Copyright (C) 1994-1996, 1999-2002, 2004-2013 Free Software Foundation,
Inc.
Copying and distribution of this file, with or without modification,
are permitted in any medium without royalty provided the copyright
notice and this notice are preserved. This file is offered as-is,
without warranty of any kind.
Basic Installation
==================
Briefly, the shell command `./configure && make && make install'
should configure, build, and install this package. The following
more-detailed instructions are generic; see the `README' file for
instructions specific to this package. Some packages provide this
`INSTALL' file but do not implement all of the features documented
below. The lack of an optional feature in a given package is not
necessarily a bug. More recommendations for GNU packages can be found
in *note Makefile Conventions: (standards)Makefile Conventions.
The `configure' shell script attempts to guess correct values for
various system-dependent variables used during compilation. It uses
those values to create a `Makefile' in each directory of the package.
It may also create one or more `.h' files containing system-dependent
definitions. Finally, it creates a shell script `config.status' that
you can run in the future to recreate the current configuration, and a
file `config.log' containing compiler output (useful mainly for
debugging `configure').
It can also use an optional file (typically called `config.cache'
and enabled with `--cache-file=config.cache' or simply `-C') that saves
the results of its tests to speed up reconfiguring. Caching is
disabled by default to prevent problems with accidental use of stale
cache files.
If you need to do unusual things to compile the package, please try
to figure out how `configure' could check whether to do them, and mail
diffs or instructions to the address given in the `README' so they can
be considered for the next release. If you are using the cache, and at
some point `config.cache' contains results you don't want to keep, you
may remove or edit it.
The file `configure.ac' (or `configure.in') is used to create
`configure' by a program called `autoconf'. You need `configure.ac' if
you want to change it or regenerate `configure' using a newer version
of `autoconf'.
The simplest way to compile this package is:
1. `cd' to the directory containing the package's source code and type
`./configure' to configure the package for your system.
Running `configure' might take a while. While running, it prints
some messages telling which features it is checking for.
2. Type `make' to compile the package.
3. Optionally, type `make check' to run any self-tests that come with
the package, generally using the just-built uninstalled binaries.
4. Type `make install' to install the programs and any data files and
documentation. When installing into a prefix owned by root, it is
recommended that the package be configured and built as a regular
user, and only the `make install' phase executed with root
privileges.
5. Optionally, type `make installcheck' to repeat any self-tests, but
this time using the binaries in their final installed location.
This target does not install anything. Running this target as a
regular user, particularly if the prior `make install' required
root privileges, verifies that the installation completed
correctly.
6. You can remove the program binaries and object files from the
source code directory by typing `make clean'. To also remove the
files that `configure' created (so you can compile the package for
a different kind of computer), type `make distclean'. There is
also a `make maintainer-clean' target, but that is intended mainly
for the package's developers. If you use it, you may have to get
all sorts of other programs in order to regenerate files that came
with the distribution.
7. Often, you can also type `make uninstall' to remove the installed
files again. In practice, not all packages have tested that
uninstallation works correctly, even though it is required by the
GNU Coding Standards.
8. Some packages, particularly those that use Automake, provide `make
distcheck', which can by used by developers to test that all other
targets like `make install' and `make uninstall' work correctly.
This target is generally not run by end users.
Compilers and Options
=====================
Some systems require unusual options for compilation or linking that
the `configure' script does not know about. Run `./configure --help'
for details on some of the pertinent environment variables.
You can give `configure' initial values for configuration parameters
by setting variables in the command line or in the environment. Here
is an example:
./configure CC=c99 CFLAGS=-g LIBS=-lposix
*Note Defining Variables::, for more details.
Compiling For Multiple Architectures
====================================
You can compile the package for more than one kind of computer at the
same time, by placing the object files for each architecture in their
own directory. To do this, you can use GNU `make'. `cd' to the
directory where you want the object files and executables to go and run
the `configure' script. `configure' automatically checks for the
source code in the directory that `configure' is in and in `..'. This
is known as a "VPATH" build.
With a non-GNU `make', it is safer to compile the package for one
architecture at a time in the source code directory. After you have
installed the package for one architecture, use `make distclean' before
reconfiguring for another architecture.
On MacOS X 10.5 and later systems, you can create libraries and
executables that work on multiple system types--known as "fat" or
"universal" binaries--by specifying multiple `-arch' options to the
compiler but only a single `-arch' option to the preprocessor. Like
this:
./configure CC="gcc -arch i386 -arch x86_64 -arch ppc -arch ppc64" \
CXX="g++ -arch i386 -arch x86_64 -arch ppc -arch ppc64" \
CPP="gcc -E" CXXCPP="g++ -E"
This is not guaranteed to produce working output in all cases, you
may have to build one architecture at a time and combine the results
using the `lipo' tool if you have problems.
Installation Names
==================
By default, `make install' installs the package's commands under
`/usr/local/bin', include files under `/usr/local/include', etc. You
can specify an installation prefix other than `/usr/local' by giving
`configure' the option `--prefix=PREFIX', where PREFIX must be an
absolute file name.
You can specify separate installation prefixes for
architecture-specific files and architecture-independent files. If you
pass the option `--exec-prefix=PREFIX' to `configure', the package uses
PREFIX as the prefix for installing programs and libraries.
Documentation and other data files still use the regular prefix.
In addition, if you use an unusual directory layout you can give
options like `--bindir=DIR' to specify different values for particular
kinds of files. Run `configure --help' for a list of the directories
you can set and what kinds of files go in them. In general, the
default for these options is expressed in terms of `${prefix}', so that
specifying just `--prefix' will affect all of the other directory
specifications that were not explicitly provided.
The most portable way to affect installation locations is to pass the
correct locations to `configure'; however, many packages provide one or
both of the following shortcuts of passing variable assignments to the
`make install' command line to change installation locations without
having to reconfigure or recompile.
The first method involves providing an override variable for each
affected directory. For example, `make install
prefix=/alternate/directory' will choose an alternate location for all
directory configuration variables that were expressed in terms of
`${prefix}'. Any directories that were specified during `configure',
but not in terms of `${prefix}', must each be overridden at install
time for the entire installation to be relocated. The approach of
makefile variable overrides for each directory variable is required by
the GNU Coding Standards, and ideally causes no recompilation.
However, some platforms have known limitations with the semantics of
shared libraries that end up requiring recompilation when using this
method, particularly noticeable in packages that use GNU Libtool.
The second method involves providing the `DESTDIR' variable. For
example, `make install DESTDIR=/alternate/directory' will prepend
`/alternate/directory' before all installation names. The approach of
`DESTDIR' overrides is not required by the GNU Coding Standards, and
does not work on platforms that have drive letters. On the other hand,
it does better at avoiding recompilation issues, and works well even
when some directory options were not specified in terms of `${prefix}'
at `configure' time.
Optional Features
=================
If the package supports it, you can cause programs to be installed
with an extra prefix or suffix on their names by giving `configure' the
option `--program-prefix=PREFIX' or `--program-suffix=SUFFIX'.
Some packages pay attention to `--enable-FEATURE' options to
`configure', where FEATURE indicates an optional part of the package.
They may also pay attention to `--with-PACKAGE' options, where PACKAGE
is something like `gnu-as' or `x' (for the X Window System). The
`README' should mention any `--enable-' and `--with-' options that the
package recognizes.
For packages that use the X Window System, `configure' can usually
find the X include and library files automatically, but if it doesn't,
you can use the `configure' options `--x-includes=DIR' and
`--x-libraries=DIR' to specify their locations.
Some packages offer the ability to configure how verbose the
execution of `make' will be. For these packages, running `./configure
--enable-silent-rules' sets the default to minimal output, which can be
overridden with `make V=1'; while running `./configure
--disable-silent-rules' sets the default to verbose, which can be
overridden with `make V=0'.
Particular systems
==================
On HP-UX, the default C compiler is not ANSI C compatible. If GNU
CC is not installed, it is recommended to use the following options in
order to use an ANSI C compiler:
./configure CC="cc -Ae -D_XOPEN_SOURCE=500"
and if that doesn't work, install pre-built binaries of GCC for HP-UX.
HP-UX `make' updates targets which have the same time stamps as
their prerequisites, which makes it generally unusable when shipped
generated files such as `configure' are involved. Use GNU `make'
instead.
On OSF/1 a.k.a. Tru64, some versions of the default C compiler cannot
parse its `<wchar.h>' header file. The option `-nodtk' can be used as
a workaround. If GNU CC is not installed, it is therefore recommended
to try
./configure CC="cc"
and if that doesn't work, try
./configure CC="cc -nodtk"
On Solaris, don't put `/usr/ucb' early in your `PATH'. This
directory contains several dysfunctional programs; working variants of
these programs are available in `/usr/bin'. So, if you need `/usr/ucb'
in your `PATH', put it _after_ `/usr/bin'.
On Haiku, software installed for all users goes in `/boot/common',
not `/usr/local'. It is recommended to use the following options:
./configure --prefix=/boot/common
Specifying the System Type
==========================
There may be some features `configure' cannot figure out
automatically, but needs to determine by the type of machine the package
will run on. Usually, assuming the package is built to be run on the
_same_ architectures, `configure' can figure that out, but if it prints
a message saying it cannot guess the machine type, give it the
`--build=TYPE' option. TYPE can either be a short name for the system
type, such as `sun4', or a canonical name which has the form:
CPU-COMPANY-SYSTEM
where SYSTEM can have one of these forms:
OS
KERNEL-OS
See the file `config.sub' for the possible values of each field. If
`config.sub' isn't included in this package, then this package doesn't
need to know the machine type.
If you are _building_ compiler tools for cross-compiling, you should
use the option `--target=TYPE' to select the type of system they will
produce code for.
If you want to _use_ a cross compiler, that generates code for a
platform different from the build platform, you should specify the
"host" platform (i.e., that on which the generated programs will
eventually be run) with `--host=TYPE'.
Sharing Defaults
================
If you want to set default values for `configure' scripts to share,
you can create a site shell script called `config.site' that gives
default values for variables like `CC', `cache_file', and `prefix'.
`configure' looks for `PREFIX/share/config.site' if it exists, then
`PREFIX/etc/config.site' if it exists. Or, you can set the
`CONFIG_SITE' environment variable to the location of the site script.
A warning: not all `configure' scripts look for a site script.
Defining Variables
==================
Variables not defined in a site shell script can be set in the
environment passed to `configure'. However, some packages may run
configure again during the build, and the customized values of these
variables may be lost. In order to avoid this problem, you should set
them in the `configure' command line, using `VAR=value'. For example:
./configure CC=/usr/local2/bin/gcc
causes the specified `gcc' to be used as the C compiler (unless it is
overridden in the site shell script).
Unfortunately, this technique does not work for `CONFIG_SHELL' due to
an Autoconf limitation. Until the limitation is lifted, you can use
this workaround:
CONFIG_SHELL=/bin/bash ./configure CONFIG_SHELL=/bin/bash
`configure' Invocation
======================
`configure' recognizes the following options to control how it
operates.
`--help'
`-h'
Print a summary of all of the options to `configure', and exit.
`--help=short'
`--help=recursive'
Print a summary of the options unique to this package's
`configure', and exit. The `short' variant lists options used
only in the top level, while the `recursive' variant lists options
also present in any nested packages.
`--version'
`-V'
Print the version of Autoconf used to generate the `configure'
script, and exit.
`--cache-file=FILE'
Enable the cache: use and save the results of the tests in FILE,
traditionally `config.cache'. FILE defaults to `/dev/null' to
disable caching.
`--config-cache'
`-C'
Alias for `--cache-file=config.cache'.
`--quiet'
`--silent'
`-q'
Do not print messages saying which checks are being made. To
suppress all normal output, redirect it to `/dev/null' (any error
messages will still be shown).
`--srcdir=DIR'
Look for the package's source code in directory DIR. Usually
`configure' can determine that directory automatically.
`--prefix=DIR'
Use DIR as the installation prefix. *note Installation Names::
for more details, including other options available for fine-tuning
the installation locations.
`--no-create'
`-n'
Run the configure checks, but stop before creating any output
files.
`configure' also accepts some other, not widely useful, options. Run
`configure --help' for more details.

190
Frameworks/Opus/Opus/opus/Makefile.am
View file

@ -1,10 +1,16 @@
# Provide the full test output for failed tests when using the parallel
# test suite (which is enabled by default with automake 1.13+).
export VERBOSE = yes
AUTOMAKE_OPTIONS = subdir-objects
ACLOCAL_AMFLAGS = -I m4
lib_LTLIBRARIES = libopus.la
DIST_SUBDIRS = doc
INCLUDES = -I$(top_srcdir)/include -I$(top_srcdir)/celt -I$(top_srcdir)/silk -I$(top_srcdir)/silk/float -I$(top_srcdir)/silk/fixed
AM_CPPFLAGS = -I$(top_srcdir)/include -I$(top_srcdir)/celt -I$(top_srcdir)/silk \
-I$(top_srcdir)/silk/float -I$(top_srcdir)/silk/fixed $(NE10_CFLAGS)
include celt_sources.mk
include silk_sources.mk
@ -12,9 +18,54 @@ include opus_sources.mk
if FIXED_POINT
SILK_SOURCES += $(SILK_SOURCES_FIXED)
if HAVE_SSE4_1
SILK_SOURCES += $(SILK_SOURCES_SSE4_1) $(SILK_SOURCES_FIXED_SSE4_1)
endif
else
SILK_SOURCES += $(SILK_SOURCES_FLOAT)
if HAVE_SSE4_1
SILK_SOURCES += $(SILK_SOURCES_SSE4_1)
endif
endif
if DISABLE_FLOAT_API
else
OPUS_SOURCES += $(OPUS_SOURCES_FLOAT)
endif
if HAVE_SSE
CELT_SOURCES += $(CELT_SOURCES_SSE)
endif
if HAVE_SSE2
CELT_SOURCES += $(CELT_SOURCES_SSE2)
endif
if HAVE_SSE4_1
CELT_SOURCES += $(CELT_SOURCES_SSE4_1)
endif
if CPU_ARM
CELT_SOURCES += $(CELT_SOURCES_ARM)
SILK_SOURCES += $(SILK_SOURCES_ARM)
if OPUS_ARM_NEON_INTR
CELT_SOURCES += $(CELT_SOURCES_ARM_NEON_INTR)
endif
if HAVE_ARM_NE10
CELT_SOURCES += $(CELT_SOURCES_ARM_NE10)
endif
if OPUS_ARM_EXTERNAL_ASM
noinst_LTLIBRARIES = libarmasm.la
libarmasm_la_SOURCES = $(CELT_SOURCES_ARM_ASM:.s=-gnu.S)
BUILT_SOURCES = $(CELT_SOURCES_ARM_ASM:.s=-gnu.S) \
$(CELT_AM_SOURCES_ARM_ASM:.s.in=.s) \
$(CELT_AM_SOURCES_ARM_ASM:.s.in=-gnu.S)
endif
endif
CLEANFILES = $(CELT_SOURCES_ARM_ASM:.s=-gnu.S) \
$(CELT_AM_SOURCES_ARM_ASM:.s.in=-gnu.S)
include celt_headers.mk
include silk_headers.mk
@ -22,91 +73,117 @@ include opus_headers.mk
libopus_la_SOURCES = $(CELT_SOURCES) $(SILK_SOURCES) $(OPUS_SOURCES)
libopus_la_LDFLAGS = -no-undefined -version-info @OPUS_LT_CURRENT@:@OPUS_LT_REVISION@:@OPUS_LT_AGE@
libopus_la_LIBADD = $(NE10_LIBS) $(LIBM)
if OPUS_ARM_EXTERNAL_ASM
libopus_la_LIBADD += libarmasm.la
endif
pkginclude_HEADERS = include/opus.h include/opus_multistream.h include/opus_types.h include/opus_defines.h
noinst_HEADERS = $(OPUS_HEAD) $(SILK_HEAD) $(CELT_HEAD)
noinst_PROGRAMS = opus_demo repacketizer_demo opus_compare tests/test_opus_api tests/test_opus_encode tests/test_opus_decode celt/tests/test_unit_cwrs32 celt/tests/test_unit_dft celt/tests/test_unit_entropy celt/tests/test_unit_laplace celt/tests/test_unit_mathops celt/tests/test_unit_mdct celt/tests/test_unit_rotation celt/tests/test_unit_types
if EXTRA_PROGRAMS
noinst_PROGRAMS = opus_demo repacketizer_demo opus_compare tests/test_opus_api tests/test_opus_encode tests/test_opus_decode tests/test_opus_padding celt/tests/test_unit_cwrs32 celt/tests/test_unit_dft celt/tests/test_unit_entropy celt/tests/test_unit_laplace celt/tests/test_unit_mathops celt/tests/test_unit_mdct celt/tests/test_unit_rotation celt/tests/test_unit_types
TESTS = celt/tests/test_unit_types celt/tests/test_unit_mathops celt/tests/test_unit_entropy celt/tests/test_unit_laplace celt/tests/test_unit_dft celt/tests/test_unit_mdct celt/tests/test_unit_rotation celt/tests/test_unit_cwrs32 tests/test_opus_api tests/test_opus_decode tests/test_opus_encode
TESTS = celt/tests/test_unit_types celt/tests/test_unit_mathops celt/tests/test_unit_entropy celt/tests/test_unit_laplace celt/tests/test_unit_dft celt/tests/test_unit_mdct celt/tests/test_unit_rotation celt/tests/test_unit_cwrs32 tests/test_opus_api tests/test_opus_decode tests/test_opus_encode tests/test_opus_padding
opus_demo_SOURCES = src/opus_demo.c
opus_demo_LDADD = libopus.la -lm
opus_demo_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
repacketizer_demo_SOURCES = src/repacketizer_demo.c
repacketizer_demo_LDADD = libopus.la -lm
repacketizer_demo_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
opus_compare_SOURCES = src/opus_compare.c
opus_compare_LDADD = -lm
opus_compare_LDADD = $(LIBM)
tests_test_opus_api_SOURCES = tests/test_opus_api.c tests/test_opus_common.h
tests_test_opus_api_LDADD = libopus.la -lm
tests_test_opus_api_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
tests_test_opus_encode_SOURCES = tests/test_opus_encode.c tests/test_opus_common.h
tests_test_opus_encode_LDADD = libopus.la -lm
tests_test_opus_encode_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
tests_test_opus_decode_SOURCES = tests/test_opus_decode.c tests/test_opus_common.h
tests_test_opus_decode_LDADD = libopus.la -lm
tests_test_opus_decode_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
tests_test_opus_padding_SOURCES = tests/test_opus_padding.c tests/test_opus_common.h
tests_test_opus_padding_LDADD = libopus.la $(NE10_LIBS) $(LIBM)
celt_tests_test_unit_cwrs32_SOURCES = celt/tests/test_unit_cwrs32.c
celt_tests_test_unit_cwrs32_LDADD = -lm
celt_tests_test_unit_cwrs32_LDADD = $(LIBM)
celt_tests_test_unit_dft_SOURCES = celt/tests/test_unit_dft.c
celt_tests_test_unit_dft_LDADD = -lm
celt_tests_test_unit_dft_LDADD = $(NE10_LIBS) $(LIBM)
if OPUS_ARM_EXTERNAL_ASM
celt_tests_test_unit_dft_LDADD += libarmasm.la
endif
celt_tests_test_unit_entropy_SOURCES = celt/tests/test_unit_entropy.c
celt_tests_test_unit_entropy_LDADD = -lm
celt_tests_test_unit_entropy_LDADD = $(LIBM)
celt_tests_test_unit_laplace_SOURCES = celt/tests/test_unit_laplace.c
celt_tests_test_unit_laplace_LDADD = -lm
celt_tests_test_unit_laplace_LDADD = $(LIBM)
celt_tests_test_unit_mathops_SOURCES = celt/tests/test_unit_mathops.c
celt_tests_test_unit_mathops_LDADD = -lm
celt_tests_test_unit_mathops_LDADD = $(NE10_LIBS) $(LIBM)
if OPUS_ARM_EXTERNAL_ASM
celt_tests_test_unit_mathops_LDADD += libarmasm.la
endif
celt_tests_test_unit_mdct_SOURCES = celt/tests/test_unit_mdct.c
celt_tests_test_unit_mdct_LDADD = -lm
celt_tests_test_unit_mdct_LDADD = $(NE10_LIBS) $(LIBM)
if OPUS_ARM_EXTERNAL_ASM
celt_tests_test_unit_mdct_LDADD += libarmasm.la
endif
celt_tests_test_unit_rotation_SOURCES = celt/tests/test_unit_rotation.c
celt_tests_test_unit_rotation_LDADD = -lm
celt_tests_test_unit_rotation_LDADD = $(NE10_LIBS) $(LIBM)
if OPUS_ARM_EXTERNAL_ASM
celt_tests_test_unit_rotation_LDADD += libarmasm.la
endif
celt_tests_test_unit_types_SOURCES = celt/tests/test_unit_types.c
celt_tests_test_unit_types_LDADD = -lm
celt_tests_test_unit_types_LDADD = $(LIBM)
endif
if CUSTOM_MODES
pkginclude_HEADERS += include/opus_custom.h
if EXTRA_PROGRAMS
noinst_PROGRAMS += opus_custom_demo
opus_custom_demo_SOURCES = celt/opus_custom_demo.c
opus_custom_demo_LDADD = libopus.la -lm
opus_custom_demo_LDADD = libopus.la $(LIBM)
endif
endif
EXTRA_DIST = version.mk \
opus.pc.in \
opus-uninstalled.pc.in \
opus.m4 \
Makefile.mips \
Makefile.unix \
tests/run_vectors.sh \
opus.sln \
celt/celt.vcxproj \
celt/celt.vcxproj.filters \
src/opus_demo.vcxproj \
src/opus.vcxproj \
src/opus.vcxproj.filters \
src/opus_demo.vcxproj.filters \
tests/test_opus_decode.vcxproj.filters \
tests/test_opus_decode.vcxproj \
tests/test_opus_encode.vcxproj.filters \
tests/test_opus_encode.vcxproj \
tests/test_opus_api.vcxproj.filters \
tests/test_opus_api.vcxproj \
silk/float/silk_float.vcxproj.filters \
silk/float/silk_float.vcxproj \
silk/fixed/silk_fixed.vcxproj.filters \
silk/fixed/silk_fixed.vcxproj \
silk/silk_common.vcxproj \
silk/silk_common.vcxproj.filters \
celt/arm/arm2gnu.pl \
celt/arm/celt_pitch_xcorr_arm.s \
win32/VS2010/silk_float.vcxproj \
win32/VS2010/celt.vcxproj.filters \
win32/VS2010/opus.vcxproj \
win32/VS2010/silk_common.vcxproj.filters \
win32/VS2010/silk_float.vcxproj.filters \
win32/VS2010/test_opus_encode.vcxproj.filters \
win32/VS2010/silk_common.vcxproj \
win32/VS2010/test_opus_encode.vcxproj \
win32/VS2010/opus_demo.vcxproj \
win32/VS2010/test_opus_api.vcxproj.filters \
win32/VS2010/test_opus_api.vcxproj \
win32/VS2010/test_opus_decode.vcxproj.filters \
win32/VS2010/silk_fixed.vcxproj.filters \
win32/VS2010/opus_demo.vcxproj.filters \
win32/VS2010/silk_fixed.vcxproj \
win32/VS2010/opus.vcxproj.filters \
win32/VS2010/test_opus_decode.vcxproj \
win32/VS2010/celt.vcxproj \
win32/VS2010/opus.sln \
win32/genversion.bat \
win32/config.h
@ -183,3 +260,42 @@ dist-hook:
.PHONY: opus check-opus install-opus docs install-docs
# automake doesn't do dependency tracking for asm files, that I can tell
$(CELT_SOURCES_ARM_ASM:%.s=%-gnu.S): celt/arm/armopts-gnu.S
$(CELT_SOURCES_ARM_ASM:%.s=%-gnu.S): $(top_srcdir)/celt/arm/arm2gnu.pl
# convert ARM asm to GNU as format
%-gnu.S: $(top_srcdir)/%.s
$(top_srcdir)/celt/arm/arm2gnu.pl @ARM2GNU_PARAMS@ < $< > $@
# For autoconf-modified sources (e.g., armopts.s)
%-gnu.S: %.s
$(top_srcdir)/celt/arm/arm2gnu.pl @ARM2GNU_PARAMS@ < $< > $@
OPT_UNIT_TEST_OBJ = $(celt_tests_test_unit_mathops_SOURCES:.c=.o) \
$(celt_tests_test_unit_rotation_SOURCES:.c=.o) \
$(celt_tests_test_unit_mdct_SOURCES:.c=.o) \
$(celt_tests_test_unit_dft_SOURCES:.c=.o)
if HAVE_SSE
SSE_OBJ = $(CELT_SOURCES_SSE:.c=.lo)
$(SSE_OBJ) $(OPT_UNIT_TEST_OBJ): CFLAGS += $(OPUS_X86_SSE_CFLAGS)
endif
if HAVE_SSE2
SSE2_OBJ = $(CELT_SOURCES_SSE2:.c=.lo)
$(SSE2_OBJ) $(OPT_UNIT_TEST_OBJ): CFLAGS += $(OPUS_X86_SSE2_CFLAGS)
endif
if HAVE_SSE4_1
SSE4_1_OBJ = $(CELT_SOURCES_SSE4_1:.c=.lo) \
$(SILK_SOURCES_SSE4_1:.c=.lo) \
$(SILK_SOURCES_FIXED_SSE4_1:.c=.lo)
$(SSE4_1_OBJ) $(OPT_UNIT_TEST_OBJ): CFLAGS += $(OPUS_X86_SSE4_1_CFLAGS)
endif
if OPUS_ARM_NEON_INTR
CELT_ARM_NEON_INTR_OBJ = $(CELT_SOURCES_ARM_NEON_INTR:.c=.lo)
$(CELT_ARM_NEON_INTR_OBJ) $(OPT_UNIT_TEST_OBJ): CFLAGS += \
$(OPUS_ARM_NEON_INTR_CFLAGS) $(NE10_CFLAGS)
endif

2095
Frameworks/Opus/Opus/opus/Makefile.in
View file

File diff suppressed because it is too large Load diff

156
Frameworks/Opus/Opus/opus/Makefile.mips Normal file
View file

@ -0,0 +1,156 @@
#################### COMPILE OPTIONS #######################
# Uncomment this for fixed-point build
FIXED_POINT=1
# It is strongly recommended to uncomment one of these
# VAR_ARRAYS: Use C99 variable-length arrays for stack allocation
# USE_ALLOCA: Use alloca() for stack allocation
# If none is defined, then the fallback is a non-threadsafe global array
CFLAGS := -DUSE_ALLOCA $(CFLAGS)
#CFLAGS := -DVAR_ARRAYS $(CFLAGS)
# These options affect performance
# HAVE_LRINTF: Use C99 intrinsics to speed up float-to-int conversion
#CFLAGS := -DHAVE_LRINTF $(CFLAGS)
###################### END OF OPTIONS ######################
-include package_version
include silk_sources.mk
include celt_sources.mk
include opus_sources.mk
ifdef FIXED_POINT
SILK_SOURCES += $(SILK_SOURCES_FIXED)
else
SILK_SOURCES += $(SILK_SOURCES_FLOAT)
OPUS_SOURCES += $(OPUS_SOURCES_FLOAT)
endif
EXESUFFIX =
LIBPREFIX = lib
LIBSUFFIX = .a
OBJSUFFIX = .o
CC = $(TOOLCHAIN_PREFIX)cc$(TOOLCHAIN_SUFFIX)
AR = $(TOOLCHAIN_PREFIX)ar
RANLIB = $(TOOLCHAIN_PREFIX)ranlib
CP = $(TOOLCHAIN_PREFIX)cp
cppflags-from-defines = $(addprefix -D,$(1))
cppflags-from-includes = $(addprefix -I,$(1))
ldflags-from-ldlibdirs = $(addprefix -L,$(1))
ldlibs-from-libs = $(addprefix -l,$(1))
WARNINGS = -Wall -W -Wstrict-prototypes -Wextra -Wcast-align -Wnested-externs -Wshadow
CFLAGS += -mips32r2 -mno-mips16 -std=gnu99 -O2 -g $(WARNINGS) -DENABLE_ASSERTIONS -DMIPSr1_ASM -DOPUS_BUILD -mdspr2 -march=74kc -mtune=74kc -mmt -mgp32
CINCLUDES = include silk celt
ifdef FIXED_POINT
CFLAGS += -DFIXED_POINT=1 -DDISABLE_FLOAT_API
CINCLUDES += silk/fixed
else
CINCLUDES += silk/float
endif
LIBS = m
LDLIBDIRS = ./
CFLAGS += $(call cppflags-from-defines,$(CDEFINES))
CFLAGS += $(call cppflags-from-includes,$(CINCLUDES))
LDFLAGS += $(call ldflags-from-ldlibdirs,$(LDLIBDIRS))
LDLIBS += $(call ldlibs-from-libs,$(LIBS))
COMPILE.c.cmdline = $(CC) -c $(CFLAGS) -o $@ $<
LINK.o = $(CC) $(LDPREFLAGS) $(LDFLAGS)
LINK.o.cmdline = $(LINK.o) $^ $(LDLIBS) -o $@$(EXESUFFIX)
ARCHIVE.cmdline = $(AR) $(ARFLAGS) $@ $^ && $(RANLIB) $@
%$(OBJSUFFIX):%.c
$(COMPILE.c.cmdline)
%$(OBJSUFFIX):%.cpp
$(COMPILE.cpp.cmdline)
# Directives
# Variable definitions
LIB_NAME = opus
TARGET = $(LIBPREFIX)$(LIB_NAME)$(LIBSUFFIX)
SRCS_C = $(SILK_SOURCES) $(CELT_SOURCES) $(OPUS_SOURCES)
OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(SRCS_C))
OPUSDEMO_SRCS_C = src/opus_demo.c
OPUSDEMO_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(OPUSDEMO_SRCS_C))
TESTOPUSAPI_SRCS_C = tests/test_opus_api.c
TESTOPUSAPI_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(TESTOPUSAPI_SRCS_C))
TESTOPUSDECODE_SRCS_C = tests/test_opus_decode.c
TESTOPUSDECODE_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(TESTOPUSDECODE_SRCS_C))
TESTOPUSENCODE_SRCS_C = tests/test_opus_encode.c
TESTOPUSENCODE_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(TESTOPUSENCODE_SRCS_C))
TESTOPUSPADDING_SRCS_C = tests/test_opus_padding.c
TESTOPUSPADDING_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(TESTOPUSPADDING_SRCS_C))
OPUSCOMPARE_SRCS_C = src/opus_compare.c
OPUSCOMPARE_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(OPUSCOMPARE_SRCS_C))
# Rules
all: lib opus_demo opus_compare test_opus_api test_opus_decode test_opus_encode test_opus_padding
lib: $(TARGET)
$(TARGET): $(OBJS)
$(ARCHIVE.cmdline)
opus_demo$(EXESUFFIX): $(OPUSDEMO_OBJS) $(TARGET)
$(LINK.o.cmdline)
test_opus_api$(EXESUFFIX): $(TESTOPUSAPI_OBJS) $(TARGET)
$(LINK.o.cmdline)
test_opus_decode$(EXESUFFIX): $(TESTOPUSDECODE_OBJS) $(TARGET)
$(LINK.o.cmdline)
test_opus_encode$(EXESUFFIX): $(TESTOPUSENCODE_OBJS) $(TARGET)
$(LINK.o.cmdline)
test_opus_padding$(EXESUFFIX): $(TESTOPUSPADDING_OBJS) $(TARGET)
$(LINK.o.cmdline)
opus_compare$(EXESUFFIX): $(OPUSCOMPARE_OBJS)
$(LINK.o.cmdline)
celt/celt.o: CFLAGS += -DPACKAGE_VERSION='$(PACKAGE_VERSION)'
celt/celt.o: package_version
package_version: force
@if [ -x ./update_version ]; then \
./update_version || true; \
elif [ ! -e ./package_version ]; then \
echo 'PACKAGE_VERSION="unknown"' > ./package_version; \
fi
force:
clean:
rm -f opus_demo$(EXESUFFIX) opus_compare$(EXESUFFIX) $(TARGET) \
test_opus_api$(EXESUFFIX) test_opus_decode$(EXESUFFIX) \
test_opus_encode$(EXESUFFIX) test_opus_padding$(EXESUFFIX) \
$(OBJS) $(OPUSDEMO_OBJS) $(OPUSCOMPARE_OBJS) $(TESTOPUSAPI_OBJS) \
$(TESTOPUSDECODE_OBJS) $(TESTOPUSENCODE_OBJS) $(TESTOPUSPADDING_OBJS)
.PHONY: all lib clean

36
Frameworks/Opus/Opus/opus/Makefile.unix
View file

@ -12,11 +12,7 @@ CFLAGS := -DUSE_ALLOCA $(CFLAGS)
# These options affect performance
# HAVE_LRINTF: Use C99 intrinsics to speed up float-to-int conversion
# inline: Don't use the 'inline' keyword (for ANSI C compilers)
# restrict: Don't use the 'restrict' keyword (for pre-C99 compilers)
#CFLAGS := -DHAVE_LRINTF $(CFLAGS)
#CFLAGS := -Dinline= $(CFLAGS)
CFLAGS := -Drestrict= $(CFLAGS)
###################### END OF OPTIONS ######################
@ -30,6 +26,7 @@ ifdef FIXED_POINT
SILK_SOURCES += $(SILK_SOURCES_FIXED)
else
SILK_SOURCES += $(SILK_SOURCES_FLOAT)
OPUS_SOURCES += $(OPUS_SOURCES_FLOAT)
endif
EXESUFFIX =
@ -94,11 +91,23 @@ OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(SRCS_C))
OPUSDEMO_SRCS_C = src/opus_demo.c
OPUSDEMO_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(OPUSDEMO_SRCS_C))
TESTOPUSAPI_SRCS_C = tests/test_opus_api.c
TESTOPUSAPI_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(TESTOPUSAPI_SRCS_C))
TESTOPUSDECODE_SRCS_C = tests/test_opus_decode.c
TESTOPUSDECODE_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(TESTOPUSDECODE_SRCS_C))
TESTOPUSENCODE_SRCS_C = tests/test_opus_encode.c
TESTOPUSENCODE_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(TESTOPUSENCODE_SRCS_C))
TESTOPUSPADDING_SRCS_C = tests/test_opus_padding.c
TESTOPUSPADDING_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(TESTOPUSPADDING_SRCS_C))
OPUSCOMPARE_SRCS_C = src/opus_compare.c
OPUSCOMPARE_OBJS := $(patsubst %.c,%$(OBJSUFFIX),$(OPUSCOMPARE_SRCS_C))
# Rules
all: lib opus_demo opus_compare
all: lib opus_demo opus_compare test_opus_api test_opus_decode test_opus_encode test_opus_padding
lib: $(TARGET)
@ -108,6 +117,18 @@ $(TARGET): $(OBJS)
opus_demo$(EXESUFFIX): $(OPUSDEMO_OBJS) $(TARGET)
$(LINK.o.cmdline)
test_opus_api$(EXESUFFIX): $(TESTOPUSAPI_OBJS) $(TARGET)
$(LINK.o.cmdline)
test_opus_decode$(EXESUFFIX): $(TESTOPUSDECODE_OBJS) $(TARGET)
$(LINK.o.cmdline)
test_opus_encode$(EXESUFFIX): $(TESTOPUSENCODE_OBJS) $(TARGET)
$(LINK.o.cmdline)
test_opus_padding$(EXESUFFIX): $(TESTOPUSPADDING_OBJS) $(TARGET)
$(LINK.o.cmdline)
opus_compare$(EXESUFFIX): $(OPUSCOMPARE_OBJS)
$(LINK.o.cmdline)
@ -125,6 +146,9 @@ force:
clean:
rm -f opus_demo$(EXESUFFIX) opus_compare$(EXESUFFIX) $(TARGET) \
$(OBJS) $(OPUSDEMO_OBJS) $(OPUSCOMPARE_OBJS)
test_opus_api$(EXESUFFIX) test_opus_decode$(EXESUFFIX) \
test_opus_encode$(EXESUFFIX) test_opus_padding$(EXESUFFIX) \
$(OBJS) $(OPUSDEMO_OBJS) $(OPUSCOMPARE_OBJS) $(TESTOPUSAPI_OBJS) \
$(TESTOPUSDECODE_OBJS) $(TESTOPUSENCODE_OBJS) $(TESTOPUSPADDING_OBJS)
.PHONY: all lib clean

14
Frameworks/Opus/Opus/opus/README
View file

@ -13,14 +13,14 @@ soundtracks, portable music players, jukeboxes, and other applications that
have historically used high latency formats such as MP3, AAC, or Vorbis.
Opus is specified by IETF RFC 6716:
http://tools.ietf.org/html/rfc6716
https://tools.ietf.org/html/rfc6716
The Opus format and this implementation of it are subject to the royalty-
free patent and copyright licenses specified in the file COPYING.
This package implements a shared library for encoding and decoding raw Opus
bitstreams. Raw Opus bitstreams should be used over RTP according to
http://tools.ietf.org/html/draft-spittka-payload-rtp-opus
https://tools.ietf.org/html/rfc7587
The package also includes a number of test tools used for testing the
correct operation of the library. The bitstreams read/written by these
@ -29,7 +29,7 @@ additional debugging data and cannot support seeking.
Opus stored in files should use the Ogg encapsulation for Opus which is
described at:
http://wiki.xiph.org/OggOpus
https://wiki.xiph.org/OggOpus
An opus-tools package is available which provides encoding and decoding of
Ogg encapsulated Opus files and includes a number of useful features.
@ -37,7 +37,7 @@ Ogg encapsulated Opus files and includes a number of useful features.
Opus-tools can be found at:
https://git.xiph.org/?p=opus-tools.git
or on the main Opus website:
http://opus-codec.org/
https://opus-codec.org/
== Compiling libopus ==
@ -50,7 +50,7 @@ To build from the git repository, the following steps are necessary:
1) Clone the repository:
% git clone git://git.opus-codec.org/opus.git
% git clone https://git.xiph.org/opus.git
% cd opus
2) Compiling the source
@ -106,11 +106,11 @@ To run the integrated tests:
There is also collection of standard test vectors which are not
included in this package for size reasons but can be obtained from:
http://opus-codec.org/testvectors/opus_testvectors.tar.gz
https://opus-codec.org/testvectors/opus_testvectors.tar.gz
To run compare the code to these test vectors:
% curl -O http://opus-codec.org/testvectors/opus_testvectors.tar.gz
% curl -O https://opus-codec.org/testvectors/opus_testvectors.tar.gz
% tar -zxf opus_testvectors.tar.gz
% ./tests/run_vectors.sh ./ opus_testvectors 48000

625
Frameworks/Opus/Opus/opus/aclocal.m4 vendored
View file

File diff suppressed because it is too large Load diff

15
Frameworks/Opus/Opus/opus/celt/_kiss_fft_guts.h
View file

@ -65,10 +65,6 @@
do{ (m).r = ADD32(S_MUL((a).r,(b).r) , S_MUL((a).i,(b).i)); \
(m).i = SUB32(S_MUL((a).i,(b).r) , S_MUL((a).r,(b).i)); }while(0)
# define C_MUL4(m,a,b) \
do{ (m).r = SHR32(SUB32(S_MUL((a).r,(b).r) , S_MUL((a).i,(b).i)),2); \
(m).i = SHR32(ADD32(S_MUL((a).r,(b).i) , S_MUL((a).i,(b).r)),2); }while(0)
# define C_MULBYSCALAR( c, s ) \
do{ (c).r = S_MUL( (c).r , s ) ;\
(c).i = S_MUL( (c).i , s ) ; }while(0)
@ -94,6 +90,17 @@
do {(res).r = ADD32((res).r,(a).r); (res).i = SUB32((res).i,(a).i); \
}while(0)
#if defined(OPUS_ARM_INLINE_ASM)
#include "arm/kiss_fft_armv4.h"
#endif
#if defined(OPUS_ARM_INLINE_EDSP)
#include "arm/kiss_fft_armv5e.h"
#endif
#if defined(MIPSr1_ASM)
#include "mips/kiss_fft_mipsr1.h"
#endif
#else /* not FIXED_POINT*/
# define S_MUL(a,b) ( (a)*(b) )

48
Frameworks/Opus/Opus/opus/celt/arch.h
View file

@ -35,6 +35,7 @@
#define ARCH_H
#include "opus_types.h"
#include "opus_defines.h"
# if !defined(__GNUC_PREREQ)
# if defined(__GNUC__)&&defined(__GNUC_MINOR__)
@ -54,7 +55,7 @@
#ifdef __GNUC__
__attribute__((noreturn))
#endif
static inline void _celt_fatal(const char *str, const char *file, int line)
static OPUS_INLINE void _celt_fatal(const char *str, const char *file, int line)
{
fprintf (stderr, "Fatal (internal) error in %s, line %d: %s\n", file, line, str);
abort();
@ -68,11 +69,8 @@ static inline void _celt_fatal(const char *str, const char *file, int line)
#define IMUL32(a,b) ((a)*(b))
#define ABS(x) ((x) < 0 ? (-(x)) : (x)) /**< Absolute integer value. */
#define ABS16(x) ((x) < 0 ? (-(x)) : (x)) /**< Absolute 16-bit value. */
#define MIN16(a,b) ((a) < (b) ? (a) : (b)) /**< Minimum 16-bit value. */
#define MAX16(a,b) ((a) > (b) ? (a) : (b)) /**< Maximum 16-bit value. */
#define ABS32(x) ((x) < 0 ? (-(x)) : (x)) /**< Absolute 32-bit value. */
#define MIN32(a,b) ((a) < (b) ? (a) : (b)) /**< Minimum 32-bit value. */
#define MAX32(a,b) ((a) > (b) ? (a) : (b)) /**< Maximum 32-bit value. */
#define IMIN(a,b) ((a) < (b) ? (a) : (b)) /**< Minimum int value. */
@ -100,22 +98,30 @@ typedef opus_val32 celt_ener;
#define DB_SHIFT 10
#define EPSILON 1
#define VERY_SMALL 0
#define VERY_LARGE16 ((opus_val16)32767)
#define Q15_ONE ((opus_val16)32767)
#define SCALEIN(a) (a)
#define SCALEOUT(a) (a)
#define ABS16(x) ((x) < 0 ? (-(x)) : (x))
#define ABS32(x) ((x) < 0 ? (-(x)) : (x))
static OPUS_INLINE opus_int16 SAT16(opus_int32 x) {
return x > 32767 ? 32767 : x < -32768 ? -32768 : (opus_int16)x;
}
#ifdef FIXED_DEBUG
#include "fixed_debug.h"
#else
#include "fixed_generic.h"
#ifdef ARM5E_ASM
#include "fixed_arm5e.h"
#elif defined (ARM4_ASM)
#include "fixed_arm4.h"
#ifdef OPUS_ARM_INLINE_EDSP
#include "arm/fixed_armv5e.h"
#elif defined (OPUS_ARM_INLINE_ASM)
#include "arm/fixed_armv4.h"
#elif defined (BFIN_ASM)
#include "fixed_bfin.h"
#elif defined (TI_C5X_ASM)
@ -135,14 +141,35 @@ typedef float celt_sig;
typedef float celt_norm;
typedef float celt_ener;
#ifdef FLOAT_APPROX
/* This code should reliably detect NaN/inf even when -ffast-math is used.
Assumes IEEE 754 format. */
static OPUS_INLINE int celt_isnan(float x)
{
union {float f; opus_uint32 i;} in;
in.f = x;
return ((in.i>>23)&0xFF)==0xFF && (in.i&0x007FFFFF)!=0;
}
#else
#ifdef __FAST_MATH__
#error Cannot build libopus with -ffast-math unless FLOAT_APPROX is defined. This could result in crashes on extreme (e.g. NaN) input
#endif
#define celt_isnan(x) ((x)!=(x))
#endif
#define Q15ONE 1.0f
#define NORM_SCALING 1.f
#define EPSILON 1e-15f
#define VERY_SMALL 1e-30f
#define VERY_LARGE16 1e15f
#define Q15_ONE ((opus_val16)1.f)
/* This appears to be the same speed as C99's fabsf() but it's more portable. */
#define ABS16(x) ((float)fabs(x))
#define ABS32(x) ((float)fabs(x))
#define QCONST16(x,bits) (x)
#define QCONST32(x,bits) (x)
@ -161,6 +188,7 @@ typedef float celt_ener;
#define SHR(a,shift) (a)
#define SHL(a,shift) (a)
#define SATURATE(x,a) (x)
#define SATURATE16(x) (x)
#define ROUND16(a,shift) (a)
#define HALF16(x) (.5f*(x))
@ -180,8 +208,10 @@ typedef float celt_ener;
#define MULT32_32_Q31(a,b) ((a)*(b))
#define MAC16_32_Q15(c,a,b) ((c)+(a)*(b))
#define MAC16_32_Q16(c,a,b) ((c)+(a)*(b))
#define MULT16_16_Q11_32(a,b) ((a)*(b))
#define MULT16_16_Q11(a,b) ((a)*(b))
#define MULT16_16_Q13(a,b) ((a)*(b))
#define MULT16_16_Q14(a,b) ((a)*(b))
#define MULT16_16_Q15(a,b) ((a)*(b))
@ -196,6 +226,8 @@ typedef float celt_ener;
#define SCALEIN(a) ((a)*CELT_SIG_SCALE)
#define SCALEOUT(a) ((a)*(1/CELT_SIG_SCALE))
#define SIG2WORD16(x) (x)
#endif /* !FIXED_POINT */
#ifndef GLOBAL_STACK_SIZE

353
Frameworks/Opus/Opus/opus/celt/arm/arm2gnu.pl Executable file
View file

@ -0,0 +1,353 @@
#!/usr/bin/perl
# Copyright (C) 2002-2013 Xiph.org Foundation
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
#
# - Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
#
# - Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
# OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
# LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
# NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
my $bigend; # little/big endian
my $nxstack;
my $apple = 0;
my $symprefix = "";
$nxstack = 0;
eval 'exec /usr/local/bin/perl -S $0 ${1+"$@"}'
if $running_under_some_shell;
while ($ARGV[0] =~ /^-/) {
$_ = shift;
last if /^--$/;
if (/^-n$/) {
$nflag++;
next;
}
if (/^--apple$/) {
$apple = 1;
$symprefix = "_";
next;
}
die "I don't recognize this switch: $_\\n";
}
$printit++ unless $nflag;
$\ = "\n"; # automatically add newline on print
$n=0;
$thumb = 0; # ARM mode by default, not Thumb.
@proc_stack = ();
printf (" .syntax unified\n");
LINE:
while (<>) {
# For ADRLs we need to add a new line after the substituted one.
$addPadding = 0;
# First, we do not dare to touch *anything* inside double quotes, do we?
# Second, if you want a dollar character in the string,
# insert two of them -- that's how ARM C and assembler treat strings.
s/^([A-Za-z_]\w*)[ \t]+DCB[ \t]*\"/$1: .ascii \"/ && do { s/\$\$/\$/g; next };
s/\bDCB\b[ \t]*\"/.ascii \"/ && do { s/\$\$/\$/g; next };
s/^(\S+)\s+RN\s+(\S+)/$1 .req r$2/ && do { s/\$\$/\$/g; next };
# If there's nothing on a line but a comment, don't try to apply any further
# substitutions (this is a cheap hack to avoid mucking up the license header)
s/^([ \t]*);/$1@/ && do { s/\$\$/\$/g; next };
# If substituted -- leave immediately !
s/@/,:/;
s/;/@/;
while ( /@.*'/ ) {
s/(@.*)'/$1/g;
}
s/\{FALSE\}/0/g;
s/\{TRUE\}/1/g;
s/\{(\w\w\w\w+)\}/$1/g;
s/\bINCLUDE[ \t]*([^ \t\n]+)/.include \"$1\"/;
s/\bGET[ \t]*([^ \t\n]+)/.include \"${ my $x=$1; $x =~ s|\.s|-gnu.S|; \$x }\"/;
s/\bIMPORT\b/.extern/;
s/\bEXPORT\b\s*/.global $symprefix/;
s/^(\s+)\[/$1IF/;
s/^(\s+)\|/$1ELSE/;
s/^(\s+)\]/$1ENDIF/;
s/IF *:DEF:/ .ifdef/;
s/IF *:LNOT: *:DEF:/ .ifndef/;
s/ELSE/ .else/;
s/ENDIF/ .endif/;
if( /\bIF\b/ ) {
s/\bIF\b/ .if/;
s/=/==/;
}
if ( $n == 2) {
s/\$/\\/g;
}
if ($n == 1) {
s/\$//g;
s/label//g;
$n = 2;
}
if ( /MACRO/ ) {
s/MACRO *\n/.macro/;
$n=1;
}
if ( /\bMEND\b/ ) {
s/\bMEND\b/.endm/;
$n=0;
}
# ".rdata" doesn't work in 'as' version 2.13.2, as it is ".rodata" there.
#
if ( /\bAREA\b/ ) {
my $align;
$align = "2";
if ( /ALIGN=(\d+)/ ) {
$align = $1;
}
if ( /CODE/ ) {
$nxstack = 1;
}
s/^(.+)CODE(.+)READONLY(.*)/ .text/;
s/^(.+)DATA(.+)READONLY(.*)/ .section .rdata/;
s/^(.+)\|\|\.data\|\|(.+)/ .data/;
s/^(.+)\|\|\.bss\|\|(.+)/ .bss/;
s/$/; .p2align $align/;
# Enable NEON instructions but don't produce a binary that requires
# ARMv7. RVCT does not have equivalent directives, so we just do this
# for all CODE areas.
if ( /.text/ ) {
# Separating .arch, .fpu, etc., by semicolons does not work (gas
# thinks the semicolon is part of the arch name, even when there's
# whitespace separating them). Sadly this means our line numbers
# won't match the original source file (we could use the .line
# directive, which is documented to be obsolete, but then gdb will
# show the wrong line in the translated source file).
s/$/; .arch armv7-a\n .fpu neon\n .object_arch armv4t/ unless ($apple);
}
}
s/\|\|\.constdata\$(\d+)\|\|/.L_CONST$1/; # ||.constdata$3||
s/\|\|\.bss\$(\d+)\|\|/.L_BSS$1/; # ||.bss$2||
s/\|\|\.data\$(\d+)\|\|/.L_DATA$1/; # ||.data$2||
s/\|\|([a-zA-Z0-9_]+)\@([a-zA-Z0-9_]+)\|\|/@ $&/;
s/^(\s+)\%(\s)/ .space $1/;
s/\|(.+)\.(\d+)\|/\.$1_$2/; # |L80.123| -> .L80_123
s/\bCODE32\b/.code 32/ && do {$thumb = 0};
s/\bCODE16\b/.code 16/ && do {$thumb = 1};
if (/\bPROC\b/)
{
my $prefix;
my $proc;
/^([A-Za-z_\.]\w+)\b/;
$proc = $1;
$prefix = "";
if ($proc)
{
$prefix = $prefix.sprintf("\t.type\t%s, %%function; ",$proc) unless ($apple);
# Make sure we $prefix isn't empty here (for the $apple case).
# We handle mangling the label here, make sure it doesn't match
# the label handling below (if $prefix would be empty).
$prefix = "; ";
push(@proc_stack, $proc);
s/^[A-Za-z_\.]\w+/$symprefix$&:/;
}
$prefix = $prefix."\t.thumb_func; " if ($thumb);
s/\bPROC\b/@ $&/;
$_ = $prefix.$_;
}
s/^(\s*)(S|Q|SH|U|UQ|UH)ASX\b/$1$2ADDSUBX/;
s/^(\s*)(S|Q|SH|U|UQ|UH)SAX\b/$1$2SUBADDX/;
if (/\bENDP\b/)
{
my $proc;
s/\bENDP\b/@ $&/;
$proc = pop(@proc_stack);
$_ = "\t.size $proc, .-$proc".$_ if ($proc && !$apple);
}
s/\bSUBT\b/@ $&/;
s/\bDATA\b/@ $&/; # DATA directive is deprecated -- Asm guide, p.7-25
s/\bKEEP\b/@ $&/;
s/\bEXPORTAS\b/@ $&/;
s/\|\|(.)+\bEQU\b/@ $&/;
s/\|\|([\w\$]+)\|\|/$1/;
s/\bENTRY\b/@ $&/;
s/\bASSERT\b/@ $&/;
s/\bGBLL\b/@ $&/;
s/\bGBLA\b/@ $&/;
s/^\W+OPT\b/@ $&/;
s/:OR:/|/g;
s/:SHL:/<</g;
s/:SHR:/>>/g;
s/:AND:/&/g;
s/:LAND:/&&/g;
s/CPSR/cpsr/;
s/SPSR/spsr/;
s/ALIGN$/.balign 4/;
s/ALIGN\s+([0-9x]+)$/.balign $1/;
s/psr_cxsf/psr_all/;
s/LTORG/.ltorg/;
s/^([A-Za-z_]\w*)[ \t]+EQU/ .set $1,/;
s/^([A-Za-z_]\w*)[ \t]+SETL/ .set $1,/;
s/^([A-Za-z_]\w*)[ \t]+SETA/ .set $1,/;
s/^([A-Za-z_]\w*)[ \t]+\*/ .set $1,/;
# {PC} + 0xdeadfeed --> . + 0xdeadfeed
s/\{PC\} \+/ \. +/;
# Single hex constant on the line !
#
# >>> NOTE <<<
# Double-precision floats in gcc are always mixed-endian, which means
# bytes in two words are little-endian, but words are big-endian.
# So, 0x0000deadfeed0000 would be stored as 0x0000dead at low address
# and 0xfeed0000 at high address.
#
s/\bDCFD\b[ \t]+0x([a-fA-F0-9]{8})([a-fA-F0-9]{8})/.long 0x$1, 0x$2/;
# Only decimal constants on the line, no hex !
s/\bDCFD\b[ \t]+([0-9\.\-]+)/.double $1/;
# Single hex constant on the line !
# s/\bDCFS\b[ \t]+0x([a-f0-9]{8})([a-f0-9]{8})/.long 0x$1, 0x$2/;
# Only decimal constants on the line, no hex !
# s/\bDCFS\b[ \t]+([0-9\.\-]+)/.double $1/;
s/\bDCFS[ \t]+0x/.word 0x/;
s/\bDCFS\b/.float/;
s/^([A-Za-z_]\w*)[ \t]+DCD/$1 .word/;
s/\bDCD\b/.word/;
s/^([A-Za-z_]\w*)[ \t]+DCW/$1 .short/;
s/\bDCW\b/.short/;
s/^([A-Za-z_]\w*)[ \t]+DCB/$1 .byte/;
s/\bDCB\b/.byte/;
s/^([A-Za-z_]\w*)[ \t]+\%/.comm $1,/;
s/^[A-Za-z_\.]\w+/$&:/;
s/^(\d+)/$1:/;
s/\%(\d+)/$1b_or_f/;
s/\%[Bb](\d+)/$1b/;
s/\%[Ff](\d+)/$1f/;
s/\%[Ff][Tt](\d+)/$1f/;
s/&([\dA-Fa-f]+)/0x$1/;
if ( /\b2_[01]+\b/ ) {
s/\b2_([01]+)\b/conv$1&&&&/g;
while ( /[01][01][01][01]&&&&/ ) {
s/0000&&&&/&&&&0/g;
s/0001&&&&/&&&&1/g;
s/0010&&&&/&&&&2/g;
s/0011&&&&/&&&&3/g;
s/0100&&&&/&&&&4/g;
s/0101&&&&/&&&&5/g;
s/0110&&&&/&&&&6/g;
s/0111&&&&/&&&&7/g;
s/1000&&&&/&&&&8/g;
s/1001&&&&/&&&&9/g;
s/1010&&&&/&&&&A/g;
s/1011&&&&/&&&&B/g;
s/1100&&&&/&&&&C/g;
s/1101&&&&/&&&&D/g;
s/1110&&&&/&&&&E/g;
s/1111&&&&/&&&&F/g;
}
s/000&&&&/&&&&0/g;
s/001&&&&/&&&&1/g;
s/010&&&&/&&&&2/g;
s/011&&&&/&&&&3/g;
s/100&&&&/&&&&4/g;
s/101&&&&/&&&&5/g;
s/110&&&&/&&&&6/g;
s/111&&&&/&&&&7/g;
s/00&&&&/&&&&0/g;
s/01&&&&/&&&&1/g;
s/10&&&&/&&&&2/g;
s/11&&&&/&&&&3/g;
s/0&&&&/&&&&0/g;
s/1&&&&/&&&&1/g;
s/conv&&&&/0x/g;
}
if ( /commandline/)
{
if( /-bigend/)
{
$bigend=1;
}
}
if ( /\bDCDU\b/ )
{
my $cmd=$_;
my $value;
my $prefix;
my $w1;
my $w2;
my $w3;
my $w4;
s/\s+DCDU\b/@ $&/;
$cmd =~ /\bDCDU\b\s+0x(\d+)/;
$value = $1;
$value =~ /(\w\w)(\w\w)(\w\w)(\w\w)/;
$w1 = $1;
$w2 = $2;
$w3 = $3;
$w4 = $4;
if( $bigend ne "")
{
# big endian
$prefix = "\t.byte\t0x".$w1.";".
"\t.byte\t0x".$w2.";".
"\t.byte\t0x".$w3.";".
"\t.byte\t0x".$w4."; ";
}
else
{
# little endian
$prefix = "\t.byte\t0x".$w4.";".
"\t.byte\t0x".$w3.";".
"\t.byte\t0x".$w2.";".
"\t.byte\t0x".$w1."; ";
}
$_=$prefix.$_;
}
if ( /\badrl\b/i )
{
s/\badrl\s+(\w+)\s*,\s*(\w+)/ldr $1,=$2/i;
$addPadding = 1;
}
s/\bEND\b/@ END/;
} continue {
printf ("%s", $_) if $printit;
if ($addPadding != 0)
{
printf (" mov r0,r0\n");
$addPadding = 0;
}
}
#If we had a code section, mark that this object doesn't need an executable
# stack.
if ($nxstack && !$apple) {
printf (" .section\t.note.GNU-stack,\"\",\%\%progbits\n");
}

121
Frameworks/Opus/Opus/opus/celt/arm/arm_celt_map.c Normal file
View file

@ -0,0 +1,121 @@
/* Copyright (c) 2010 Xiph.Org Foundation
* Copyright (c) 2013 Parrot */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "pitch.h"
#include "kiss_fft.h"
#include "mdct.h"
#if defined(OPUS_HAVE_RTCD)
# if defined(FIXED_POINT)
opus_val32 (*const CELT_PITCH_XCORR_IMPL[OPUS_ARCHMASK+1])(const opus_val16 *,
const opus_val16 *, opus_val32 *, int , int) = {
celt_pitch_xcorr_c, /* ARMv4 */
MAY_HAVE_EDSP(celt_pitch_xcorr), /* EDSP */
MAY_HAVE_MEDIA(celt_pitch_xcorr), /* Media */
MAY_HAVE_NEON(celt_pitch_xcorr) /* NEON */
};
# else /* !FIXED_POINT */
# if defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
void (*const CELT_PITCH_XCORR_IMPL[OPUS_ARCHMASK+1])(const opus_val16 *,
const opus_val16 *, opus_val32 *, int, int) = {
celt_pitch_xcorr_c, /* ARMv4 */
celt_pitch_xcorr_c, /* EDSP */
celt_pitch_xcorr_c, /* Media */
celt_pitch_xcorr_float_neon /* Neon */
};
# endif
# endif /* FIXED_POINT */
# if defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
# if defined(HAVE_ARM_NE10)
# if defined(CUSTOM_MODES)
int (*const OPUS_FFT_ALLOC_ARCH_IMPL[OPUS_ARCHMASK+1])(kiss_fft_state *st) = {
opus_fft_alloc_arch_c, /* ARMv4 */
opus_fft_alloc_arch_c, /* EDSP */
opus_fft_alloc_arch_c, /* Media */
opus_fft_alloc_arm_neon /* Neon with NE10 library support */
};
void (*const OPUS_FFT_FREE_ARCH_IMPL[OPUS_ARCHMASK+1])(kiss_fft_state *st) = {
opus_fft_free_arch_c, /* ARMv4 */
opus_fft_free_arch_c, /* EDSP */
opus_fft_free_arch_c, /* Media */
opus_fft_free_arm_neon /* Neon with NE10 */
};
# endif /* CUSTOM_MODES */
void (*const OPUS_FFT[OPUS_ARCHMASK+1])(const kiss_fft_state *cfg,
const kiss_fft_cpx *fin,
kiss_fft_cpx *fout) = {
opus_fft_c, /* ARMv4 */
opus_fft_c, /* EDSP */
opus_fft_c, /* Media */
opus_fft_neon /* Neon with NE10 */
};
void (*const OPUS_IFFT[OPUS_ARCHMASK+1])(const kiss_fft_state *cfg,
const kiss_fft_cpx *fin,
kiss_fft_cpx *fout) = {
opus_ifft_c, /* ARMv4 */
opus_ifft_c, /* EDSP */
opus_ifft_c, /* Media */
opus_ifft_neon /* Neon with NE10 */
};
void (*const CLT_MDCT_FORWARD_IMPL[OPUS_ARCHMASK+1])(const mdct_lookup *l,
kiss_fft_scalar *in,
kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 *window,
int overlap, int shift,
int stride, int arch) = {
clt_mdct_forward_c, /* ARMv4 */
clt_mdct_forward_c, /* EDSP */
clt_mdct_forward_c, /* Media */
clt_mdct_forward_neon /* Neon with NE10 */
};
void (*const CLT_MDCT_BACKWARD_IMPL[OPUS_ARCHMASK+1])(const mdct_lookup *l,
kiss_fft_scalar *in,
kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 *window,
int overlap, int shift,
int stride, int arch) = {
clt_mdct_backward_c, /* ARMv4 */
clt_mdct_backward_c, /* EDSP */
clt_mdct_backward_c, /* Media */
clt_mdct_backward_neon /* Neon with NE10 */
};
# endif /* HAVE_ARM_NE10 */
# endif /* OPUS_ARM_MAY_HAVE_NEON_INTR */
#endif /* OPUS_HAVE_RTCD */

174
Frameworks/Opus/Opus/opus/celt/arm/armcpu.c Normal file
View file

@ -0,0 +1,174 @@
/* Copyright (c) 2010 Xiph.Org Foundation
* Copyright (c) 2013 Parrot */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
/* Original code from libtheora modified to suit to Opus */
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#ifdef OPUS_HAVE_RTCD
#include "armcpu.h"
#include "cpu_support.h"
#include "os_support.h"
#include "opus_types.h"
#define OPUS_CPU_ARM_V4 (1)
#define OPUS_CPU_ARM_EDSP (1<<1)
#define OPUS_CPU_ARM_MEDIA (1<<2)
#define OPUS_CPU_ARM_NEON (1<<3)
#if defined(_MSC_VER)
/*For GetExceptionCode() and EXCEPTION_ILLEGAL_INSTRUCTION.*/
# define WIN32_LEAN_AND_MEAN
# define WIN32_EXTRA_LEAN
# include <windows.h>
static OPUS_INLINE opus_uint32 opus_cpu_capabilities(void){
opus_uint32 flags;
flags=0;
/* MSVC has no OPUS_INLINE __asm support for ARM, but it does let you __emit
* instructions via their assembled hex code.
* All of these instructions should be essentially nops. */
# if defined(OPUS_ARM_MAY_HAVE_EDSP)
__try{
/*PLD [r13]*/
__emit(0xF5DDF000);
flags|=OPUS_CPU_ARM_EDSP;
}
__except(GetExceptionCode()==EXCEPTION_ILLEGAL_INSTRUCTION){
/*Ignore exception.*/
}
# if defined(OPUS_ARM_MAY_HAVE_MEDIA)
__try{
/*SHADD8 r3,r3,r3*/
__emit(0xE6333F93);
flags|=OPUS_CPU_ARM_MEDIA;
}
__except(GetExceptionCode()==EXCEPTION_ILLEGAL_INSTRUCTION){
/*Ignore exception.*/
}
# if defined(OPUS_ARM_MAY_HAVE_NEON) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
__try{
/*VORR q0,q0,q0*/
__emit(0xF2200150);
flags|=OPUS_CPU_ARM_NEON;
}
__except(GetExceptionCode()==EXCEPTION_ILLEGAL_INSTRUCTION){
/*Ignore exception.*/
}
# endif
# endif
# endif
return flags;
}
#elif defined(__linux__)
/* Linux based */
opus_uint32 opus_cpu_capabilities(void)
{
opus_uint32 flags = 0;
FILE *cpuinfo;
/* Reading /proc/self/auxv would be easier, but that doesn't work reliably on
* Android */
cpuinfo = fopen("/proc/cpuinfo", "r");
if(cpuinfo != NULL)
{
/* 512 should be enough for anybody (it's even enough for all the flags that
* x86 has accumulated... so far). */
char buf[512];
while(fgets(buf, 512, cpuinfo) != NULL)
{
# if defined(OPUS_ARM_MAY_HAVE_EDSP) || defined(OPUS_ARM_MAY_HAVE_NEON) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
/* Search for edsp and neon flag */
if(memcmp(buf, "Features", 8) == 0)
{
char *p;
# if defined(OPUS_ARM_MAY_HAVE_EDSP)
p = strstr(buf, " edsp");
if(p != NULL && (p[5] == ' ' || p[5] == '\n'))
flags |= OPUS_CPU_ARM_EDSP;
# endif
# if defined(OPUS_ARM_MAY_HAVE_NEON) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
p = strstr(buf, " neon");
if(p != NULL && (p[5] == ' ' || p[5] == '\n'))
flags |= OPUS_CPU_ARM_NEON;
# endif
}
# endif
# if defined(OPUS_ARM_MAY_HAVE_MEDIA)
/* Search for media capabilities (>= ARMv6) */
if(memcmp(buf, "CPU architecture:", 17) == 0)
{
int version;
version = atoi(buf+17);
if(version >= 6)
flags |= OPUS_CPU_ARM_MEDIA;
}
# endif
}
fclose(cpuinfo);
}
return flags;
}
#else
/* The feature registers which can tell us what the processor supports are
* accessible in priveleged modes only, so we can't have a general user-space
* detection method like on x86.*/
# error "Configured to use ARM asm but no CPU detection method available for " \
"your platform. Reconfigure with --disable-rtcd (or send patches)."
#endif
int opus_select_arch(void)
{
opus_uint32 flags = opus_cpu_capabilities();
int arch = 0;
if(!(flags & OPUS_CPU_ARM_EDSP))
return arch;
arch++;
if(!(flags & OPUS_CPU_ARM_MEDIA))
return arch;
arch++;
if(!(flags & OPUS_CPU_ARM_NEON))
return arch;
arch++;
return arch;
}
#endif

71
Frameworks/Opus/Opus/opus/celt/arm/armcpu.h Normal file
View file

@ -0,0 +1,71 @@
/* Copyright (c) 2010 Xiph.Org Foundation
* Copyright (c) 2013 Parrot */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#if !defined(ARMCPU_H)
# define ARMCPU_H
# if defined(OPUS_ARM_MAY_HAVE_EDSP)
# define MAY_HAVE_EDSP(name) name ## _edsp
# else
# define MAY_HAVE_EDSP(name) name ## _c
# endif
# if defined(OPUS_ARM_MAY_HAVE_MEDIA)
# define MAY_HAVE_MEDIA(name) name ## _media
# else
# define MAY_HAVE_MEDIA(name) MAY_HAVE_EDSP(name)
# endif
# if defined(OPUS_ARM_MAY_HAVE_NEON)
# define MAY_HAVE_NEON(name) name ## _neon
# else
# define MAY_HAVE_NEON(name) MAY_HAVE_MEDIA(name)
# endif
# if defined(OPUS_ARM_PRESUME_EDSP)
# define PRESUME_EDSP(name) name ## _edsp
# else
# define PRESUME_EDSP(name) name ## _c
# endif
# if defined(OPUS_ARM_PRESUME_MEDIA)
# define PRESUME_MEDIA(name) name ## _media
# else
# define PRESUME_MEDIA(name) PRESUME_EDSP(name)
# endif
# if defined(OPUS_ARM_PRESUME_NEON)
# define PRESUME_NEON(name) name ## _neon
# else
# define PRESUME_NEON(name) PRESUME_MEDIA(name)
# endif
# if defined(OPUS_HAVE_RTCD)
int opus_select_arch(void);
# endif
#endif

37
Frameworks/Opus/Opus/opus/celt/arm/armopts.s.in Normal file
View file

@ -0,0 +1,37 @@
/* Copyright (C) 2013 Mozilla Corporation */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
; Set the following to 1 if we have EDSP instructions
; (LDRD/STRD, etc., ARMv5E and later).
OPUS_ARM_MAY_HAVE_EDSP * @OPUS_ARM_MAY_HAVE_EDSP@
; Set the following to 1 if we have ARMv6 media instructions.
OPUS_ARM_MAY_HAVE_MEDIA * @OPUS_ARM_MAY_HAVE_MEDIA@
; Set the following to 1 if we have NEON (some ARMv7)
OPUS_ARM_MAY_HAVE_NEON * @OPUS_ARM_MAY_HAVE_NEON@
END

174
Frameworks/Opus/Opus/opus/celt/arm/celt_ne10_fft.c Normal file
View file

@ -0,0 +1,174 @@
/* Copyright (c) 2015 Xiph.Org Foundation
Written by Viswanath Puttagunta */
/**
@file celt_ne10_fft.c
@brief ARM Neon optimizations for fft using NE10 library
*/
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef SKIP_CONFIG_H
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#endif
#include <NE10_init.h>
#include <NE10_dsp.h>
#include "os_support.h"
#include "kiss_fft.h"
#include "stack_alloc.h"
#if !defined(FIXED_POINT)
# define NE10_FFT_ALLOC_C2C_TYPE_NEON ne10_fft_alloc_c2c_float32_neon
# define NE10_FFT_CFG_TYPE_T ne10_fft_cfg_float32_t
# define NE10_FFT_STATE_TYPE_T ne10_fft_state_float32_t
# define NE10_FFT_DESTROY_C2C_TYPE ne10_fft_destroy_c2c_float32
# define NE10_FFT_CPX_TYPE_T ne10_fft_cpx_float32_t
# define NE10_FFT_C2C_1D_TYPE_NEON ne10_fft_c2c_1d_float32_neon
#else
# define NE10_FFT_ALLOC_C2C_TYPE_NEON(nfft) ne10_fft_alloc_c2c_int32_neon(nfft)
# define NE10_FFT_CFG_TYPE_T ne10_fft_cfg_int32_t
# define NE10_FFT_STATE_TYPE_T ne10_fft_state_int32_t
# define NE10_FFT_DESTROY_C2C_TYPE ne10_fft_destroy_c2c_int32
# define NE10_FFT_DESTROY_C2C_TYPE ne10_fft_destroy_c2c_int32
# define NE10_FFT_CPX_TYPE_T ne10_fft_cpx_int32_t
# define NE10_FFT_C2C_1D_TYPE_NEON ne10_fft_c2c_1d_int32_neon
#endif
#if defined(CUSTOM_MODES)
/* nfft lengths in NE10 that support scaled fft */
# define NE10_FFTSCALED_SUPPORT_MAX 4
static const int ne10_fft_scaled_support[NE10_FFTSCALED_SUPPORT_MAX] = {
480, 240, 120, 60
};
int opus_fft_alloc_arm_neon(kiss_fft_state *st)
{
int i;
size_t memneeded = sizeof(struct arch_fft_state);
st->arch_fft = (arch_fft_state *)opus_alloc(memneeded);
if (!st->arch_fft)
return -1;
for (i = 0; i < NE10_FFTSCALED_SUPPORT_MAX; i++) {
if(st->nfft == ne10_fft_scaled_support[i])
break;
}
if (i == NE10_FFTSCALED_SUPPORT_MAX) {
/* This nfft length (scaled fft) is not supported in NE10 */
st->arch_fft->is_supported = 0;
st->arch_fft->priv = NULL;
}
else {
st->arch_fft->is_supported = 1;
st->arch_fft->priv = (void *)NE10_FFT_ALLOC_C2C_TYPE_NEON(st->nfft);
if (st->arch_fft->priv == NULL) {
return -1;
}
}
return 0;
}
void opus_fft_free_arm_neon(kiss_fft_state *st)
{
NE10_FFT_CFG_TYPE_T cfg;
if (!st->arch_fft)
return;
cfg = (NE10_FFT_CFG_TYPE_T)st->arch_fft->priv;
if (cfg)
NE10_FFT_DESTROY_C2C_TYPE(cfg);
opus_free(st->arch_fft);
}
#endif
void opus_fft_neon(const kiss_fft_state *st,
const kiss_fft_cpx *fin,
kiss_fft_cpx *fout)
{
NE10_FFT_STATE_TYPE_T state;
NE10_FFT_CFG_TYPE_T cfg = &state;
VARDECL(NE10_FFT_CPX_TYPE_T, buffer);
SAVE_STACK;
ALLOC(buffer, st->nfft, NE10_FFT_CPX_TYPE_T);
if (!st->arch_fft->is_supported) {
/* This nfft length (scaled fft) not supported in NE10 */
opus_fft_c(st, fin, fout);
}
else {
memcpy((void *)cfg, st->arch_fft->priv, sizeof(NE10_FFT_STATE_TYPE_T));
state.buffer = (NE10_FFT_CPX_TYPE_T *)&buffer[0];
#if !defined(FIXED_POINT)
state.is_forward_scaled = 1;
NE10_FFT_C2C_1D_TYPE_NEON((NE10_FFT_CPX_TYPE_T *)fout,
(NE10_FFT_CPX_TYPE_T *)fin,
cfg, 0);
#else
NE10_FFT_C2C_1D_TYPE_NEON((NE10_FFT_CPX_TYPE_T *)fout,
(NE10_FFT_CPX_TYPE_T *)fin,
cfg, 0, 1);
#endif
}
RESTORE_STACK;
}
void opus_ifft_neon(const kiss_fft_state *st,
const kiss_fft_cpx *fin,
kiss_fft_cpx *fout)
{
NE10_FFT_STATE_TYPE_T state;
NE10_FFT_CFG_TYPE_T cfg = &state;
VARDECL(NE10_FFT_CPX_TYPE_T, buffer);
SAVE_STACK;
ALLOC(buffer, st->nfft, NE10_FFT_CPX_TYPE_T);
if (!st->arch_fft->is_supported) {
/* This nfft length (scaled fft) not supported in NE10 */
opus_ifft_c(st, fin, fout);
}
else {
memcpy((void *)cfg, st->arch_fft->priv, sizeof(NE10_FFT_STATE_TYPE_T));
state.buffer = (NE10_FFT_CPX_TYPE_T *)&buffer[0];
#if !defined(FIXED_POINT)
state.is_backward_scaled = 0;
NE10_FFT_C2C_1D_TYPE_NEON((NE10_FFT_CPX_TYPE_T *)fout,
(NE10_FFT_CPX_TYPE_T *)fin,
cfg, 1);
#else
NE10_FFT_C2C_1D_TYPE_NEON((NE10_FFT_CPX_TYPE_T *)fout,
(NE10_FFT_CPX_TYPE_T *)fin,
cfg, 1, 0);
#endif
}
RESTORE_STACK;
}

258
Frameworks/Opus/Opus/opus/celt/arm/celt_ne10_mdct.c Normal file
View file

@ -0,0 +1,258 @@
/* Copyright (c) 2015 Xiph.Org Foundation
Written by Viswanath Puttagunta */
/**
@file celt_ne10_mdct.c
@brief ARM Neon optimizations for mdct using NE10 library
*/
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef SKIP_CONFIG_H
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#endif
#include "kiss_fft.h"
#include "_kiss_fft_guts.h"
#include "mdct.h"
#include "stack_alloc.h"
void clt_mdct_forward_neon(const mdct_lookup *l,
kiss_fft_scalar *in,
kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 *window,
int overlap, int shift, int stride, int arch)
{
int i;
int N, N2, N4;
VARDECL(kiss_fft_scalar, f);
VARDECL(kiss_fft_cpx, f2);
const kiss_fft_state *st = l->kfft[shift];
const kiss_twiddle_scalar *trig;
SAVE_STACK;
N = l->n;
trig = l->trig;
for (i=0;i<shift;i++)
{
N >>= 1;
trig += N;
}
N2 = N>>1;
N4 = N>>2;
ALLOC(f, N2, kiss_fft_scalar);
ALLOC(f2, N4, kiss_fft_cpx);
/* Consider the input to be composed of four blocks: [a, b, c, d] */
/* Window, shuffle, fold */
{
/* Temp pointers to make it really clear to the compiler what we're doing */
const kiss_fft_scalar * OPUS_RESTRICT xp1 = in+(overlap>>1);
const kiss_fft_scalar * OPUS_RESTRICT xp2 = in+N2-1+(overlap>>1);
kiss_fft_scalar * OPUS_RESTRICT yp = f;
const opus_val16 * OPUS_RESTRICT wp1 = window+(overlap>>1);
const opus_val16 * OPUS_RESTRICT wp2 = window+(overlap>>1)-1;
for(i=0;i<((overlap+3)>>2);i++)
{
/* Real part arranged as -d-cR, Imag part arranged as -b+aR*/
*yp++ = MULT16_32_Q15(*wp2, xp1[N2]) + MULT16_32_Q15(*wp1,*xp2);
*yp++ = MULT16_32_Q15(*wp1, *xp1) - MULT16_32_Q15(*wp2, xp2[-N2]);
xp1+=2;
xp2-=2;
wp1+=2;
wp2-=2;
}
wp1 = window;
wp2 = window+overlap-1;
for(;i<N4-((overlap+3)>>2);i++)
{
/* Real part arranged as a-bR, Imag part arranged as -c-dR */
*yp++ = *xp2;
*yp++ = *xp1;
xp1+=2;
xp2-=2;
}
for(;i<N4;i++)
{
/* Real part arranged as a-bR, Imag part arranged as -c-dR */
*yp++ = -MULT16_32_Q15(*wp1, xp1[-N2]) + MULT16_32_Q15(*wp2, *xp2);
*yp++ = MULT16_32_Q15(*wp2, *xp1) + MULT16_32_Q15(*wp1, xp2[N2]);
xp1+=2;
xp2-=2;
wp1+=2;
wp2-=2;
}
}
/* Pre-rotation */
{
kiss_fft_scalar * OPUS_RESTRICT yp = f;
const kiss_twiddle_scalar *t = &trig[0];
for(i=0;i<N4;i++)
{
kiss_fft_cpx yc;
kiss_twiddle_scalar t0, t1;
kiss_fft_scalar re, im, yr, yi;
t0 = t[i];
t1 = t[N4+i];
re = *yp++;
im = *yp++;
yr = S_MUL(re,t0) - S_MUL(im,t1);
yi = S_MUL(im,t0) + S_MUL(re,t1);
yc.r = yr;
yc.i = yi;
f2[i] = yc;
}
}
opus_fft(st, f2, (kiss_fft_cpx *)f, arch);
/* Post-rotate */
{
/* Temp pointers to make it really clear to the compiler what we're doing */
const kiss_fft_cpx * OPUS_RESTRICT fp = (kiss_fft_cpx *)f;
kiss_fft_scalar * OPUS_RESTRICT yp1 = out;
kiss_fft_scalar * OPUS_RESTRICT yp2 = out+stride*(N2-1);
const kiss_twiddle_scalar *t = &trig[0];
/* Temp pointers to make it really clear to the compiler what we're doing */
for(i=0;i<N4;i++)
{
kiss_fft_scalar yr, yi;
yr = S_MUL(fp->i,t[N4+i]) - S_MUL(fp->r,t[i]);
yi = S_MUL(fp->r,t[N4+i]) + S_MUL(fp->i,t[i]);
*yp1 = yr;
*yp2 = yi;
fp++;
yp1 += 2*stride;
yp2 -= 2*stride;
}
}
RESTORE_STACK;
}
void clt_mdct_backward_neon(const mdct_lookup *l,
kiss_fft_scalar *in,
kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 * OPUS_RESTRICT window,
int overlap, int shift, int stride, int arch)
{
int i;
int N, N2, N4;
VARDECL(kiss_fft_scalar, f);
const kiss_twiddle_scalar *trig;
const kiss_fft_state *st = l->kfft[shift];
N = l->n;
trig = l->trig;
for (i=0;i<shift;i++)
{
N >>= 1;
trig += N;
}
N2 = N>>1;
N4 = N>>2;
ALLOC(f, N2, kiss_fft_scalar);
/* Pre-rotate */
{
/* Temp pointers to make it really clear to the compiler what we're doing */
const kiss_fft_scalar * OPUS_RESTRICT xp1 = in;
const kiss_fft_scalar * OPUS_RESTRICT xp2 = in+stride*(N2-1);
kiss_fft_scalar * OPUS_RESTRICT yp = f;
const kiss_twiddle_scalar * OPUS_RESTRICT t = &trig[0];
for(i=0;i<N4;i++)
{
kiss_fft_scalar yr, yi;
yr = S_MUL(*xp2, t[i]) + S_MUL(*xp1, t[N4+i]);
yi = S_MUL(*xp1, t[i]) - S_MUL(*xp2, t[N4+i]);
yp[2*i] = yr;
yp[2*i+1] = yi;
xp1+=2*stride;
xp2-=2*stride;
}
}
opus_ifft(st, (kiss_fft_cpx *)f, (kiss_fft_cpx*)(out+(overlap>>1)), arch);
/* Post-rotate and de-shuffle from both ends of the buffer at once to make
it in-place. */
{
kiss_fft_scalar * yp0 = out+(overlap>>1);
kiss_fft_scalar * yp1 = out+(overlap>>1)+N2-2;
const kiss_twiddle_scalar *t = &trig[0];
/* Loop to (N4+1)>>1 to handle odd N4. When N4 is odd, the
middle pair will be computed twice. */
for(i=0;i<(N4+1)>>1;i++)
{
kiss_fft_scalar re, im, yr, yi;
kiss_twiddle_scalar t0, t1;
re = yp0[0];
im = yp0[1];
t0 = t[i];
t1 = t[N4+i];
/* We'd scale up by 2 here, but instead it's done when mixing the windows */
yr = S_MUL(re,t0) + S_MUL(im,t1);
yi = S_MUL(re,t1) - S_MUL(im,t0);
re = yp1[0];
im = yp1[1];
yp0[0] = yr;
yp1[1] = yi;
t0 = t[(N4-i-1)];
t1 = t[(N2-i-1)];
/* We'd scale up by 2 here, but instead it's done when mixing the windows */
yr = S_MUL(re,t0) + S_MUL(im,t1);
yi = S_MUL(re,t1) - S_MUL(im,t0);
yp1[0] = yr;
yp0[1] = yi;
yp0 += 2;
yp1 -= 2;
}
}
/* Mirror on both sides for TDAC */
{
kiss_fft_scalar * OPUS_RESTRICT xp1 = out+overlap-1;
kiss_fft_scalar * OPUS_RESTRICT yp1 = out;
const opus_val16 * OPUS_RESTRICT wp1 = window;
const opus_val16 * OPUS_RESTRICT wp2 = window+overlap-1;
for(i = 0; i < overlap/2; i++)
{
kiss_fft_scalar x1, x2;
x1 = *xp1;
x2 = *yp1;
*yp1++ = MULT16_32_Q15(*wp2, x2) - MULT16_32_Q15(*wp1, x1);
*xp1-- = MULT16_32_Q15(*wp1, x2) + MULT16_32_Q15(*wp2, x1);
wp1++;
wp2--;
}
}
RESTORE_STACK;
}

252
Frameworks/Opus/Opus/opus/celt/arm/celt_neon_intr.c Normal file
View file

@ -0,0 +1,252 @@
/* Copyright (c) 2014-2015 Xiph.Org Foundation
Written by Viswanath Puttagunta */
/**
@file celt_neon_intr.c
@brief ARM Neon Intrinsic optimizations for celt
*/
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <arm_neon.h>
#include "../pitch.h"
#if !defined(FIXED_POINT)
/*
* Function: xcorr_kernel_neon_float
* ---------------------------------
* Computes 4 correlation values and stores them in sum[4]
*/
static void xcorr_kernel_neon_float(const float32_t *x, const float32_t *y,
float32_t sum[4], int len) {
float32x4_t YY[3];
float32x4_t YEXT[3];
float32x4_t XX[2];
float32x2_t XX_2;
float32x4_t SUMM;
const float32_t *xi = x;
const float32_t *yi = y;
celt_assert(len>0);
YY[0] = vld1q_f32(yi);
SUMM = vdupq_n_f32(0);
/* Consume 8 elements in x vector and 12 elements in y
* vector. However, the 12'th element never really gets
* touched in this loop. So, if len == 8, then we only
* must access y[0] to y[10]. y[11] must not be accessed
* hence make sure len > 8 and not len >= 8
*/
while (len > 8) {
yi += 4;
YY[1] = vld1q_f32(yi);
yi += 4;
YY[2] = vld1q_f32(yi);
XX[0] = vld1q_f32(xi);
xi += 4;
XX[1] = vld1q_f32(xi);
xi += 4;
SUMM = vmlaq_lane_f32(SUMM, YY[0], vget_low_f32(XX[0]), 0);
YEXT[0] = vextq_f32(YY[0], YY[1], 1);
SUMM = vmlaq_lane_f32(SUMM, YEXT[0], vget_low_f32(XX[0]), 1);
YEXT[1] = vextq_f32(YY[0], YY[1], 2);
SUMM = vmlaq_lane_f32(SUMM, YEXT[1], vget_high_f32(XX[0]), 0);
YEXT[2] = vextq_f32(YY[0], YY[1], 3);
SUMM = vmlaq_lane_f32(SUMM, YEXT[2], vget_high_f32(XX[0]), 1);
SUMM = vmlaq_lane_f32(SUMM, YY[1], vget_low_f32(XX[1]), 0);
YEXT[0] = vextq_f32(YY[1], YY[2], 1);
SUMM = vmlaq_lane_f32(SUMM, YEXT[0], vget_low_f32(XX[1]), 1);
YEXT[1] = vextq_f32(YY[1], YY[2], 2);
SUMM = vmlaq_lane_f32(SUMM, YEXT[1], vget_high_f32(XX[1]), 0);
YEXT[2] = vextq_f32(YY[1], YY[2], 3);
SUMM = vmlaq_lane_f32(SUMM, YEXT[2], vget_high_f32(XX[1]), 1);
YY[0] = YY[2];
len -= 8;
}
/* Consume 4 elements in x vector and 8 elements in y
* vector. However, the 8'th element in y never really gets
* touched in this loop. So, if len == 4, then we only
* must access y[0] to y[6]. y[7] must not be accessed
* hence make sure len>4 and not len>=4
*/
if (len > 4) {
yi += 4;
YY[1] = vld1q_f32(yi);
XX[0] = vld1q_f32(xi);
xi += 4;
SUMM = vmlaq_lane_f32(SUMM, YY[0], vget_low_f32(XX[0]), 0);
YEXT[0] = vextq_f32(YY[0], YY[1], 1);
SUMM = vmlaq_lane_f32(SUMM, YEXT[0], vget_low_f32(XX[0]), 1);
YEXT[1] = vextq_f32(YY[0], YY[1], 2);
SUMM = vmlaq_lane_f32(SUMM, YEXT[1], vget_high_f32(XX[0]), 0);
YEXT[2] = vextq_f32(YY[0], YY[1], 3);
SUMM = vmlaq_lane_f32(SUMM, YEXT[2], vget_high_f32(XX[0]), 1);
YY[0] = YY[1];
len -= 4;
}
while (--len > 0) {
XX_2 = vld1_dup_f32(xi++);
SUMM = vmlaq_lane_f32(SUMM, YY[0], XX_2, 0);
YY[0]= vld1q_f32(++yi);
}
XX_2 = vld1_dup_f32(xi);
SUMM = vmlaq_lane_f32(SUMM, YY[0], XX_2, 0);
vst1q_f32(sum, SUMM);
}
/*
* Function: xcorr_kernel_neon_float_process1
* ---------------------------------
* Computes single correlation values and stores in *sum
*/
static void xcorr_kernel_neon_float_process1(const float32_t *x,
const float32_t *y, float32_t *sum, int len) {
float32x4_t XX[4];
float32x4_t YY[4];
float32x2_t XX_2;
float32x2_t YY_2;
float32x4_t SUMM;
float32x2_t SUMM_2[2];
const float32_t *xi = x;
const float32_t *yi = y;
SUMM = vdupq_n_f32(0);
/* Work on 16 values per iteration */
while (len >= 16) {
XX[0] = vld1q_f32(xi);
xi += 4;
XX[1] = vld1q_f32(xi);
xi += 4;
XX[2] = vld1q_f32(xi);
xi += 4;
XX[3] = vld1q_f32(xi);
xi += 4;
YY[0] = vld1q_f32(yi);
yi += 4;
YY[1] = vld1q_f32(yi);
yi += 4;
YY[2] = vld1q_f32(yi);
yi += 4;
YY[3] = vld1q_f32(yi);
yi += 4;
SUMM = vmlaq_f32(SUMM, YY[0], XX[0]);
SUMM = vmlaq_f32(SUMM, YY[1], XX[1]);
SUMM = vmlaq_f32(SUMM, YY[2], XX[2]);
SUMM = vmlaq_f32(SUMM, YY[3], XX[3]);
len -= 16;
}
/* Work on 8 values */
if (len >= 8) {
XX[0] = vld1q_f32(xi);
xi += 4;
XX[1] = vld1q_f32(xi);
xi += 4;
YY[0] = vld1q_f32(yi);
yi += 4;
YY[1] = vld1q_f32(yi);
yi += 4;
SUMM = vmlaq_f32(SUMM, YY[0], XX[0]);
SUMM = vmlaq_f32(SUMM, YY[1], XX[1]);
len -= 8;
}
/* Work on 4 values */
if (len >= 4) {
XX[0] = vld1q_f32(xi);
xi += 4;
YY[0] = vld1q_f32(yi);
yi += 4;
SUMM = vmlaq_f32(SUMM, YY[0], XX[0]);
len -= 4;
}
/* Start accumulating results */
SUMM_2[0] = vget_low_f32(SUMM);
if (len >= 2) {
/* While at it, consume 2 more values if available */
XX_2 = vld1_f32(xi);
xi += 2;
YY_2 = vld1_f32(yi);
yi += 2;
SUMM_2[0] = vmla_f32(SUMM_2[0], YY_2, XX_2);
len -= 2;
}
SUMM_2[1] = vget_high_f32(SUMM);
SUMM_2[0] = vadd_f32(SUMM_2[0], SUMM_2[1]);
SUMM_2[0] = vpadd_f32(SUMM_2[0], SUMM_2[0]);
/* Ok, now we have result accumulated in SUMM_2[0].0 */
if (len > 0) {
/* Case when you have one value left */
XX_2 = vld1_dup_f32(xi);
YY_2 = vld1_dup_f32(yi);
SUMM_2[0] = vmla_f32(SUMM_2[0], XX_2, YY_2);
}
vst1_lane_f32(sum, SUMM_2[0], 0);
}
void celt_pitch_xcorr_float_neon(const opus_val16 *_x, const opus_val16 *_y,
opus_val32 *xcorr, int len, int max_pitch) {
int i;
celt_assert(max_pitch > 0);
celt_assert((((unsigned char *)_x-(unsigned char *)NULL)&3)==0);
for (i = 0; i < (max_pitch-3); i += 4) {
xcorr_kernel_neon_float((const float32_t *)_x, (const float32_t *)_y+i,
(float32_t *)xcorr+i, len);
}
/* In case max_pitch isn't multiple of 4
* compute single correlation value per iteration
*/
for (; i < max_pitch; i++) {
xcorr_kernel_neon_float_process1((const float32_t *)_x,
(const float32_t *)_y+i, (float32_t *)xcorr+i, len);
}
}
#endif

551
Frameworks/Opus/Opus/opus/celt/arm/celt_pitch_xcorr_arm-gnu.S Normal file
View file

@ -0,0 +1,551 @@
.syntax unified
@ Copyright (c) 2007-2008 CSIRO
@ Copyright (c) 2007-2009 Xiph.Org Foundation
@ Copyright (c) 2013 Parrot
@ Written by Aurélien Zanelli
@
@ Redistribution and use in source and binary forms, with or without
@ modification, are permitted provided that the following conditions
@ are met:
@
@ - Redistributions of source code must retain the above copyright
@ notice, this list of conditions and the following disclaimer.
@
@ - Redistributions in binary form must reproduce the above copyright
@ notice, this list of conditions and the following disclaimer in the
@ documentation and/or other materials provided with the distribution.
@
@ THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
@ ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
@ LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
@ A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
@ OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
@ EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
@ PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
@ PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
@ LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
@ NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
@ SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
.text; .p2align 2; .arch armv7-a
.fpu neon
.object_arch armv4t
.include "celt/arm/armopts-gnu.S"
.if OPUS_ARM_MAY_HAVE_EDSP
.global celt_pitch_xcorr_edsp
.endif
.if OPUS_ARM_MAY_HAVE_NEON
.global celt_pitch_xcorr_neon
.endif
.if OPUS_ARM_MAY_HAVE_NEON
@ Compute sum[k]=sum(x[j]*y[j+k],j=0...len-1), k=0...3
; xcorr_kernel_neon: @ PROC
xcorr_kernel_neon_start:
@ input:
@ r3 = int len
@ r4 = opus_val16 *x
@ r5 = opus_val16 *y
@ q0 = opus_val32 sum[4]
@ output:
@ q0 = opus_val32 sum[4]
@ preserved: r0-r3, r6-r11, d2, q4-q7, q9-q15
@ internal usage:
@ r12 = int j
@ d3 = y_3|y_2|y_1|y_0
@ q2 = y_B|y_A|y_9|y_8|y_7|y_6|y_5|y_4
@ q3 = x_7|x_6|x_5|x_4|x_3|x_2|x_1|x_0
@ q8 = scratch
@
@ Load y[0...3]
@ This requires len>0 to always be valid (which we assert in the C code).
VLD1.16 {d5}, [r5]!
SUBS r12, r3, #8
BLE xcorr_kernel_neon_process4
@ Process 8 samples at a time.
@ This loop loads one y value more than we actually need. Therefore we have to
@ stop as soon as there are 8 or fewer samples left (instead of 7), to avoid
@ reading past the end of the array.
xcorr_kernel_neon_process8:
@ This loop has 19 total instructions (10 cycles to issue, minimum), with
@ - 2 cycles of ARM insrtuctions,
@ - 10 cycles of load/store/byte permute instructions, and
@ - 9 cycles of data processing instructions.
@ On a Cortex A8, we dual-issue the maximum amount (9 cycles) between the
@ latter two categories, meaning the whole loop should run in 10 cycles per
@ iteration, barring cache misses.
@
@ Load x[0...7]
VLD1.16 {d6, d7}, [r4]!
@ Unlike VMOV, VAND is a data processsing instruction (and doesn't get
@ assembled to VMOV, like VORR would), so it dual-issues with the prior VLD1.
VAND d3, d5, d5
SUBS r12, r12, #8
@ Load y[4...11]
VLD1.16 {d4, d5}, [r5]!
VMLAL.S16 q0, d3, d6[0]
VEXT.16 d16, d3, d4, #1
VMLAL.S16 q0, d4, d7[0]
VEXT.16 d17, d4, d5, #1
VMLAL.S16 q0, d16, d6[1]
VEXT.16 d16, d3, d4, #2
VMLAL.S16 q0, d17, d7[1]
VEXT.16 d17, d4, d5, #2
VMLAL.S16 q0, d16, d6[2]
VEXT.16 d16, d3, d4, #3
VMLAL.S16 q0, d17, d7[2]
VEXT.16 d17, d4, d5, #3
VMLAL.S16 q0, d16, d6[3]
VMLAL.S16 q0, d17, d7[3]
BGT xcorr_kernel_neon_process8
@ Process 4 samples here if we have > 4 left (still reading one extra y value).
xcorr_kernel_neon_process4:
ADDS r12, r12, #4
BLE xcorr_kernel_neon_process2
@ Load x[0...3]
VLD1.16 d6, [r4]!
@ Use VAND since it's a data processing instruction again.
VAND d4, d5, d5
SUB r12, r12, #4
@ Load y[4...7]
VLD1.16 d5, [r5]!
VMLAL.S16 q0, d4, d6[0]
VEXT.16 d16, d4, d5, #1
VMLAL.S16 q0, d16, d6[1]
VEXT.16 d16, d4, d5, #2
VMLAL.S16 q0, d16, d6[2]
VEXT.16 d16, d4, d5, #3
VMLAL.S16 q0, d16, d6[3]
@ Process 2 samples here if we have > 2 left (still reading one extra y value).
xcorr_kernel_neon_process2:
ADDS r12, r12, #2
BLE xcorr_kernel_neon_process1
@ Load x[0...1]
VLD2.16 {d6[],d7[]}, [r4]!
@ Use VAND since it's a data processing instruction again.
VAND d4, d5, d5
SUB r12, r12, #2
@ Load y[4...5]
VLD1.32 {d5[]}, [r5]!
VMLAL.S16 q0, d4, d6
VEXT.16 d16, d4, d5, #1
@ Replace bottom copy of {y5,y4} in d5 with {y3,y2} from d4, using VSRI
@ instead of VEXT, since it's a data-processing instruction.
VSRI.64 d5, d4, #32
VMLAL.S16 q0, d16, d7
@ Process 1 sample using the extra y value we loaded above.
xcorr_kernel_neon_process1:
@ Load next *x
VLD1.16 {d6[]}, [r4]!
ADDS r12, r12, #1
@ y[0...3] are left in d5 from prior iteration(s) (if any)
VMLAL.S16 q0, d5, d6
MOVLE pc, lr
@ Now process 1 last sample, not reading ahead.
@ Load last *y
VLD1.16 {d4[]}, [r5]!
VSRI.64 d4, d5, #16
@ Load last *x
VLD1.16 {d6[]}, [r4]!
VMLAL.S16 q0, d4, d6
MOV pc, lr
.size xcorr_kernel_neon, .-xcorr_kernel_neon @ ENDP
@ opus_val32 celt_pitch_xcorr_neon(opus_val16 *_x, opus_val16 *_y,
@ opus_val32 *xcorr, int len, int max_pitch)
; celt_pitch_xcorr_neon: @ PROC
@ input:
@ r0 = opus_val16 *_x
@ r1 = opus_val16 *_y
@ r2 = opus_val32 *xcorr
@ r3 = int len
@ output:
@ r0 = int maxcorr
@ internal usage:
@ r4 = opus_val16 *x (for xcorr_kernel_neon())
@ r5 = opus_val16 *y (for xcorr_kernel_neon())
@ r6 = int max_pitch
@ r12 = int j
@ q15 = int maxcorr[4] (q15 is not used by xcorr_kernel_neon())
STMFD sp!, {r4-r6, lr}
LDR r6, [sp, #16]
VMOV.S32 q15, #1
@ if (max_pitch < 4) goto celt_pitch_xcorr_neon_process4_done
SUBS r6, r6, #4
BLT celt_pitch_xcorr_neon_process4_done
celt_pitch_xcorr_neon_process4:
@ xcorr_kernel_neon parameters:
@ r3 = len, r4 = _x, r5 = _y, q0 = {0, 0, 0, 0}
MOV r4, r0
MOV r5, r1
VEOR q0, q0, q0
@ xcorr_kernel_neon only modifies r4, r5, r12, and q0...q3.
@ So we don't save/restore any other registers.
BL xcorr_kernel_neon_start
SUBS r6, r6, #4
VST1.32 {q0}, [r2]!
@ _y += 4
ADD r1, r1, #8
VMAX.S32 q15, q15, q0
@ if (max_pitch < 4) goto celt_pitch_xcorr_neon_process4_done
BGE celt_pitch_xcorr_neon_process4
@ We have less than 4 sums left to compute.
celt_pitch_xcorr_neon_process4_done:
ADDS r6, r6, #4
@ Reduce maxcorr to a single value
VMAX.S32 d30, d30, d31
VPMAX.S32 d30, d30, d30
@ if (max_pitch <= 0) goto celt_pitch_xcorr_neon_done
BLE celt_pitch_xcorr_neon_done
@ Now compute each remaining sum one at a time.
celt_pitch_xcorr_neon_process_remaining:
MOV r4, r0
MOV r5, r1
VMOV.I32 q0, #0
SUBS r12, r3, #8
BLT celt_pitch_xcorr_neon_process_remaining4
@ Sum terms 8 at a time.
celt_pitch_xcorr_neon_process_remaining_loop8:
@ Load x[0...7]
VLD1.16 {q1}, [r4]!
@ Load y[0...7]
VLD1.16 {q2}, [r5]!
SUBS r12, r12, #8
VMLAL.S16 q0, d4, d2
VMLAL.S16 q0, d5, d3
BGE celt_pitch_xcorr_neon_process_remaining_loop8
@ Sum terms 4 at a time.
celt_pitch_xcorr_neon_process_remaining4:
ADDS r12, r12, #4
BLT celt_pitch_xcorr_neon_process_remaining4_done
@ Load x[0...3]
VLD1.16 {d2}, [r4]!
@ Load y[0...3]
VLD1.16 {d3}, [r5]!
SUB r12, r12, #4
VMLAL.S16 q0, d3, d2
celt_pitch_xcorr_neon_process_remaining4_done:
@ Reduce the sum to a single value.
VADD.S32 d0, d0, d1
VPADDL.S32 d0, d0
ADDS r12, r12, #4
BLE celt_pitch_xcorr_neon_process_remaining_loop_done
@ Sum terms 1 at a time.
celt_pitch_xcorr_neon_process_remaining_loop1:
VLD1.16 {d2[]}, [r4]!
VLD1.16 {d3[]}, [r5]!
SUBS r12, r12, #1
VMLAL.S16 q0, d2, d3
BGT celt_pitch_xcorr_neon_process_remaining_loop1
celt_pitch_xcorr_neon_process_remaining_loop_done:
VST1.32 {d0[0]}, [r2]!
VMAX.S32 d30, d30, d0
SUBS r6, r6, #1
@ _y++
ADD r1, r1, #2
@ if (--max_pitch > 0) goto celt_pitch_xcorr_neon_process_remaining
BGT celt_pitch_xcorr_neon_process_remaining
celt_pitch_xcorr_neon_done:
VMOV.32 r0, d30[0]
LDMFD sp!, {r4-r6, pc}
.size celt_pitch_xcorr_neon, .-celt_pitch_xcorr_neon @ ENDP
.endif
.if OPUS_ARM_MAY_HAVE_EDSP
@ This will get used on ARMv7 devices without NEON, so it has been optimized
@ to take advantage of dual-issuing where possible.
; xcorr_kernel_edsp: @ PROC
xcorr_kernel_edsp_start:
@ input:
@ r3 = int len
@ r4 = opus_val16 *_x (must be 32-bit aligned)
@ r5 = opus_val16 *_y (must be 32-bit aligned)
@ r6...r9 = opus_val32 sum[4]
@ output:
@ r6...r9 = opus_val32 sum[4]
@ preserved: r0-r5
@ internal usage
@ r2 = int j
@ r12,r14 = opus_val16 x[4]
@ r10,r11 = opus_val16 y[4]
STMFD sp!, {r2,r4,r5,lr}
LDR r10, [r5], #4 @ Load y[0...1]
SUBS r2, r3, #4 @ j = len-4
LDR r11, [r5], #4 @ Load y[2...3]
BLE xcorr_kernel_edsp_process4_done
LDR r12, [r4], #4 @ Load x[0...1]
@ Stall
xcorr_kernel_edsp_process4:
@ The multiplies must issue from pipeline 0, and can't dual-issue with each
@ other. Every other instruction here dual-issues with a multiply, and is
@ thus "free". There should be no stalls in the body of the loop.
SMLABB r6, r12, r10, r6 @ sum[0] = MAC16_16(sum[0],x_0,y_0)
LDR r14, [r4], #4 @ Load x[2...3]
SMLABT r7, r12, r10, r7 @ sum[1] = MAC16_16(sum[1],x_0,y_1)
SUBS r2, r2, #4 @ j-=4
SMLABB r8, r12, r11, r8 @ sum[2] = MAC16_16(sum[2],x_0,y_2)
SMLABT r9, r12, r11, r9 @ sum[3] = MAC16_16(sum[3],x_0,y_3)
SMLATT r6, r12, r10, r6 @ sum[0] = MAC16_16(sum[0],x_1,y_1)
LDR r10, [r5], #4 @ Load y[4...5]
SMLATB r7, r12, r11, r7 @ sum[1] = MAC16_16(sum[1],x_1,y_2)
SMLATT r8, r12, r11, r8 @ sum[2] = MAC16_16(sum[2],x_1,y_3)
SMLATB r9, r12, r10, r9 @ sum[3] = MAC16_16(sum[3],x_1,y_4)
LDRGT r12, [r4], #4 @ Load x[0...1]
SMLABB r6, r14, r11, r6 @ sum[0] = MAC16_16(sum[0],x_2,y_2)
SMLABT r7, r14, r11, r7 @ sum[1] = MAC16_16(sum[1],x_2,y_3)
SMLABB r8, r14, r10, r8 @ sum[2] = MAC16_16(sum[2],x_2,y_4)
SMLABT r9, r14, r10, r9 @ sum[3] = MAC16_16(sum[3],x_2,y_5)
SMLATT r6, r14, r11, r6 @ sum[0] = MAC16_16(sum[0],x_3,y_3)
LDR r11, [r5], #4 @ Load y[6...7]
SMLATB r7, r14, r10, r7 @ sum[1] = MAC16_16(sum[1],x_3,y_4)
SMLATT r8, r14, r10, r8 @ sum[2] = MAC16_16(sum[2],x_3,y_5)
SMLATB r9, r14, r11, r9 @ sum[3] = MAC16_16(sum[3],x_3,y_6)
BGT xcorr_kernel_edsp_process4
xcorr_kernel_edsp_process4_done:
ADDS r2, r2, #4
BLE xcorr_kernel_edsp_done
LDRH r12, [r4], #2 @ r12 = *x++
SUBS r2, r2, #1 @ j--
@ Stall
SMLABB r6, r12, r10, r6 @ sum[0] = MAC16_16(sum[0],x,y_0)
LDRHGT r14, [r4], #2 @ r14 = *x++
SMLABT r7, r12, r10, r7 @ sum[1] = MAC16_16(sum[1],x,y_1)
SMLABB r8, r12, r11, r8 @ sum[2] = MAC16_16(sum[2],x,y_2)
SMLABT r9, r12, r11, r9 @ sum[3] = MAC16_16(sum[3],x,y_3)
BLE xcorr_kernel_edsp_done
SMLABT r6, r14, r10, r6 @ sum[0] = MAC16_16(sum[0],x,y_1)
SUBS r2, r2, #1 @ j--
SMLABB r7, r14, r11, r7 @ sum[1] = MAC16_16(sum[1],x,y_2)
LDRH r10, [r5], #2 @ r10 = y_4 = *y++
SMLABT r8, r14, r11, r8 @ sum[2] = MAC16_16(sum[2],x,y_3)
LDRHGT r12, [r4], #2 @ r12 = *x++
SMLABB r9, r14, r10, r9 @ sum[3] = MAC16_16(sum[3],x,y_4)
BLE xcorr_kernel_edsp_done
SMLABB r6, r12, r11, r6 @ sum[0] = MAC16_16(sum[0],tmp,y_2)
CMP r2, #1 @ j--
SMLABT r7, r12, r11, r7 @ sum[1] = MAC16_16(sum[1],tmp,y_3)
LDRH r2, [r5], #2 @ r2 = y_5 = *y++
SMLABB r8, r12, r10, r8 @ sum[2] = MAC16_16(sum[2],tmp,y_4)
LDRHGT r14, [r4] @ r14 = *x
SMLABB r9, r12, r2, r9 @ sum[3] = MAC16_16(sum[3],tmp,y_5)
BLE xcorr_kernel_edsp_done
SMLABT r6, r14, r11, r6 @ sum[0] = MAC16_16(sum[0],tmp,y_3)
LDRH r11, [r5] @ r11 = y_6 = *y
SMLABB r7, r14, r10, r7 @ sum[1] = MAC16_16(sum[1],tmp,y_4)
SMLABB r8, r14, r2, r8 @ sum[2] = MAC16_16(sum[2],tmp,y_5)
SMLABB r9, r14, r11, r9 @ sum[3] = MAC16_16(sum[3],tmp,y_6)
xcorr_kernel_edsp_done:
LDMFD sp!, {r2,r4,r5,pc}
.size xcorr_kernel_edsp, .-xcorr_kernel_edsp @ ENDP
; celt_pitch_xcorr_edsp: @ PROC
@ input:
@ r0 = opus_val16 *_x (must be 32-bit aligned)
@ r1 = opus_val16 *_y (only needs to be 16-bit aligned)
@ r2 = opus_val32 *xcorr
@ r3 = int len
@ output:
@ r0 = maxcorr
@ internal usage
@ r4 = opus_val16 *x
@ r5 = opus_val16 *y
@ r6 = opus_val32 sum0
@ r7 = opus_val32 sum1
@ r8 = opus_val32 sum2
@ r9 = opus_val32 sum3
@ r1 = int max_pitch
@ r12 = int j
STMFD sp!, {r4-r11, lr}
MOV r5, r1
LDR r1, [sp, #36]
MOV r4, r0
TST r5, #3
@ maxcorr = 1
MOV r0, #1
BEQ celt_pitch_xcorr_edsp_process1u_done
@ Compute one sum at the start to make y 32-bit aligned.
SUBS r12, r3, #4
@ r14 = sum = 0
MOV r14, #0
LDRH r8, [r5], #2
BLE celt_pitch_xcorr_edsp_process1u_loop4_done
LDR r6, [r4], #4
MOV r8, r8, LSL #16
celt_pitch_xcorr_edsp_process1u_loop4:
LDR r9, [r5], #4
SMLABT r14, r6, r8, r14 @ sum = MAC16_16(sum, x_0, y_0)
LDR r7, [r4], #4
SMLATB r14, r6, r9, r14 @ sum = MAC16_16(sum, x_1, y_1)
LDR r8, [r5], #4
SMLABT r14, r7, r9, r14 @ sum = MAC16_16(sum, x_2, y_2)
SUBS r12, r12, #4 @ j-=4
SMLATB r14, r7, r8, r14 @ sum = MAC16_16(sum, x_3, y_3)
LDRGT r6, [r4], #4
BGT celt_pitch_xcorr_edsp_process1u_loop4
MOV r8, r8, LSR #16
celt_pitch_xcorr_edsp_process1u_loop4_done:
ADDS r12, r12, #4
celt_pitch_xcorr_edsp_process1u_loop1:
LDRHGE r6, [r4], #2
@ Stall
SMLABBGE r14, r6, r8, r14 @ sum = MAC16_16(sum, *x, *y)
SUBSGE r12, r12, #1
LDRHGT r8, [r5], #2
BGT celt_pitch_xcorr_edsp_process1u_loop1
@ Restore _x
SUB r4, r4, r3, LSL #1
@ Restore and advance _y
SUB r5, r5, r3, LSL #1
@ maxcorr = max(maxcorr, sum)
CMP r0, r14
ADD r5, r5, #2
MOVLT r0, r14
SUBS r1, r1, #1
@ xcorr[i] = sum
STR r14, [r2], #4
BLE celt_pitch_xcorr_edsp_done
celt_pitch_xcorr_edsp_process1u_done:
@ if (max_pitch < 4) goto celt_pitch_xcorr_edsp_process2
SUBS r1, r1, #4
BLT celt_pitch_xcorr_edsp_process2
celt_pitch_xcorr_edsp_process4:
@ xcorr_kernel_edsp parameters:
@ r3 = len, r4 = _x, r5 = _y, r6...r9 = sum[4] = {0, 0, 0, 0}
MOV r6, #0
MOV r7, #0
MOV r8, #0
MOV r9, #0
BL xcorr_kernel_edsp_start @ xcorr_kernel_edsp(_x, _y+i, xcorr+i, len)
@ maxcorr = max(maxcorr, sum0, sum1, sum2, sum3)
CMP r0, r6
@ _y+=4
ADD r5, r5, #8
MOVLT r0, r6
CMP r0, r7
MOVLT r0, r7
CMP r0, r8
MOVLT r0, r8
CMP r0, r9
MOVLT r0, r9
STMIA r2!, {r6-r9}
SUBS r1, r1, #4
BGE celt_pitch_xcorr_edsp_process4
celt_pitch_xcorr_edsp_process2:
ADDS r1, r1, #2
BLT celt_pitch_xcorr_edsp_process1a
SUBS r12, r3, #4
@ {r10, r11} = {sum0, sum1} = {0, 0}
MOV r10, #0
MOV r11, #0
LDR r8, [r5], #4
BLE celt_pitch_xcorr_edsp_process2_loop_done
LDR r6, [r4], #4
LDR r9, [r5], #4
celt_pitch_xcorr_edsp_process2_loop4:
SMLABB r10, r6, r8, r10 @ sum0 = MAC16_16(sum0, x_0, y_0)
LDR r7, [r4], #4
SMLABT r11, r6, r8, r11 @ sum1 = MAC16_16(sum1, x_0, y_1)
SUBS r12, r12, #4 @ j-=4
SMLATT r10, r6, r8, r10 @ sum0 = MAC16_16(sum0, x_1, y_1)
LDR r8, [r5], #4
SMLATB r11, r6, r9, r11 @ sum1 = MAC16_16(sum1, x_1, y_2)
LDRGT r6, [r4], #4
SMLABB r10, r7, r9, r10 @ sum0 = MAC16_16(sum0, x_2, y_2)
SMLABT r11, r7, r9, r11 @ sum1 = MAC16_16(sum1, x_2, y_3)
SMLATT r10, r7, r9, r10 @ sum0 = MAC16_16(sum0, x_3, y_3)
LDRGT r9, [r5], #4
SMLATB r11, r7, r8, r11 @ sum1 = MAC16_16(sum1, x_3, y_4)
BGT celt_pitch_xcorr_edsp_process2_loop4
celt_pitch_xcorr_edsp_process2_loop_done:
ADDS r12, r12, #2
BLE celt_pitch_xcorr_edsp_process2_1
LDR r6, [r4], #4
@ Stall
SMLABB r10, r6, r8, r10 @ sum0 = MAC16_16(sum0, x_0, y_0)
LDR r9, [r5], #4
SMLABT r11, r6, r8, r11 @ sum1 = MAC16_16(sum1, x_0, y_1)
SUB r12, r12, #2
SMLATT r10, r6, r8, r10 @ sum0 = MAC16_16(sum0, x_1, y_1)
MOV r8, r9
SMLATB r11, r6, r9, r11 @ sum1 = MAC16_16(sum1, x_1, y_2)
celt_pitch_xcorr_edsp_process2_1:
LDRH r6, [r4], #2
ADDS r12, r12, #1
@ Stall
SMLABB r10, r6, r8, r10 @ sum0 = MAC16_16(sum0, x_0, y_0)
LDRHGT r7, [r4], #2
SMLABT r11, r6, r8, r11 @ sum1 = MAC16_16(sum1, x_0, y_1)
BLE celt_pitch_xcorr_edsp_process2_done
LDRH r9, [r5], #2
SMLABT r10, r7, r8, r10 @ sum0 = MAC16_16(sum0, x_0, y_1)
SMLABB r11, r7, r9, r11 @ sum1 = MAC16_16(sum1, x_0, y_2)
celt_pitch_xcorr_edsp_process2_done:
@ Restore _x
SUB r4, r4, r3, LSL #1
@ Restore and advance _y
SUB r5, r5, r3, LSL #1
@ maxcorr = max(maxcorr, sum0)
CMP r0, r10
ADD r5, r5, #2
MOVLT r0, r10
SUB r1, r1, #2
@ maxcorr = max(maxcorr, sum1)
CMP r0, r11
@ xcorr[i] = sum
STR r10, [r2], #4
MOVLT r0, r11
STR r11, [r2], #4
celt_pitch_xcorr_edsp_process1a:
ADDS r1, r1, #1
BLT celt_pitch_xcorr_edsp_done
SUBS r12, r3, #4
@ r14 = sum = 0
MOV r14, #0
BLT celt_pitch_xcorr_edsp_process1a_loop_done
LDR r6, [r4], #4
LDR r8, [r5], #4
LDR r7, [r4], #4
LDR r9, [r5], #4
celt_pitch_xcorr_edsp_process1a_loop4:
SMLABB r14, r6, r8, r14 @ sum = MAC16_16(sum, x_0, y_0)
SUBS r12, r12, #4 @ j-=4
SMLATT r14, r6, r8, r14 @ sum = MAC16_16(sum, x_1, y_1)
LDRGE r6, [r4], #4
SMLABB r14, r7, r9, r14 @ sum = MAC16_16(sum, x_2, y_2)
LDRGE r8, [r5], #4
SMLATT r14, r7, r9, r14 @ sum = MAC16_16(sum, x_3, y_3)
LDRGE r7, [r4], #4
LDRGE r9, [r5], #4
BGE celt_pitch_xcorr_edsp_process1a_loop4
celt_pitch_xcorr_edsp_process1a_loop_done:
ADDS r12, r12, #2
LDRGE r6, [r4], #4
LDRGE r8, [r5], #4
@ Stall
SMLABBGE r14, r6, r8, r14 @ sum = MAC16_16(sum, x_0, y_0)
SUBGE r12, r12, #2
SMLATTGE r14, r6, r8, r14 @ sum = MAC16_16(sum, x_1, y_1)
ADDS r12, r12, #1
LDRHGE r6, [r4], #2
LDRHGE r8, [r5], #2
@ Stall
SMLABBGE r14, r6, r8, r14 @ sum = MAC16_16(sum, *x, *y)
@ maxcorr = max(maxcorr, sum)
CMP r0, r14
@ xcorr[i] = sum
STR r14, [r2], #4
MOVLT r0, r14
celt_pitch_xcorr_edsp_done:
LDMFD sp!, {r4-r11, pc}
.size celt_pitch_xcorr_edsp, .-celt_pitch_xcorr_edsp @ ENDP
.endif
@ END:
.section .note.GNU-stack,"",%progbits

547
Frameworks/Opus/Opus/opus/celt/arm/celt_pitch_xcorr_arm.s Normal file
View file

@ -0,0 +1,547 @@
; Copyright (c) 2007-2008 CSIRO
; Copyright (c) 2007-2009 Xiph.Org Foundation
; Copyright (c) 2013 Parrot
; Written by Aurélien Zanelli
;
; Redistribution and use in source and binary forms, with or without
; modification, are permitted provided that the following conditions
; are met:
;
; - Redistributions of source code must retain the above copyright
; notice, this list of conditions and the following disclaimer.
;
; - Redistributions in binary form must reproduce the above copyright
; notice, this list of conditions and the following disclaimer in the
; documentation and/or other materials provided with the distribution.
;
; THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
; ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
; LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
; A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
; OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
; EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
; PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
; PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
; LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
; NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
; SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
AREA |.text|, CODE, READONLY
GET celt/arm/armopts.s
IF OPUS_ARM_MAY_HAVE_EDSP
EXPORT celt_pitch_xcorr_edsp
ENDIF
IF OPUS_ARM_MAY_HAVE_NEON
EXPORT celt_pitch_xcorr_neon
ENDIF
IF OPUS_ARM_MAY_HAVE_NEON
; Compute sum[k]=sum(x[j]*y[j+k],j=0...len-1), k=0...3
xcorr_kernel_neon PROC
xcorr_kernel_neon_start
; input:
; r3 = int len
; r4 = opus_val16 *x
; r5 = opus_val16 *y
; q0 = opus_val32 sum[4]
; output:
; q0 = opus_val32 sum[4]
; preserved: r0-r3, r6-r11, d2, q4-q7, q9-q15
; internal usage:
; r12 = int j
; d3 = y_3|y_2|y_1|y_0
; q2 = y_B|y_A|y_9|y_8|y_7|y_6|y_5|y_4
; q3 = x_7|x_6|x_5|x_4|x_3|x_2|x_1|x_0
; q8 = scratch
;
; Load y[0...3]
; This requires len>0 to always be valid (which we assert in the C code).
VLD1.16 {d5}, [r5]!
SUBS r12, r3, #8
BLE xcorr_kernel_neon_process4
; Process 8 samples at a time.
; This loop loads one y value more than we actually need. Therefore we have to
; stop as soon as there are 8 or fewer samples left (instead of 7), to avoid
; reading past the end of the array.
xcorr_kernel_neon_process8
; This loop has 19 total instructions (10 cycles to issue, minimum), with
; - 2 cycles of ARM insrtuctions,
; - 10 cycles of load/store/byte permute instructions, and
; - 9 cycles of data processing instructions.
; On a Cortex A8, we dual-issue the maximum amount (9 cycles) between the
; latter two categories, meaning the whole loop should run in 10 cycles per
; iteration, barring cache misses.
;
; Load x[0...7]
VLD1.16 {d6, d7}, [r4]!
; Unlike VMOV, VAND is a data processsing instruction (and doesn't get
; assembled to VMOV, like VORR would), so it dual-issues with the prior VLD1.
VAND d3, d5, d5
SUBS r12, r12, #8
; Load y[4...11]
VLD1.16 {d4, d5}, [r5]!
VMLAL.S16 q0, d3, d6[0]
VEXT.16 d16, d3, d4, #1
VMLAL.S16 q0, d4, d7[0]
VEXT.16 d17, d4, d5, #1
VMLAL.S16 q0, d16, d6[1]
VEXT.16 d16, d3, d4, #2
VMLAL.S16 q0, d17, d7[1]
VEXT.16 d17, d4, d5, #2
VMLAL.S16 q0, d16, d6[2]
VEXT.16 d16, d3, d4, #3
VMLAL.S16 q0, d17, d7[2]
VEXT.16 d17, d4, d5, #3
VMLAL.S16 q0, d16, d6[3]
VMLAL.S16 q0, d17, d7[3]
BGT xcorr_kernel_neon_process8
; Process 4 samples here if we have > 4 left (still reading one extra y value).
xcorr_kernel_neon_process4
ADDS r12, r12, #4
BLE xcorr_kernel_neon_process2
; Load x[0...3]
VLD1.16 d6, [r4]!
; Use VAND since it's a data processing instruction again.
VAND d4, d5, d5
SUB r12, r12, #4
; Load y[4...7]
VLD1.16 d5, [r5]!
VMLAL.S16 q0, d4, d6[0]
VEXT.16 d16, d4, d5, #1
VMLAL.S16 q0, d16, d6[1]
VEXT.16 d16, d4, d5, #2
VMLAL.S16 q0, d16, d6[2]
VEXT.16 d16, d4, d5, #3
VMLAL.S16 q0, d16, d6[3]
; Process 2 samples here if we have > 2 left (still reading one extra y value).
xcorr_kernel_neon_process2
ADDS r12, r12, #2
BLE xcorr_kernel_neon_process1
; Load x[0...1]
VLD2.16 {d6[],d7[]}, [r4]!
; Use VAND since it's a data processing instruction again.
VAND d4, d5, d5
SUB r12, r12, #2
; Load y[4...5]
VLD1.32 {d5[]}, [r5]!
VMLAL.S16 q0, d4, d6
VEXT.16 d16, d4, d5, #1
; Replace bottom copy of {y5,y4} in d5 with {y3,y2} from d4, using VSRI
; instead of VEXT, since it's a data-processing instruction.
VSRI.64 d5, d4, #32
VMLAL.S16 q0, d16, d7
; Process 1 sample using the extra y value we loaded above.
xcorr_kernel_neon_process1
; Load next *x
VLD1.16 {d6[]}, [r4]!
ADDS r12, r12, #1
; y[0...3] are left in d5 from prior iteration(s) (if any)
VMLAL.S16 q0, d5, d6
MOVLE pc, lr
; Now process 1 last sample, not reading ahead.
; Load last *y
VLD1.16 {d4[]}, [r5]!
VSRI.64 d4, d5, #16
; Load last *x
VLD1.16 {d6[]}, [r4]!
VMLAL.S16 q0, d4, d6
MOV pc, lr
ENDP
; opus_val32 celt_pitch_xcorr_neon(opus_val16 *_x, opus_val16 *_y,
; opus_val32 *xcorr, int len, int max_pitch)
celt_pitch_xcorr_neon PROC
; input:
; r0 = opus_val16 *_x
; r1 = opus_val16 *_y
; r2 = opus_val32 *xcorr
; r3 = int len
; output:
; r0 = int maxcorr
; internal usage:
; r4 = opus_val16 *x (for xcorr_kernel_neon())
; r5 = opus_val16 *y (for xcorr_kernel_neon())
; r6 = int max_pitch
; r12 = int j
; q15 = int maxcorr[4] (q15 is not used by xcorr_kernel_neon())
STMFD sp!, {r4-r6, lr}
LDR r6, [sp, #16]
VMOV.S32 q15, #1
; if (max_pitch < 4) goto celt_pitch_xcorr_neon_process4_done
SUBS r6, r6, #4
BLT celt_pitch_xcorr_neon_process4_done
celt_pitch_xcorr_neon_process4
; xcorr_kernel_neon parameters:
; r3 = len, r4 = _x, r5 = _y, q0 = {0, 0, 0, 0}
MOV r4, r0
MOV r5, r1
VEOR q0, q0, q0
; xcorr_kernel_neon only modifies r4, r5, r12, and q0...q3.
; So we don't save/restore any other registers.
BL xcorr_kernel_neon_start
SUBS r6, r6, #4
VST1.32 {q0}, [r2]!
; _y += 4
ADD r1, r1, #8
VMAX.S32 q15, q15, q0
; if (max_pitch < 4) goto celt_pitch_xcorr_neon_process4_done
BGE celt_pitch_xcorr_neon_process4
; We have less than 4 sums left to compute.
celt_pitch_xcorr_neon_process4_done
ADDS r6, r6, #4
; Reduce maxcorr to a single value
VMAX.S32 d30, d30, d31
VPMAX.S32 d30, d30, d30
; if (max_pitch <= 0) goto celt_pitch_xcorr_neon_done
BLE celt_pitch_xcorr_neon_done
; Now compute each remaining sum one at a time.
celt_pitch_xcorr_neon_process_remaining
MOV r4, r0
MOV r5, r1
VMOV.I32 q0, #0
SUBS r12, r3, #8
BLT celt_pitch_xcorr_neon_process_remaining4
; Sum terms 8 at a time.
celt_pitch_xcorr_neon_process_remaining_loop8
; Load x[0...7]
VLD1.16 {q1}, [r4]!
; Load y[0...7]
VLD1.16 {q2}, [r5]!
SUBS r12, r12, #8
VMLAL.S16 q0, d4, d2
VMLAL.S16 q0, d5, d3
BGE celt_pitch_xcorr_neon_process_remaining_loop8
; Sum terms 4 at a time.
celt_pitch_xcorr_neon_process_remaining4
ADDS r12, r12, #4
BLT celt_pitch_xcorr_neon_process_remaining4_done
; Load x[0...3]
VLD1.16 {d2}, [r4]!
; Load y[0...3]
VLD1.16 {d3}, [r5]!
SUB r12, r12, #4
VMLAL.S16 q0, d3, d2
celt_pitch_xcorr_neon_process_remaining4_done
; Reduce the sum to a single value.
VADD.S32 d0, d0, d1
VPADDL.S32 d0, d0
ADDS r12, r12, #4
BLE celt_pitch_xcorr_neon_process_remaining_loop_done
; Sum terms 1 at a time.
celt_pitch_xcorr_neon_process_remaining_loop1
VLD1.16 {d2[]}, [r4]!
VLD1.16 {d3[]}, [r5]!
SUBS r12, r12, #1
VMLAL.S16 q0, d2, d3
BGT celt_pitch_xcorr_neon_process_remaining_loop1
celt_pitch_xcorr_neon_process_remaining_loop_done
VST1.32 {d0[0]}, [r2]!
VMAX.S32 d30, d30, d0
SUBS r6, r6, #1
; _y++
ADD r1, r1, #2
; if (--max_pitch > 0) goto celt_pitch_xcorr_neon_process_remaining
BGT celt_pitch_xcorr_neon_process_remaining
celt_pitch_xcorr_neon_done
VMOV.32 r0, d30[0]
LDMFD sp!, {r4-r6, pc}
ENDP
ENDIF
IF OPUS_ARM_MAY_HAVE_EDSP
; This will get used on ARMv7 devices without NEON, so it has been optimized
; to take advantage of dual-issuing where possible.
xcorr_kernel_edsp PROC
xcorr_kernel_edsp_start
; input:
; r3 = int len
; r4 = opus_val16 *_x (must be 32-bit aligned)
; r5 = opus_val16 *_y (must be 32-bit aligned)
; r6...r9 = opus_val32 sum[4]
; output:
; r6...r9 = opus_val32 sum[4]
; preserved: r0-r5
; internal usage
; r2 = int j
; r12,r14 = opus_val16 x[4]
; r10,r11 = opus_val16 y[4]
STMFD sp!, {r2,r4,r5,lr}
LDR r10, [r5], #4 ; Load y[0...1]
SUBS r2, r3, #4 ; j = len-4
LDR r11, [r5], #4 ; Load y[2...3]
BLE xcorr_kernel_edsp_process4_done
LDR r12, [r4], #4 ; Load x[0...1]
; Stall
xcorr_kernel_edsp_process4
; The multiplies must issue from pipeline 0, and can't dual-issue with each
; other. Every other instruction here dual-issues with a multiply, and is
; thus "free". There should be no stalls in the body of the loop.
SMLABB r6, r12, r10, r6 ; sum[0] = MAC16_16(sum[0],x_0,y_0)
LDR r14, [r4], #4 ; Load x[2...3]
SMLABT r7, r12, r10, r7 ; sum[1] = MAC16_16(sum[1],x_0,y_1)
SUBS r2, r2, #4 ; j-=4
SMLABB r8, r12, r11, r8 ; sum[2] = MAC16_16(sum[2],x_0,y_2)
SMLABT r9, r12, r11, r9 ; sum[3] = MAC16_16(sum[3],x_0,y_3)
SMLATT r6, r12, r10, r6 ; sum[0] = MAC16_16(sum[0],x_1,y_1)
LDR r10, [r5], #4 ; Load y[4...5]
SMLATB r7, r12, r11, r7 ; sum[1] = MAC16_16(sum[1],x_1,y_2)
SMLATT r8, r12, r11, r8 ; sum[2] = MAC16_16(sum[2],x_1,y_3)
SMLATB r9, r12, r10, r9 ; sum[3] = MAC16_16(sum[3],x_1,y_4)
LDRGT r12, [r4], #4 ; Load x[0...1]
SMLABB r6, r14, r11, r6 ; sum[0] = MAC16_16(sum[0],x_2,y_2)
SMLABT r7, r14, r11, r7 ; sum[1] = MAC16_16(sum[1],x_2,y_3)
SMLABB r8, r14, r10, r8 ; sum[2] = MAC16_16(sum[2],x_2,y_4)
SMLABT r9, r14, r10, r9 ; sum[3] = MAC16_16(sum[3],x_2,y_5)
SMLATT r6, r14, r11, r6 ; sum[0] = MAC16_16(sum[0],x_3,y_3)
LDR r11, [r5], #4 ; Load y[6...7]
SMLATB r7, r14, r10, r7 ; sum[1] = MAC16_16(sum[1],x_3,y_4)
SMLATT r8, r14, r10, r8 ; sum[2] = MAC16_16(sum[2],x_3,y_5)
SMLATB r9, r14, r11, r9 ; sum[3] = MAC16_16(sum[3],x_3,y_6)
BGT xcorr_kernel_edsp_process4
xcorr_kernel_edsp_process4_done
ADDS r2, r2, #4
BLE xcorr_kernel_edsp_done
LDRH r12, [r4], #2 ; r12 = *x++
SUBS r2, r2, #1 ; j--
; Stall
SMLABB r6, r12, r10, r6 ; sum[0] = MAC16_16(sum[0],x,y_0)
LDRHGT r14, [r4], #2 ; r14 = *x++
SMLABT r7, r12, r10, r7 ; sum[1] = MAC16_16(sum[1],x,y_1)
SMLABB r8, r12, r11, r8 ; sum[2] = MAC16_16(sum[2],x,y_2)
SMLABT r9, r12, r11, r9 ; sum[3] = MAC16_16(sum[3],x,y_3)
BLE xcorr_kernel_edsp_done
SMLABT r6, r14, r10, r6 ; sum[0] = MAC16_16(sum[0],x,y_1)
SUBS r2, r2, #1 ; j--
SMLABB r7, r14, r11, r7 ; sum[1] = MAC16_16(sum[1],x,y_2)
LDRH r10, [r5], #2 ; r10 = y_4 = *y++
SMLABT r8, r14, r11, r8 ; sum[2] = MAC16_16(sum[2],x,y_3)
LDRHGT r12, [r4], #2 ; r12 = *x++
SMLABB r9, r14, r10, r9 ; sum[3] = MAC16_16(sum[3],x,y_4)
BLE xcorr_kernel_edsp_done
SMLABB r6, r12, r11, r6 ; sum[0] = MAC16_16(sum[0],tmp,y_2)
CMP r2, #1 ; j--
SMLABT r7, r12, r11, r7 ; sum[1] = MAC16_16(sum[1],tmp,y_3)
LDRH r2, [r5], #2 ; r2 = y_5 = *y++
SMLABB r8, r12, r10, r8 ; sum[2] = MAC16_16(sum[2],tmp,y_4)
LDRHGT r14, [r4] ; r14 = *x
SMLABB r9, r12, r2, r9 ; sum[3] = MAC16_16(sum[3],tmp,y_5)
BLE xcorr_kernel_edsp_done
SMLABT r6, r14, r11, r6 ; sum[0] = MAC16_16(sum[0],tmp,y_3)
LDRH r11, [r5] ; r11 = y_6 = *y
SMLABB r7, r14, r10, r7 ; sum[1] = MAC16_16(sum[1],tmp,y_4)
SMLABB r8, r14, r2, r8 ; sum[2] = MAC16_16(sum[2],tmp,y_5)
SMLABB r9, r14, r11, r9 ; sum[3] = MAC16_16(sum[3],tmp,y_6)
xcorr_kernel_edsp_done
LDMFD sp!, {r2,r4,r5,pc}
ENDP
celt_pitch_xcorr_edsp PROC
; input:
; r0 = opus_val16 *_x (must be 32-bit aligned)
; r1 = opus_val16 *_y (only needs to be 16-bit aligned)
; r2 = opus_val32 *xcorr
; r3 = int len
; output:
; r0 = maxcorr
; internal usage
; r4 = opus_val16 *x
; r5 = opus_val16 *y
; r6 = opus_val32 sum0
; r7 = opus_val32 sum1
; r8 = opus_val32 sum2
; r9 = opus_val32 sum3
; r1 = int max_pitch
; r12 = int j
STMFD sp!, {r4-r11, lr}
MOV r5, r1
LDR r1, [sp, #36]
MOV r4, r0
TST r5, #3
; maxcorr = 1
MOV r0, #1
BEQ celt_pitch_xcorr_edsp_process1u_done
; Compute one sum at the start to make y 32-bit aligned.
SUBS r12, r3, #4
; r14 = sum = 0
MOV r14, #0
LDRH r8, [r5], #2
BLE celt_pitch_xcorr_edsp_process1u_loop4_done
LDR r6, [r4], #4
MOV r8, r8, LSL #16
celt_pitch_xcorr_edsp_process1u_loop4
LDR r9, [r5], #4
SMLABT r14, r6, r8, r14 ; sum = MAC16_16(sum, x_0, y_0)
LDR r7, [r4], #4
SMLATB r14, r6, r9, r14 ; sum = MAC16_16(sum, x_1, y_1)
LDR r8, [r5], #4
SMLABT r14, r7, r9, r14 ; sum = MAC16_16(sum, x_2, y_2)
SUBS r12, r12, #4 ; j-=4
SMLATB r14, r7, r8, r14 ; sum = MAC16_16(sum, x_3, y_3)
LDRGT r6, [r4], #4
BGT celt_pitch_xcorr_edsp_process1u_loop4
MOV r8, r8, LSR #16
celt_pitch_xcorr_edsp_process1u_loop4_done
ADDS r12, r12, #4
celt_pitch_xcorr_edsp_process1u_loop1
LDRHGE r6, [r4], #2
; Stall
SMLABBGE r14, r6, r8, r14 ; sum = MAC16_16(sum, *x, *y)
SUBSGE r12, r12, #1
LDRHGT r8, [r5], #2
BGT celt_pitch_xcorr_edsp_process1u_loop1
; Restore _x
SUB r4, r4, r3, LSL #1
; Restore and advance _y
SUB r5, r5, r3, LSL #1
; maxcorr = max(maxcorr, sum)
CMP r0, r14
ADD r5, r5, #2
MOVLT r0, r14
SUBS r1, r1, #1
; xcorr[i] = sum
STR r14, [r2], #4
BLE celt_pitch_xcorr_edsp_done
celt_pitch_xcorr_edsp_process1u_done
; if (max_pitch < 4) goto celt_pitch_xcorr_edsp_process2
SUBS r1, r1, #4
BLT celt_pitch_xcorr_edsp_process2
celt_pitch_xcorr_edsp_process4
; xcorr_kernel_edsp parameters:
; r3 = len, r4 = _x, r5 = _y, r6...r9 = sum[4] = {0, 0, 0, 0}
MOV r6, #0
MOV r7, #0
MOV r8, #0
MOV r9, #0
BL xcorr_kernel_edsp_start ; xcorr_kernel_edsp(_x, _y+i, xcorr+i, len)
; maxcorr = max(maxcorr, sum0, sum1, sum2, sum3)
CMP r0, r6
; _y+=4
ADD r5, r5, #8
MOVLT r0, r6
CMP r0, r7
MOVLT r0, r7
CMP r0, r8
MOVLT r0, r8
CMP r0, r9
MOVLT r0, r9
STMIA r2!, {r6-r9}
SUBS r1, r1, #4
BGE celt_pitch_xcorr_edsp_process4
celt_pitch_xcorr_edsp_process2
ADDS r1, r1, #2
BLT celt_pitch_xcorr_edsp_process1a
SUBS r12, r3, #4
; {r10, r11} = {sum0, sum1} = {0, 0}
MOV r10, #0
MOV r11, #0
LDR r8, [r5], #4
BLE celt_pitch_xcorr_edsp_process2_loop_done
LDR r6, [r4], #4
LDR r9, [r5], #4
celt_pitch_xcorr_edsp_process2_loop4
SMLABB r10, r6, r8, r10 ; sum0 = MAC16_16(sum0, x_0, y_0)
LDR r7, [r4], #4
SMLABT r11, r6, r8, r11 ; sum1 = MAC16_16(sum1, x_0, y_1)
SUBS r12, r12, #4 ; j-=4
SMLATT r10, r6, r8, r10 ; sum0 = MAC16_16(sum0, x_1, y_1)
LDR r8, [r5], #4
SMLATB r11, r6, r9, r11 ; sum1 = MAC16_16(sum1, x_1, y_2)
LDRGT r6, [r4], #4
SMLABB r10, r7, r9, r10 ; sum0 = MAC16_16(sum0, x_2, y_2)
SMLABT r11, r7, r9, r11 ; sum1 = MAC16_16(sum1, x_2, y_3)
SMLATT r10, r7, r9, r10 ; sum0 = MAC16_16(sum0, x_3, y_3)
LDRGT r9, [r5], #4
SMLATB r11, r7, r8, r11 ; sum1 = MAC16_16(sum1, x_3, y_4)
BGT celt_pitch_xcorr_edsp_process2_loop4
celt_pitch_xcorr_edsp_process2_loop_done
ADDS r12, r12, #2
BLE celt_pitch_xcorr_edsp_process2_1
LDR r6, [r4], #4
; Stall
SMLABB r10, r6, r8, r10 ; sum0 = MAC16_16(sum0, x_0, y_0)
LDR r9, [r5], #4
SMLABT r11, r6, r8, r11 ; sum1 = MAC16_16(sum1, x_0, y_1)
SUB r12, r12, #2
SMLATT r10, r6, r8, r10 ; sum0 = MAC16_16(sum0, x_1, y_1)
MOV r8, r9
SMLATB r11, r6, r9, r11 ; sum1 = MAC16_16(sum1, x_1, y_2)
celt_pitch_xcorr_edsp_process2_1
LDRH r6, [r4], #2
ADDS r12, r12, #1
; Stall
SMLABB r10, r6, r8, r10 ; sum0 = MAC16_16(sum0, x_0, y_0)
LDRHGT r7, [r4], #2
SMLABT r11, r6, r8, r11 ; sum1 = MAC16_16(sum1, x_0, y_1)
BLE celt_pitch_xcorr_edsp_process2_done
LDRH r9, [r5], #2
SMLABT r10, r7, r8, r10 ; sum0 = MAC16_16(sum0, x_0, y_1)
SMLABB r11, r7, r9, r11 ; sum1 = MAC16_16(sum1, x_0, y_2)
celt_pitch_xcorr_edsp_process2_done
; Restore _x
SUB r4, r4, r3, LSL #1
; Restore and advance _y
SUB r5, r5, r3, LSL #1
; maxcorr = max(maxcorr, sum0)
CMP r0, r10
ADD r5, r5, #2
MOVLT r0, r10
SUB r1, r1, #2
; maxcorr = max(maxcorr, sum1)
CMP r0, r11
; xcorr[i] = sum
STR r10, [r2], #4
MOVLT r0, r11
STR r11, [r2], #4
celt_pitch_xcorr_edsp_process1a
ADDS r1, r1, #1
BLT celt_pitch_xcorr_edsp_done
SUBS r12, r3, #4
; r14 = sum = 0
MOV r14, #0
BLT celt_pitch_xcorr_edsp_process1a_loop_done
LDR r6, [r4], #4
LDR r8, [r5], #4
LDR r7, [r4], #4
LDR r9, [r5], #4
celt_pitch_xcorr_edsp_process1a_loop4
SMLABB r14, r6, r8, r14 ; sum = MAC16_16(sum, x_0, y_0)
SUBS r12, r12, #4 ; j-=4
SMLATT r14, r6, r8, r14 ; sum = MAC16_16(sum, x_1, y_1)
LDRGE r6, [r4], #4
SMLABB r14, r7, r9, r14 ; sum = MAC16_16(sum, x_2, y_2)
LDRGE r8, [r5], #4
SMLATT r14, r7, r9, r14 ; sum = MAC16_16(sum, x_3, y_3)
LDRGE r7, [r4], #4
LDRGE r9, [r5], #4
BGE celt_pitch_xcorr_edsp_process1a_loop4
celt_pitch_xcorr_edsp_process1a_loop_done
ADDS r12, r12, #2
LDRGE r6, [r4], #4
LDRGE r8, [r5], #4
; Stall
SMLABBGE r14, r6, r8, r14 ; sum = MAC16_16(sum, x_0, y_0)
SUBGE r12, r12, #2
SMLATTGE r14, r6, r8, r14 ; sum = MAC16_16(sum, x_1, y_1)
ADDS r12, r12, #1
LDRHGE r6, [r4], #2
LDRHGE r8, [r5], #2
; Stall
SMLABBGE r14, r6, r8, r14 ; sum = MAC16_16(sum, *x, *y)
; maxcorr = max(maxcorr, sum)
CMP r0, r14
; xcorr[i] = sum
STR r14, [r2], #4
MOVLT r0, r14
celt_pitch_xcorr_edsp_done
LDMFD sp!, {r4-r11, pc}
ENDP
ENDIF
END

72
Frameworks/Opus/Opus/opus/celt/arm/fft_arm.h Normal file
View file

@ -0,0 +1,72 @@
/* Copyright (c) 2015 Xiph.Org Foundation
Written by Viswanath Puttagunta */
/**
@file fft_arm.h
@brief ARM Neon Intrinsic optimizations for fft using NE10 library
*/
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#if !defined(FFT_ARM_H)
#define FFT_ARM_H
#include "config.h"
#include "kiss_fft.h"
#if defined(HAVE_ARM_NE10)
int opus_fft_alloc_arm_neon(kiss_fft_state *st);
void opus_fft_free_arm_neon(kiss_fft_state *st);
void opus_fft_neon(const kiss_fft_state *st,
const kiss_fft_cpx *fin,
kiss_fft_cpx *fout);
void opus_ifft_neon(const kiss_fft_state *st,
const kiss_fft_cpx *fin,
kiss_fft_cpx *fout);
#if !defined(OPUS_HAVE_RTCD)
#define OVERRIDE_OPUS_FFT (1)
#define opus_fft_alloc_arch(_st, arch) \
((void)(arch), opus_fft_alloc_arm_neon(_st))
#define opus_fft_free_arch(_st, arch) \
((void)(arch), opus_fft_free_arm_neon(_st))
#define opus_fft(_st, _fin, _fout, arch) \
((void)(arch), opus_fft_neon(_st, _fin, _fout))
#define opus_ifft(_st, _fin, _fout, arch) \
((void)(arch), opus_ifft_neon(_st, _fin, _fout))
#endif /* OPUS_HAVE_RTCD */
#endif /* HAVE_ARM_NE10 */
#endif

80
Frameworks/Opus/Opus/opus/celt/arm/fixed_armv4.h Normal file
View file

@ -0,0 +1,80 @@
/* Copyright (C) 2013 Xiph.Org Foundation and contributors */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef FIXED_ARMv4_H
#define FIXED_ARMv4_H
/** 16x32 multiplication, followed by a 16-bit shift right. Results fits in 32 bits */
#undef MULT16_32_Q16
static OPUS_INLINE opus_val32 MULT16_32_Q16_armv4(opus_val16 a, opus_val32 b)
{
unsigned rd_lo;
int rd_hi;
__asm__(
"#MULT16_32_Q16\n\t"
"smull %0, %1, %2, %3\n\t"
: "=&r"(rd_lo), "=&r"(rd_hi)
: "%r"(b),"r"(a<<16)
);
return rd_hi;
}
#define MULT16_32_Q16(a, b) (MULT16_32_Q16_armv4(a, b))
/** 16x32 multiplication, followed by a 15-bit shift right. Results fits in 32 bits */
#undef MULT16_32_Q15
static OPUS_INLINE opus_val32 MULT16_32_Q15_armv4(opus_val16 a, opus_val32 b)
{
unsigned rd_lo;
int rd_hi;
__asm__(
"#MULT16_32_Q15\n\t"
"smull %0, %1, %2, %3\n\t"
: "=&r"(rd_lo), "=&r"(rd_hi)
: "%r"(b), "r"(a<<16)
);
/*We intentionally don't OR in the high bit of rd_lo for speed.*/
return rd_hi<<1;
}
#define MULT16_32_Q15(a, b) (MULT16_32_Q15_armv4(a, b))
/** 16x32 multiply, followed by a 15-bit shift right and 32-bit add.
b must fit in 31 bits.
Result fits in 32 bits. */
#undef MAC16_32_Q15
#define MAC16_32_Q15(c, a, b) ADD32(c, MULT16_32_Q15(a, b))
/** 16x32 multiply, followed by a 16-bit shift right and 32-bit add.
Result fits in 32 bits. */
#undef MAC16_32_Q16
#define MAC16_32_Q16(c, a, b) ADD32(c, MULT16_32_Q16(a, b))
/** 32x32 multiplication, followed by a 31-bit shift right. Results fits in 32 bits */
#undef MULT32_32_Q31
#define MULT32_32_Q31(a,b) (opus_val32)((((opus_int64)(a)) * ((opus_int64)(b)))>>31)
#endif

151
Frameworks/Opus/Opus/opus/celt/arm/fixed_armv5e.h Normal file
View file

@ -0,0 +1,151 @@
/* Copyright (C) 2007-2009 Xiph.Org Foundation
Copyright (C) 2003-2008 Jean-Marc Valin
Copyright (C) 2007-2008 CSIRO
Copyright (C) 2013 Parrot */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef FIXED_ARMv5E_H
#define FIXED_ARMv5E_H
#include "fixed_armv4.h"
/** 16x32 multiplication, followed by a 16-bit shift right. Results fits in 32 bits */
#undef MULT16_32_Q16
static OPUS_INLINE opus_val32 MULT16_32_Q16_armv5e(opus_val16 a, opus_val32 b)
{
int res;
__asm__(
"#MULT16_32_Q16\n\t"
"smulwb %0, %1, %2\n\t"
: "=r"(res)
: "r"(b),"r"(a)
);
return res;
}
#define MULT16_32_Q16(a, b) (MULT16_32_Q16_armv5e(a, b))
/** 16x32 multiplication, followed by a 15-bit shift right. Results fits in 32 bits */
#undef MULT16_32_Q15
static OPUS_INLINE opus_val32 MULT16_32_Q15_armv5e(opus_val16 a, opus_val32 b)
{
int res;
__asm__(
"#MULT16_32_Q15\n\t"
"smulwb %0, %1, %2\n\t"
: "=r"(res)
: "r"(b), "r"(a)
);
return res<<1;
}
#define MULT16_32_Q15(a, b) (MULT16_32_Q15_armv5e(a, b))
/** 16x32 multiply, followed by a 15-bit shift right and 32-bit add.
b must fit in 31 bits.
Result fits in 32 bits. */
#undef MAC16_32_Q15
static OPUS_INLINE opus_val32 MAC16_32_Q15_armv5e(opus_val32 c, opus_val16 a,
opus_val32 b)
{
int res;
__asm__(
"#MAC16_32_Q15\n\t"
"smlawb %0, %1, %2, %3;\n"
: "=r"(res)
: "r"(b<<1), "r"(a), "r"(c)
);
return res;
}
#define MAC16_32_Q15(c, a, b) (MAC16_32_Q15_armv5e(c, a, b))
/** 16x32 multiply, followed by a 16-bit shift right and 32-bit add.
Result fits in 32 bits. */
#undef MAC16_32_Q16
static OPUS_INLINE opus_val32 MAC16_32_Q16_armv5e(opus_val32 c, opus_val16 a,
opus_val32 b)
{
int res;
__asm__(
"#MAC16_32_Q16\n\t"
"smlawb %0, %1, %2, %3;\n"
: "=r"(res)
: "r"(b), "r"(a), "r"(c)
);
return res;
}
#define MAC16_32_Q16(c, a, b) (MAC16_32_Q16_armv5e(c, a, b))
/** 16x16 multiply-add where the result fits in 32 bits */
#undef MAC16_16
static OPUS_INLINE opus_val32 MAC16_16_armv5e(opus_val32 c, opus_val16 a,
opus_val16 b)
{
int res;
__asm__(
"#MAC16_16\n\t"
"smlabb %0, %1, %2, %3;\n"
: "=r"(res)
: "r"(a), "r"(b), "r"(c)
);
return res;
}
#define MAC16_16(c, a, b) (MAC16_16_armv5e(c, a, b))
/** 16x16 multiplication where the result fits in 32 bits */
#undef MULT16_16
static OPUS_INLINE opus_val32 MULT16_16_armv5e(opus_val16 a, opus_val16 b)
{
int res;
__asm__(
"#MULT16_16\n\t"
"smulbb %0, %1, %2;\n"
: "=r"(res)
: "r"(a), "r"(b)
);
return res;
}
#define MULT16_16(a, b) (MULT16_16_armv5e(a, b))
#ifdef OPUS_ARM_INLINE_MEDIA
#undef SIG2WORD16
static OPUS_INLINE opus_val16 SIG2WORD16_armv6(opus_val32 x)
{
celt_sig res;
__asm__(
"#SIG2WORD16\n\t"
"ssat %0, #16, %1, ASR #12\n\t"
: "=r"(res)
: "r"(x+2048)
);
return EXTRACT16(res);
}
#define SIG2WORD16(x) (SIG2WORD16_armv6(x))
#endif /* OPUS_ARM_INLINE_MEDIA */
#endif

121
Frameworks/Opus/Opus/opus/celt/arm/kiss_fft_armv4.h Normal file
View file

@ -0,0 +1,121 @@
/*Copyright (c) 2013, Xiph.Org Foundation and contributors.
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.*/
#ifndef KISS_FFT_ARMv4_H
#define KISS_FFT_ARMv4_H
#if !defined(KISS_FFT_GUTS_H)
#error "This file should only be included from _kiss_fft_guts.h"
#endif
#ifdef FIXED_POINT
#undef C_MUL
#define C_MUL(m,a,b) \
do{ \
int br__; \
int bi__; \
int tt__; \
__asm__ __volatile__( \
"#C_MUL\n\t" \
"ldrsh %[br], [%[bp], #0]\n\t" \
"ldm %[ap], {r0,r1}\n\t" \
"ldrsh %[bi], [%[bp], #2]\n\t" \
"smull %[tt], %[mi], r1, %[br]\n\t" \
"smlal %[tt], %[mi], r0, %[bi]\n\t" \
"rsb %[bi], %[bi], #0\n\t" \
"smull %[br], %[mr], r0, %[br]\n\t" \
"mov %[tt], %[tt], lsr #15\n\t" \
"smlal %[br], %[mr], r1, %[bi]\n\t" \
"orr %[mi], %[tt], %[mi], lsl #17\n\t" \
"mov %[br], %[br], lsr #15\n\t" \
"orr %[mr], %[br], %[mr], lsl #17\n\t" \
: [mr]"=r"((m).r), [mi]"=r"((m).i), \
[br]"=&r"(br__), [bi]"=r"(bi__), [tt]"=r"(tt__) \
: [ap]"r"(&(a)), [bp]"r"(&(b)) \
: "r0", "r1" \
); \
} \
while(0)
#undef C_MUL4
#define C_MUL4(m,a,b) \
do{ \
int br__; \
int bi__; \
int tt__; \
__asm__ __volatile__( \
"#C_MUL4\n\t" \
"ldrsh %[br], [%[bp], #0]\n\t" \
"ldm %[ap], {r0,r1}\n\t" \
"ldrsh %[bi], [%[bp], #2]\n\t" \
"smull %[tt], %[mi], r1, %[br]\n\t" \
"smlal %[tt], %[mi], r0, %[bi]\n\t" \
"rsb %[bi], %[bi], #0\n\t" \
"smull %[br], %[mr], r0, %[br]\n\t" \
"mov %[tt], %[tt], lsr #17\n\t" \
"smlal %[br], %[mr], r1, %[bi]\n\t" \
"orr %[mi], %[tt], %[mi], lsl #15\n\t" \
"mov %[br], %[br], lsr #17\n\t" \
"orr %[mr], %[br], %[mr], lsl #15\n\t" \
: [mr]"=r"((m).r), [mi]"=r"((m).i), \
[br]"=&r"(br__), [bi]"=r"(bi__), [tt]"=r"(tt__) \
: [ap]"r"(&(a)), [bp]"r"(&(b)) \
: "r0", "r1" \
); \
} \
while(0)
#undef C_MULC
#define C_MULC(m,a,b) \
do{ \
int br__; \
int bi__; \
int tt__; \
__asm__ __volatile__( \
"#C_MULC\n\t" \
"ldrsh %[br], [%[bp], #0]\n\t" \
"ldm %[ap], {r0,r1}\n\t" \
"ldrsh %[bi], [%[bp], #2]\n\t" \
"smull %[tt], %[mr], r0, %[br]\n\t" \
"smlal %[tt], %[mr], r1, %[bi]\n\t" \
"rsb %[bi], %[bi], #0\n\t" \
"smull %[br], %[mi], r1, %[br]\n\t" \
"mov %[tt], %[tt], lsr #15\n\t" \
"smlal %[br], %[mi], r0, %[bi]\n\t" \
"orr %[mr], %[tt], %[mr], lsl #17\n\t" \
"mov %[br], %[br], lsr #15\n\t" \
"orr %[mi], %[br], %[mi], lsl #17\n\t" \
: [mr]"=r"((m).r), [mi]"=r"((m).i), \
[br]"=&r"(br__), [bi]"=r"(bi__), [tt]"=r"(tt__) \
: [ap]"r"(&(a)), [bp]"r"(&(b)) \
: "r0", "r1" \
); \
} \
while(0)
#endif /* FIXED_POINT */
#endif /* KISS_FFT_ARMv4_H */

118
Frameworks/Opus/Opus/opus/celt/arm/kiss_fft_armv5e.h Normal file
View file

@ -0,0 +1,118 @@
/*Copyright (c) 2013, Xiph.Org Foundation and contributors.
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.*/
#ifndef KISS_FFT_ARMv5E_H
#define KISS_FFT_ARMv5E_H
#if !defined(KISS_FFT_GUTS_H)
#error "This file should only be included from _kiss_fft_guts.h"
#endif
#ifdef FIXED_POINT
#if defined(__thumb__)||defined(__thumb2__)
#define LDRD_CONS "Q"
#else
#define LDRD_CONS "Uq"
#endif
#undef C_MUL
#define C_MUL(m,a,b) \
do{ \
int mr1__; \
int mr2__; \
int mi__; \
long long aval__; \
int bval__; \
__asm__( \
"#C_MUL\n\t" \
"ldrd %[aval], %H[aval], %[ap]\n\t" \
"ldr %[bval], %[bp]\n\t" \
"smulwb %[mi], %H[aval], %[bval]\n\t" \
"smulwb %[mr1], %[aval], %[bval]\n\t" \
"smulwt %[mr2], %H[aval], %[bval]\n\t" \
"smlawt %[mi], %[aval], %[bval], %[mi]\n\t" \
: [mr1]"=r"(mr1__), [mr2]"=r"(mr2__), [mi]"=r"(mi__), \
[aval]"=&r"(aval__), [bval]"=r"(bval__) \
: [ap]LDRD_CONS(a), [bp]"m"(b) \
); \
(m).r = SHL32(SUB32(mr1__, mr2__), 1); \
(m).i = SHL32(mi__, 1); \
} \
while(0)
#undef C_MUL4
#define C_MUL4(m,a,b) \
do{ \
int mr1__; \
int mr2__; \
int mi__; \
long long aval__; \
int bval__; \
__asm__( \
"#C_MUL4\n\t" \
"ldrd %[aval], %H[aval], %[ap]\n\t" \
"ldr %[bval], %[bp]\n\t" \
"smulwb %[mi], %H[aval], %[bval]\n\t" \
"smulwb %[mr1], %[aval], %[bval]\n\t" \
"smulwt %[mr2], %H[aval], %[bval]\n\t" \
"smlawt %[mi], %[aval], %[bval], %[mi]\n\t" \
: [mr1]"=r"(mr1__), [mr2]"=r"(mr2__), [mi]"=r"(mi__), \
[aval]"=&r"(aval__), [bval]"=r"(bval__) \
: [ap]LDRD_CONS(a), [bp]"m"(b) \
); \
(m).r = SHR32(SUB32(mr1__, mr2__), 1); \
(m).i = SHR32(mi__, 1); \
} \
while(0)
#undef C_MULC
#define C_MULC(m,a,b) \
do{ \
int mr__; \
int mi1__; \
int mi2__; \
long long aval__; \
int bval__; \
__asm__( \
"#C_MULC\n\t" \
"ldrd %[aval], %H[aval], %[ap]\n\t" \
"ldr %[bval], %[bp]\n\t" \
"smulwb %[mr], %[aval], %[bval]\n\t" \
"smulwb %[mi1], %H[aval], %[bval]\n\t" \
"smulwt %[mi2], %[aval], %[bval]\n\t" \
"smlawt %[mr], %H[aval], %[bval], %[mr]\n\t" \
: [mr]"=r"(mr__), [mi1]"=r"(mi1__), [mi2]"=r"(mi2__), \
[aval]"=&r"(aval__), [bval]"=r"(bval__) \
: [ap]LDRD_CONS(a), [bp]"m"(b) \
); \
(m).r = SHL32(mr__, 1); \
(m).i = SHL32(SUB32(mi1__, mi2__), 1); \
} \
while(0)
#endif /* FIXED_POINT */
#endif /* KISS_FFT_GUTS_H */

60
Frameworks/Opus/Opus/opus/celt/arm/mdct_arm.h Normal file
View file

@ -0,0 +1,60 @@
/* Copyright (c) 2015 Xiph.Org Foundation
Written by Viswanath Puttagunta */
/**
@file arm_mdct.h
@brief ARM Neon Intrinsic optimizations for mdct using NE10 library
*/
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#if !defined(MDCT_ARM_H)
#define MDCT_ARM_H
#include "config.h"
#include "mdct.h"
#if defined(HAVE_ARM_NE10)
/** Compute a forward MDCT and scale by 4/N, trashes the input array */
void clt_mdct_forward_neon(const mdct_lookup *l, kiss_fft_scalar *in,
kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 *window, int overlap,
int shift, int stride, int arch);
void clt_mdct_backward_neon(const mdct_lookup *l, kiss_fft_scalar *in,
kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 *window, int overlap,
int shift, int stride, int arch);
#if !defined(OPUS_HAVE_RTCD)
#define OVERRIDE_OPUS_MDCT (1)
#define clt_mdct_forward(_l, _in, _out, _window, _int, _shift, _stride, _arch) \
clt_mdct_forward_neon(_l, _in, _out, _window, _int, _shift, _stride, _arch)
#define clt_mdct_backward(_l, _in, _out, _window, _int, _shift, _stride, _arch) \
clt_mdct_backward_neon(_l, _in, _out, _window, _int, _shift, _stride, _arch)
#endif /* OPUS_HAVE_RTCD */
#endif /* HAVE_ARM_NE10 */
#endif

68
Frameworks/Opus/Opus/opus/celt/arm/pitch_arm.h Normal file
View file

@ -0,0 +1,68 @@
/* Copyright (c) 2010 Xiph.Org Foundation
* Copyright (c) 2013 Parrot */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#if !defined(PITCH_ARM_H)
# define PITCH_ARM_H
# include "armcpu.h"
# if defined(FIXED_POINT)
# if defined(OPUS_ARM_MAY_HAVE_NEON)
opus_val32 celt_pitch_xcorr_neon(const opus_val16 *_x, const opus_val16 *_y,
opus_val32 *xcorr, int len, int max_pitch);
# endif
# if defined(OPUS_ARM_MAY_HAVE_MEDIA)
# define celt_pitch_xcorr_media MAY_HAVE_EDSP(celt_pitch_xcorr)
# endif
# if defined(OPUS_ARM_MAY_HAVE_EDSP)
opus_val32 celt_pitch_xcorr_edsp(const opus_val16 *_x, const opus_val16 *_y,
opus_val32 *xcorr, int len, int max_pitch);
# endif
# if !defined(OPUS_HAVE_RTCD)
# define OVERRIDE_PITCH_XCORR (1)
# define celt_pitch_xcorr(_x, _y, xcorr, len, max_pitch, arch) \
((void)(arch),PRESUME_NEON(celt_pitch_xcorr)(_x, _y, xcorr, len, max_pitch))
# endif
#else /* Start !FIXED_POINT */
/* Float case */
#if defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
void celt_pitch_xcorr_float_neon(const opus_val16 *_x, const opus_val16 *_y,
opus_val32 *xcorr, int len, int max_pitch);
#if !defined(OPUS_HAVE_RTCD) || defined(OPUS_ARM_PRESUME_NEON_INTR)
#define OVERRIDE_PITCH_XCORR (1)
# define celt_pitch_xcorr(_x, _y, xcorr, len, max_pitch, arch) \
((void)(arch),celt_pitch_xcorr_float_neon(_x, _y, xcorr, len, max_pitch))
#endif
#endif
#endif /* end !FIXED_POINT */
#endif

1033
Frameworks/Opus/Opus/opus/celt/bands.c
View file

File diff suppressed because it is too large Load diff

51
Frameworks/Opus/Opus/opus/celt/bands.h
View file

@ -39,9 +39,9 @@
/** Compute the amplitude (sqrt energy) in each of the bands
* @param m Mode data
* @param X Spectrum
* @param bands Square root of the energy for each band (returned)
* @param bandE Square root of the energy for each band (returned)
*/
void compute_band_energies(const CELTMode *m, const celt_sig *X, celt_ener *bandE, int end, int C, int M);
void compute_band_energies(const CELTMode *m, const celt_sig *X, celt_ener *bandE, int end, int C, int LM);
/*void compute_noise_energies(const CELTMode *m, const celt_sig *X, const opus_val16 *tonality, celt_ener *bandE);*/
@ -49,23 +49,25 @@ void compute_band_energies(const CELTMode *m, const celt_sig *X, celt_ener *band
equal to 1
* @param m Mode data
* @param X Spectrum (returned normalised)
* @param bands Square root of the energy for each band
* @param bandE Square root of the energy for each band
*/
void normalise_bands(const CELTMode *m, const celt_sig * OPUS_RESTRICT freq, celt_norm * OPUS_RESTRICT X, const celt_ener *bandE, int end, int C, int M);
/** Denormalise each band of X to restore full amplitude
* @param m Mode data
* @param X Spectrum (returned de-normalised)
* @param bands Square root of the energy for each band
* @param bandE Square root of the energy for each band
*/
void denormalise_bands(const CELTMode *m, const celt_norm * OPUS_RESTRICT X, celt_sig * OPUS_RESTRICT freq, const celt_ener *bandE, int end, int C, int M);
void denormalise_bands(const CELTMode *m, const celt_norm * OPUS_RESTRICT X,
celt_sig * OPUS_RESTRICT freq, const opus_val16 *bandE, int start,
int end, int M, int downsample, int silence);
#define SPREAD_NONE (0)
#define SPREAD_LIGHT (1)
#define SPREAD_NORMAL (2)
#define SPREAD_AGGRESSIVE (3)
int spreading_decision(const CELTMode *m, celt_norm *X, int *average,
int spreading_decision(const CELTMode *m, const celt_norm *X, int *average,
int last_decision, int *hf_average, int *tapset_decision, int update_hf,
int end, int C, int M);
@ -76,20 +78,43 @@ void measure_norm_mse(const CELTMode *m, float *X, float *X0, float *bandE, floa
void haar1(celt_norm *X, int N0, int stride);
/** Quantisation/encoding of the residual spectrum
* @param encode flag that indicates whether we're encoding (1) or decoding (0)
* @param m Mode data
* @param start First band to process
* @param end Last band to process + 1
* @param X Residual (normalised)
* @param Y Residual (normalised) for second channel (or NULL for mono)
* @param collapse_masks Anti-collapse tracking mask
* @param bandE Square root of the energy for each band
* @param pulses Bit allocation (per band) for PVQ
* @param shortBlocks Zero for long blocks, non-zero for short blocks
* @param spread Amount of spreading to use
* @param dual_stereo Zero for MS stereo, non-zero for dual stereo
* @param intensity First band to use intensity stereo
* @param tf_res Time-frequency resolution change
* @param total_bits Total number of bits that can be used for the frame (including the ones already spent)
* @param enc Entropy encoder
* @param balance Number of unallocated bits
* @param en Entropy coder state
* @param LM log2() of the number of 2.5 subframes in the frame
* @param codedBands Last band to receive bits + 1
* @param seed Random generator seed
* @param arch Run-time architecture (see opus_select_arch())
*/
void quant_all_bands(int encode, const CELTMode *m, int start, int end,
celt_norm * X, celt_norm * Y, unsigned char *collapse_masks, const celt_ener *bandE, int *pulses,
int time_domain, int fold, int dual_stereo, int intensity, int *tf_res,
opus_int32 total_bits, opus_int32 balance, ec_ctx *ec, int M, int codedBands, opus_uint32 *seed);
celt_norm * X, celt_norm * Y, unsigned char *collapse_masks,
const celt_ener *bandE, int *pulses, int shortBlocks, int spread,
int dual_stereo, int intensity, int *tf_res, opus_int32 total_bits,
opus_int32 balance, ec_ctx *ec, int M, int codedBands, opus_uint32 *seed,
int arch);
void anti_collapse(const CELTMode *m, celt_norm *X_, unsigned char *collapse_masks, int LM, int C, int size,
int start, int end, opus_val16 *logE, opus_val16 *prev1logE,
opus_val16 *prev2logE, int *pulses, opus_uint32 seed);
void anti_collapse(const CELTMode *m, celt_norm *X_,
unsigned char *collapse_masks, int LM, int C, int size, int start,
int end, const opus_val16 *logE, const opus_val16 *prev1logE,
const opus_val16 *prev2logE, const int *pulses, opus_uint32 seed,
int arch);
opus_uint32 celt_lcg_rand(opus_uint32 seed);
int hysteresis_decision(opus_val16 val, const opus_val16 *thresholds, const opus_val16 *hysteresis, int N, int prev);
#endif /* BANDS_H */

2876
Frameworks/Opus/Opus/opus/celt/celt.c
View file

File diff suppressed because it is too large Load diff

120
Frameworks/Opus/Opus/opus/celt/celt.h
View file

@ -50,10 +50,26 @@ extern "C" {
#define CELTDecoder OpusCustomDecoder
#define CELTMode OpusCustomMode
#define _celt_check_mode_ptr_ptr(ptr) ((ptr) + ((ptr) - (const CELTMode**)(ptr)))
typedef struct {
int valid;
float tonality;
float tonality_slope;
float noisiness;
float activity;
float music_prob;
int bandwidth;
}AnalysisInfo;
#define __celt_check_mode_ptr_ptr(ptr) ((ptr) + ((ptr) - (const CELTMode**)(ptr)))
#define __celt_check_analysis_ptr(ptr) ((ptr) + ((ptr) - (const AnalysisInfo*)(ptr)))
/* Encoder/decoder Requests */
/* Expose this option again when variable framesize actually works */
#define OPUS_FRAMESIZE_VARIABLE 5010 /**< Optimize the frame size dynamically */
#define CELT_SET_PREDICTION_REQUEST 10002
/** Controls the use of interframe prediction.
0=Independent frames
@ -81,12 +97,24 @@ extern "C" {
#define CELT_GET_MODE_REQUEST 10015
/** Get the CELTMode used by an encoder or decoder */
#define CELT_GET_MODE(x) CELT_GET_MODE_REQUEST, _celt_check_mode_ptr_ptr(x)
#define CELT_GET_MODE(x) CELT_GET_MODE_REQUEST, __celt_check_mode_ptr_ptr(x)
#define CELT_SET_SIGNALLING_REQUEST 10016
#define CELT_SET_SIGNALLING(x) CELT_SET_SIGNALLING_REQUEST, __opus_check_int(x)
#define CELT_SET_TONALITY_REQUEST 10018
#define CELT_SET_TONALITY(x) CELT_SET_TONALITY_REQUEST, __opus_check_int(x)
#define CELT_SET_TONALITY_SLOPE_REQUEST 10020
#define CELT_SET_TONALITY_SLOPE(x) CELT_SET_TONALITY_SLOPE_REQUEST, __opus_check_int(x)
#define CELT_SET_ANALYSIS_REQUEST 10022
#define CELT_SET_ANALYSIS(x) CELT_SET_ANALYSIS_REQUEST, __celt_check_analysis_ptr(x)
#define OPUS_SET_LFE_REQUEST 10024
#define OPUS_SET_LFE(x) OPUS_SET_LFE_REQUEST, __opus_check_int(x)
#define OPUS_SET_ENERGY_MASK_REQUEST 10026
#define OPUS_SET_ENERGY_MASK(x) OPUS_SET_ENERGY_MASK_REQUEST, __opus_check_val16_ptr(x)
/* Encoder stuff */
@ -94,7 +122,8 @@ int celt_encoder_get_size(int channels);
int celt_encode_with_ec(OpusCustomEncoder * OPUS_RESTRICT st, const opus_val16 * pcm, int frame_size, unsigned char *compressed, int nbCompressedBytes, ec_enc *enc);
int celt_encoder_init(CELTEncoder *st, opus_int32 sampling_rate, int channels);
int celt_encoder_init(CELTEncoder *st, opus_int32 sampling_rate, int channels,
int arch);
@ -105,11 +134,94 @@ int celt_decoder_get_size(int channels);
int celt_decoder_init(CELTDecoder *st, opus_int32 sampling_rate, int channels);
int celt_decode_with_ec(OpusCustomDecoder * OPUS_RESTRICT st, const unsigned char *data, int len, opus_val16 * OPUS_RESTRICT pcm, int frame_size, ec_dec *dec);
int celt_decode_with_ec(OpusCustomDecoder * OPUS_RESTRICT st, const unsigned char *data,
int len, opus_val16 * OPUS_RESTRICT pcm, int frame_size, ec_dec *dec, int accum);
#define celt_encoder_ctl opus_custom_encoder_ctl
#define celt_decoder_ctl opus_custom_decoder_ctl
#ifdef CUSTOM_MODES
#define OPUS_CUSTOM_NOSTATIC
#else
#define OPUS_CUSTOM_NOSTATIC static OPUS_INLINE
#endif
static const unsigned char trim_icdf[11] = {126, 124, 119, 109, 87, 41, 19, 9, 4, 2, 0};
/* Probs: NONE: 21.875%, LIGHT: 6.25%, NORMAL: 65.625%, AGGRESSIVE: 6.25% */
static const unsigned char spread_icdf[4] = {25, 23, 2, 0};
static const unsigned char tapset_icdf[3]={2,1,0};
#ifdef CUSTOM_MODES
static const unsigned char toOpusTable[20] = {
0xE0, 0xE8, 0xF0, 0xF8,
0xC0, 0xC8, 0xD0, 0xD8,
0xA0, 0xA8, 0xB0, 0xB8,
0x00, 0x00, 0x00, 0x00,
0x80, 0x88, 0x90, 0x98,
};
static const unsigned char fromOpusTable[16] = {
0x80, 0x88, 0x90, 0x98,
0x40, 0x48, 0x50, 0x58,
0x20, 0x28, 0x30, 0x38,
0x00, 0x08, 0x10, 0x18
};
static OPUS_INLINE int toOpus(unsigned char c)
{
int ret=0;
if (c<0xA0)
ret = toOpusTable[c>>3];
if (ret == 0)
return -1;
else
return ret|(c&0x7);
}
static OPUS_INLINE int fromOpus(unsigned char c)
{
if (c<0x80)
return -1;
else
return fromOpusTable[(c>>3)-16] | (c&0x7);
}
#endif /* CUSTOM_MODES */
#define COMBFILTER_MAXPERIOD 1024
#define COMBFILTER_MINPERIOD 15
extern const signed char tf_select_table[4][8];
int resampling_factor(opus_int32 rate);
void celt_preemphasis(const opus_val16 * OPUS_RESTRICT pcmp, celt_sig * OPUS_RESTRICT inp,
int N, int CC, int upsample, const opus_val16 *coef, celt_sig *mem, int clip);
void comb_filter(opus_val32 *y, opus_val32 *x, int T0, int T1, int N,
opus_val16 g0, opus_val16 g1, int tapset0, int tapset1,
const opus_val16 *window, int overlap, int arch);
#ifdef NON_STATIC_COMB_FILTER_CONST_C
void comb_filter_const_c(opus_val32 *y, opus_val32 *x, int T, int N,
opus_val16 g10, opus_val16 g11, opus_val16 g12);
#endif
#ifndef OVERRIDE_COMB_FILTER_CONST
# define comb_filter_const(y, x, T, N, g10, g11, g12, arch) \
((void)(arch),comb_filter_const_c(y, x, T, N, g10, g11, g12))
#endif
void init_caps(const CELTMode *m,int *cap,int LM,int C);
#ifdef RESYNTH
void deemphasis(celt_sig *in[], opus_val16 *pcm, int N, int C, int downsample, const opus_val16 *coef, celt_sig *mem);
void celt_synthesis(const CELTMode *mode, celt_norm *X, celt_sig * out_syn[],
opus_val16 *oldBandE, int start, int effEnd, int C, int CC, int isTransient,
int LM, int downsample, int silence);
#endif
#ifdef __cplusplus
}
#endif

141
Frameworks/Opus/Opus/opus/celt/celt.vcxproj.filters
View file

@ -1,141 +0,0 @@
<?xml version="1.0" encoding="utf-8"?>
<Project ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
<ItemGroup>
<Filter Include="Source Files">
<UniqueIdentifier>{4FC737F1-C7A5-4376-A066-2A32D752A2FF}</UniqueIdentifier>
<Extensions>cpp;c;cc;cxx;def;odl;idl;hpj;bat;asm;asmx</Extensions>
</Filter>
<Filter Include="Header Files">
<UniqueIdentifier>{93995380-89BD-4b04-88EB-625FBE52EBFB}</UniqueIdentifier>
<Extensions>h;hpp;hxx;hm;inl;inc;xsd</Extensions>
</Filter>
<Filter Include="Resource Files">
<UniqueIdentifier>{67DA6AB6-F800-4c08-8B7A-83BB121AAD01}</UniqueIdentifier>
<Extensions>rc;ico;cur;bmp;dlg;rc2;rct;bin;rgs;gif;jpg;jpeg;jpe;resx;tiff;tif;png;wav;mfcribbon-ms</Extensions>
</Filter>
</ItemGroup>
<ItemGroup>
<ClInclude Include="arch.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="bands.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="celt.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="cwrs.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="ecintrin.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="entcode.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="entdec.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="entenc.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="fixed_debug.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="fixed_generic.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="float_cast.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="kiss_fft.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="laplace.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="mathops.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="mdct.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="modes.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="os_support.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="pitch.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="celt_lpc.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="quant_bands.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="rate.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="stack_alloc.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="vq.h">
<Filter>Header Files</Filter>
</ClInclude>
<ClInclude Include="..\win32\config.h">
<Filter>Header Files</Filter>
</ClInclude>
</ItemGroup>
<ItemGroup>
<ClCompile Include="bands.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="celt.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="cwrs.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="entcode.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="entdec.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="entenc.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="kiss_fft.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="laplace.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="mdct.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="modes.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="pitch.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="celt_lpc.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="quant_bands.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="rate.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="vq.c">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="mathops.c">
<Filter>Source Files</Filter>
</ClCompile>
</ItemGroup>
</Project>

1243
Frameworks/Opus/Opus/opus/celt/celt_decoder.c Normal file
View file

File diff suppressed because it is too large Load diff

2407
Frameworks/Opus/Opus/opus/celt/celt_encoder.c Normal file
View file

File diff suppressed because it is too large Load diff

191
Frameworks/Opus/Opus/opus/celt/celt_lpc.c
View file

@ -32,6 +32,7 @@
#include "celt_lpc.h"
#include "stack_alloc.h"
#include "mathops.h"
#include "pitch.h"
void _celt_lpc(
opus_val16 *_lpc, /* out: [0...p-1] LPC coefficients */
@ -87,42 +88,77 @@ int p
#endif
}
void celt_fir(const opus_val16 *x,
void celt_fir_c(
const opus_val16 *_x,
const opus_val16 *num,
opus_val16 *y,
opus_val16 *_y,
int N,
int ord,
opus_val16 *mem)
opus_val16 *mem,
int arch)
{
int i,j;
VARDECL(opus_val16, rnum);
VARDECL(opus_val16, x);
SAVE_STACK;
ALLOC(rnum, ord, opus_val16);
ALLOC(x, N+ord, opus_val16);
for(i=0;i<ord;i++)
rnum[i] = num[ord-i-1];
for(i=0;i<ord;i++)
x[i] = mem[ord-i-1];
for (i=0;i<N;i++)
x[i+ord]=_x[i];
for(i=0;i<ord;i++)
mem[i] = _x[N-i-1];
#ifdef SMALL_FOOTPRINT
(void)arch;
for (i=0;i<N;i++)
{
opus_val32 sum = SHL32(EXTEND32(x[i]), SIG_SHIFT);
opus_val32 sum = SHL32(EXTEND32(_x[i]), SIG_SHIFT);
for (j=0;j<ord;j++)
{
sum += MULT16_16(num[j],mem[j]);
sum = MAC16_16(sum,rnum[j],x[i+j]);
}
for (j=ord-1;j>=1;j--)
_y[i] = SATURATE16(PSHR32(sum, SIG_SHIFT));
}
#else
for (i=0;i<N-3;i+=4)
{
mem[j]=mem[j-1];
opus_val32 sum[4]={0,0,0,0};
xcorr_kernel(rnum, x+i, sum, ord, arch);
_y[i ] = SATURATE16(ADD32(EXTEND32(_x[i ]), PSHR32(sum[0], SIG_SHIFT)));
_y[i+1] = SATURATE16(ADD32(EXTEND32(_x[i+1]), PSHR32(sum[1], SIG_SHIFT)));
_y[i+2] = SATURATE16(ADD32(EXTEND32(_x[i+2]), PSHR32(sum[2], SIG_SHIFT)));
_y[i+3] = SATURATE16(ADD32(EXTEND32(_x[i+3]), PSHR32(sum[3], SIG_SHIFT)));
}
mem[0] = x[i];
y[i] = ROUND16(sum, SIG_SHIFT);
for (;i<N;i++)
{
opus_val32 sum = 0;
for (j=0;j<ord;j++)
sum = MAC16_16(sum,rnum[j],x[i+j]);
_y[i] = SATURATE16(ADD32(EXTEND32(_x[i]), PSHR32(sum, SIG_SHIFT)));
}
#endif
RESTORE_STACK;
}
void celt_iir(const opus_val32 *x,
void celt_iir(const opus_val32 *_x,
const opus_val16 *den,
opus_val32 *y,
opus_val32 *_y,
int N,
int ord,
opus_val16 *mem)
opus_val16 *mem,
int arch)
{
#ifdef SMALL_FOOTPRINT
int i,j;
(void)arch;
for (i=0;i<N;i++)
{
opus_val32 sum = x[i];
opus_val32 sum = _x[i];
for (j=0;j<ord;j++)
{
sum -= MULT16_16(den[j],mem[j]);
@ -132,26 +168,88 @@ void celt_iir(const opus_val32 *x,
mem[j]=mem[j-1];
}
mem[0] = ROUND16(sum,SIG_SHIFT);
y[i] = sum;
_y[i] = sum;
}
#else
int i,j;
VARDECL(opus_val16, rden);
VARDECL(opus_val16, y);
SAVE_STACK;
celt_assert((ord&3)==0);
ALLOC(rden, ord, opus_val16);
ALLOC(y, N+ord, opus_val16);
for(i=0;i<ord;i++)
rden[i] = den[ord-i-1];
for(i=0;i<ord;i++)
y[i] = -mem[ord-i-1];
for(;i<N+ord;i++)
y[i]=0;
for (i=0;i<N-3;i+=4)
{
/* Unroll by 4 as if it were an FIR filter */
opus_val32 sum[4];
sum[0]=_x[i];
sum[1]=_x[i+1];
sum[2]=_x[i+2];
sum[3]=_x[i+3];
xcorr_kernel(rden, y+i, sum, ord, arch);
/* Patch up the result to compensate for the fact that this is an IIR */
y[i+ord ] = -ROUND16(sum[0],SIG_SHIFT);
_y[i ] = sum[0];
sum[1] = MAC16_16(sum[1], y[i+ord ], den[0]);
y[i+ord+1] = -ROUND16(sum[1],SIG_SHIFT);
_y[i+1] = sum[1];
sum[2] = MAC16_16(sum[2], y[i+ord+1], den[0]);
sum[2] = MAC16_16(sum[2], y[i+ord ], den[1]);
y[i+ord+2] = -ROUND16(sum[2],SIG_SHIFT);
_y[i+2] = sum[2];
sum[3] = MAC16_16(sum[3], y[i+ord+2], den[0]);
sum[3] = MAC16_16(sum[3], y[i+ord+1], den[1]);
sum[3] = MAC16_16(sum[3], y[i+ord ], den[2]);
y[i+ord+3] = -ROUND16(sum[3],SIG_SHIFT);
_y[i+3] = sum[3];
}
for (;i<N;i++)
{
opus_val32 sum = _x[i];
for (j=0;j<ord;j++)
sum -= MULT16_16(rden[j],y[i+j]);
y[i+ord] = ROUND16(sum,SIG_SHIFT);
_y[i] = sum;
}
for(i=0;i<ord;i++)
mem[i] = _y[N-i-1];
RESTORE_STACK;
#endif
}
void _celt_autocorr(
int _celt_autocorr(
const opus_val16 *x, /* in: [0...n-1] samples x */
opus_val32 *ac, /* out: [0...lag-1] ac values */
const opus_val16 *window,
int overlap,
int lag,
int n
int n,
int arch
)
{
opus_val32 d;
int i;
int i, k;
int fastN=n-lag;
int shift;
const opus_val16 *xptr;
VARDECL(opus_val16, xx);
SAVE_STACK;
ALLOC(xx, n, opus_val16);
celt_assert(n>0);
celt_assert(overlap>=0);
if (overlap == 0)
{
xptr = x;
} else {
for (i=0;i<n;i++)
xx[i] = x[i];
for (i=0;i<overlap;i++)
@ -159,30 +257,59 @@ void _celt_autocorr(
xx[i] = MULT16_16_Q15(x[i],window[i]);
xx[n-i-1] = MULT16_16_Q15(x[n-i-1],window[i]);
}
xptr = xx;
}
shift=0;
#ifdef FIXED_POINT
{
opus_val32 ac0=0;
int shift;
for(i=0;i<n;i++)
ac0 += SHR32(MULT16_16(xx[i],xx[i]),9);
ac0 += 1+n;
opus_val32 ac0;
ac0 = 1+(n<<7);
if (n&1) ac0 += SHR32(MULT16_16(xptr[0],xptr[0]),9);
for(i=(n&1);i<n;i+=2)
{
ac0 += SHR32(MULT16_16(xptr[i],xptr[i]),9);
ac0 += SHR32(MULT16_16(xptr[i+1],xptr[i+1]),9);
}
shift = celt_ilog2(ac0)-30+10;
shift = (shift+1)/2;
shift = (shift)/2;
if (shift>0)
{
for(i=0;i<n;i++)
xx[i] = VSHR32(xx[i], shift);
xx[i] = PSHR32(xptr[i], shift);
xptr = xx;
} else
shift = 0;
}
#endif
while (lag>=0)
celt_pitch_xcorr(xptr, xptr, ac, fastN, lag+1, arch);
for (k=0;k<=lag;k++)
{
for (i = lag, d = 0; i < n; i++)
d += xx[i] * xx[i-lag];
ac[lag] = d;
/*printf ("%f ", ac[lag]);*/
lag--;
for (i = k+fastN, d = 0; i < n; i++)
d = MAC16_16(d, xptr[i], xptr[i-k]);
ac[k] += d;
}
/*printf ("\n");*/
ac[0] += 10;
#ifdef FIXED_POINT
shift = 2*shift;
if (shift<=0)
ac[0] += SHL32((opus_int32)1, -shift);
if (ac[0] < 268435456)
{
int shift2 = 29 - EC_ILOG(ac[0]);
for (i=0;i<=lag;i++)
ac[i] = SHL32(ac[i], shift2);
shift -= shift2;
} else if (ac[0] >= 536870912)
{
int shift2=1;
if (ac[0] >= 1073741824)
shift2++;
for (i=0;i<=lag;i++)
ac[i] = SHR32(ac[i], shift2);
shift += shift2;
}
#endif
RESTORE_STACK;
return shift;
}

22
Frameworks/Opus/Opus/opus/celt/celt_lpc.h
View file

@ -29,25 +29,39 @@
#define PLC_H
#include "arch.h"
#include "cpu_support.h"
#if defined(OPUS_X86_MAY_HAVE_SSE4_1)
#include "x86/celt_lpc_sse.h"
#endif
#define LPC_ORDER 24
void _celt_lpc(opus_val16 *_lpc, const opus_val32 *ac, int p);
void celt_fir(const opus_val16 *x,
void celt_fir_c(
const opus_val16 *x,
const opus_val16 *num,
opus_val16 *y,
int N,
int ord,
opus_val16 *mem);
opus_val16 *mem,
int arch);
#if !defined(OVERRIDE_CELT_FIR)
#define celt_fir(x, num, y, N, ord, mem, arch) \
(celt_fir_c(x, num, y, N, ord, mem, arch))
#endif
void celt_iir(const opus_val32 *x,
const opus_val16 *den,
opus_val32 *y,
int N,
int ord,
opus_val16 *mem);
opus_val16 *mem,
int arch);
void _celt_autocorr(const opus_val16 *x, opus_val32 *ac, const opus_val16 *window, int overlap, int lag, int n);
int _celt_autocorr(const opus_val16 *x, opus_val32 *ac,
const opus_val16 *window, int overlap, int lag, int n, int arch);
#endif /* PLC_H */

68
Frameworks/Opus/Opus/opus/celt/cpu_support.h Normal file
View file

@ -0,0 +1,68 @@
/* Copyright (c) 2010 Xiph.Org Foundation
* Copyright (c) 2013 Parrot */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef CPU_SUPPORT_H
#define CPU_SUPPORT_H
#include "opus_types.h"
#include "opus_defines.h"
#if defined(OPUS_HAVE_RTCD) && \
(defined(OPUS_ARM_ASM) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR))
#include "arm/armcpu.h"
/* We currently support 4 ARM variants:
* arch[0] -> ARMv4
* arch[1] -> ARMv5E
* arch[2] -> ARMv6
* arch[3] -> NEON
*/
#define OPUS_ARCHMASK 3
#elif (defined(OPUS_X86_MAY_HAVE_SSE) && !defined(OPUS_X86_PRESUME_SSE)) || \
(defined(OPUS_X86_MAY_HAVE_SSE2) && !defined(OPUS_X86_PRESUME_SSE2)) || \
(defined(OPUS_X86_MAY_HAVE_SSE4_1) && !defined(OPUS_X86_PRESUME_SSE4_1))
#include "x86/x86cpu.h"
/* We currently support 4 x86 variants:
* arch[0] -> non-sse
* arch[1] -> sse
* arch[2] -> sse2
* arch[3] -> sse4.1
*/
#define OPUS_ARCHMASK 3
int opus_select_arch(void);
#else
#define OPUS_ARCHMASK 0
static OPUS_INLINE int opus_select_arch(void)
{
return 0;
}
#endif
#endif

650
Frameworks/Opus/Opus/opus/celt/cwrs.c
View file

@ -71,64 +71,6 @@ int log2_frac(opus_uint32 val, int frac)
}
#endif
#ifndef SMALL_FOOTPRINT
#define MASK32 (0xFFFFFFFF)
/*INV_TABLE[i] holds the multiplicative inverse of (2*i+1) mod 2**32.*/
static const opus_uint32 INV_TABLE[53]={
0x00000001,0xAAAAAAAB,0xCCCCCCCD,0xB6DB6DB7,
0x38E38E39,0xBA2E8BA3,0xC4EC4EC5,0xEEEEEEEF,
0xF0F0F0F1,0x286BCA1B,0x3CF3CF3D,0xE9BD37A7,
0xC28F5C29,0x684BDA13,0x4F72C235,0xBDEF7BDF,
0x3E0F83E1,0x8AF8AF8B,0x914C1BAD,0x96F96F97,
0xC18F9C19,0x2FA0BE83,0xA4FA4FA5,0x677D46CF,
0x1A1F58D1,0xFAFAFAFB,0x8C13521D,0x586FB587,
0xB823EE09,0xA08AD8F3,0xC10C9715,0xBEFBEFBF,
0xC0FC0FC1,0x07A44C6B,0xA33F128D,0xE327A977,
0xC7E3F1F9,0x962FC963,0x3F2B3885,0x613716AF,
0x781948B1,0x2B2E43DB,0xFCFCFCFD,0x6FD0EB67,
0xFA3F47E9,0xD2FD2FD3,0x3F4FD3F5,0xD4E25B9F,
0x5F02A3A1,0xBF5A814B,0x7C32B16D,0xD3431B57,
0xD8FD8FD9,
};
/*Computes (_a*_b-_c)/(2*_d+1) when the quotient is known to be exact.
_a, _b, _c, and _d may be arbitrary so long as the arbitrary precision result
fits in 32 bits, but currently the table for multiplicative inverses is only
valid for _d<=52.*/
static inline opus_uint32 imusdiv32odd(opus_uint32 _a,opus_uint32 _b,
opus_uint32 _c,int _d){
celt_assert(_d<=52);
return (_a*_b-_c)*INV_TABLE[_d]&MASK32;
}
/*Computes (_a*_b-_c)/_d when the quotient is known to be exact.
_d does not actually have to be even, but imusdiv32odd will be faster when
it's odd, so you should use that instead.
_a and _d are assumed to be small (e.g., _a*_d fits in 32 bits; currently the
table for multiplicative inverses is only valid for _d<=54).
_b and _c may be arbitrary so long as the arbitrary precision reuslt fits in
32 bits.*/
static inline opus_uint32 imusdiv32even(opus_uint32 _a,opus_uint32 _b,
opus_uint32 _c,int _d){
opus_uint32 inv;
int mask;
int shift;
int one;
celt_assert(_d>0);
celt_assert(_d<=54);
shift=EC_ILOG(_d^(_d-1));
inv=INV_TABLE[(_d-1)>>shift];
shift--;
one=1<<shift;
mask=one-1;
return (_a*(_b>>shift)-(_c>>shift)+
((_a*(_b&mask)+one-(_c&mask))>>shift)-1)*inv&MASK32;
}
#endif /* SMALL_FOOTPRINT */
/*Although derived separately, the pulse vector coding scheme is equivalent to
a Pyramid Vector Quantizer \cite{Fis86}.
Some additional notes about an early version appear at
@ -248,51 +190,362 @@ static inline opus_uint32 imusdiv32even(opus_uint32 _a,opus_uint32 _b,
year=1986
}*/
#ifndef SMALL_FOOTPRINT
/*Compute U(2,_k).
Note that this may be called with _k=32768 (maxK[2]+1).*/
static inline unsigned ucwrs2(unsigned _k){
celt_assert(_k>0);
return _k+(_k-1);
#if !defined(SMALL_FOOTPRINT)
/*U(N,K) = U(K,N) := N>0?K>0?U(N-1,K)+U(N,K-1)+U(N-1,K-1):0:K>0?1:0*/
# define CELT_PVQ_U(_n,_k) (CELT_PVQ_U_ROW[IMIN(_n,_k)][IMAX(_n,_k)])
/*V(N,K) := U(N,K)+U(N,K+1) = the number of PVQ codewords for a band of size N
with K pulses allocated to it.*/
# define CELT_PVQ_V(_n,_k) (CELT_PVQ_U(_n,_k)+CELT_PVQ_U(_n,(_k)+1))
/*For each V(N,K) supported, we will access element U(min(N,K+1),max(N,K+1)).
Thus, the number of entries in row I is the larger of the maximum number of
pulses we will ever allocate for a given N=I (K=128, or however many fit in
32 bits, whichever is smaller), plus one, and the maximum N for which
K=I-1 pulses fit in 32 bits.
The largest band size in an Opus Custom mode is 208.
Otherwise, we can limit things to the set of N which can be achieved by
splitting a band from a standard Opus mode: 176, 144, 96, 88, 72, 64, 48,
44, 36, 32, 24, 22, 18, 16, 8, 4, 2).*/
#if defined(CUSTOM_MODES)
static const opus_uint32 CELT_PVQ_U_DATA[1488]={
#else
static const opus_uint32 CELT_PVQ_U_DATA[1272]={
#endif
/*N=0, K=0...176:*/
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
#if defined(CUSTOM_MODES)
/*...208:*/
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0,
#endif
/*N=1, K=1...176:*/
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
#if defined(CUSTOM_MODES)
/*...208:*/
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1,
#endif
/*N=2, K=2...176:*/
3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31, 33, 35, 37, 39, 41,
43, 45, 47, 49, 51, 53, 55, 57, 59, 61, 63, 65, 67, 69, 71, 73, 75, 77, 79,
81, 83, 85, 87, 89, 91, 93, 95, 97, 99, 101, 103, 105, 107, 109, 111, 113,
115, 117, 119, 121, 123, 125, 127, 129, 131, 133, 135, 137, 139, 141, 143,
145, 147, 149, 151, 153, 155, 157, 159, 161, 163, 165, 167, 169, 171, 173,
175, 177, 179, 181, 183, 185, 187, 189, 191, 193, 195, 197, 199, 201, 203,
205, 207, 209, 211, 213, 215, 217, 219, 221, 223, 225, 227, 229, 231, 233,
235, 237, 239, 241, 243, 245, 247, 249, 251, 253, 255, 257, 259, 261, 263,
265, 267, 269, 271, 273, 275, 277, 279, 281, 283, 285, 287, 289, 291, 293,
295, 297, 299, 301, 303, 305, 307, 309, 311, 313, 315, 317, 319, 321, 323,
325, 327, 329, 331, 333, 335, 337, 339, 341, 343, 345, 347, 349, 351,
#if defined(CUSTOM_MODES)
/*...208:*/
353, 355, 357, 359, 361, 363, 365, 367, 369, 371, 373, 375, 377, 379, 381,
383, 385, 387, 389, 391, 393, 395, 397, 399, 401, 403, 405, 407, 409, 411,
413, 415,
#endif
/*N=3, K=3...176:*/
13, 25, 41, 61, 85, 113, 145, 181, 221, 265, 313, 365, 421, 481, 545, 613,
685, 761, 841, 925, 1013, 1105, 1201, 1301, 1405, 1513, 1625, 1741, 1861,
1985, 2113, 2245, 2381, 2521, 2665, 2813, 2965, 3121, 3281, 3445, 3613, 3785,
3961, 4141, 4325, 4513, 4705, 4901, 5101, 5305, 5513, 5725, 5941, 6161, 6385,
6613, 6845, 7081, 7321, 7565, 7813, 8065, 8321, 8581, 8845, 9113, 9385, 9661,
9941, 10225, 10513, 10805, 11101, 11401, 11705, 12013, 12325, 12641, 12961,
13285, 13613, 13945, 14281, 14621, 14965, 15313, 15665, 16021, 16381, 16745,
17113, 17485, 17861, 18241, 18625, 19013, 19405, 19801, 20201, 20605, 21013,
21425, 21841, 22261, 22685, 23113, 23545, 23981, 24421, 24865, 25313, 25765,
26221, 26681, 27145, 27613, 28085, 28561, 29041, 29525, 30013, 30505, 31001,
31501, 32005, 32513, 33025, 33541, 34061, 34585, 35113, 35645, 36181, 36721,
37265, 37813, 38365, 38921, 39481, 40045, 40613, 41185, 41761, 42341, 42925,
43513, 44105, 44701, 45301, 45905, 46513, 47125, 47741, 48361, 48985, 49613,
50245, 50881, 51521, 52165, 52813, 53465, 54121, 54781, 55445, 56113, 56785,
57461, 58141, 58825, 59513, 60205, 60901, 61601,
#if defined(CUSTOM_MODES)
/*...208:*/
62305, 63013, 63725, 64441, 65161, 65885, 66613, 67345, 68081, 68821, 69565,
70313, 71065, 71821, 72581, 73345, 74113, 74885, 75661, 76441, 77225, 78013,
78805, 79601, 80401, 81205, 82013, 82825, 83641, 84461, 85285, 86113,
#endif
/*N=4, K=4...176:*/
63, 129, 231, 377, 575, 833, 1159, 1561, 2047, 2625, 3303, 4089, 4991, 6017,
7175, 8473, 9919, 11521, 13287, 15225, 17343, 19649, 22151, 24857, 27775,
30913, 34279, 37881, 41727, 45825, 50183, 54809, 59711, 64897, 70375, 76153,
82239, 88641, 95367, 102425, 109823, 117569, 125671, 134137, 142975, 152193,
161799, 171801, 182207, 193025, 204263, 215929, 228031, 240577, 253575,
267033, 280959, 295361, 310247, 325625, 341503, 357889, 374791, 392217,
410175, 428673, 447719, 467321, 487487, 508225, 529543, 551449, 573951,
597057, 620775, 645113, 670079, 695681, 721927, 748825, 776383, 804609,
833511, 863097, 893375, 924353, 956039, 988441, 1021567, 1055425, 1090023,
1125369, 1161471, 1198337, 1235975, 1274393, 1313599, 1353601, 1394407,
1436025, 1478463, 1521729, 1565831, 1610777, 1656575, 1703233, 1750759,
1799161, 1848447, 1898625, 1949703, 2001689, 2054591, 2108417, 2163175,
2218873, 2275519, 2333121, 2391687, 2451225, 2511743, 2573249, 2635751,
2699257, 2763775, 2829313, 2895879, 2963481, 3032127, 3101825, 3172583,
3244409, 3317311, 3391297, 3466375, 3542553, 3619839, 3698241, 3777767,
3858425, 3940223, 4023169, 4107271, 4192537, 4278975, 4366593, 4455399,
4545401, 4636607, 4729025, 4822663, 4917529, 5013631, 5110977, 5209575,
5309433, 5410559, 5512961, 5616647, 5721625, 5827903, 5935489, 6044391,
6154617, 6266175, 6379073, 6493319, 6608921, 6725887, 6844225, 6963943,
7085049, 7207551,
#if defined(CUSTOM_MODES)
/*...208:*/
7331457, 7456775, 7583513, 7711679, 7841281, 7972327, 8104825, 8238783,
8374209, 8511111, 8649497, 8789375, 8930753, 9073639, 9218041, 9363967,
9511425, 9660423, 9810969, 9963071, 10116737, 10271975, 10428793, 10587199,
10747201, 10908807, 11072025, 11236863, 11403329, 11571431, 11741177,
11912575,
#endif
/*N=5, K=5...176:*/
321, 681, 1289, 2241, 3649, 5641, 8361, 11969, 16641, 22569, 29961, 39041,
50049, 63241, 78889, 97281, 118721, 143529, 172041, 204609, 241601, 283401,
330409, 383041, 441729, 506921, 579081, 658689, 746241, 842249, 947241,
1061761, 1186369, 1321641, 1468169, 1626561, 1797441, 1981449, 2179241,
2391489, 2618881, 2862121, 3121929, 3399041, 3694209, 4008201, 4341801,
4695809, 5071041, 5468329, 5888521, 6332481, 6801089, 7295241, 7815849,
8363841, 8940161, 9545769, 10181641, 10848769, 11548161, 12280841, 13047849,
13850241, 14689089, 15565481, 16480521, 17435329, 18431041, 19468809,
20549801, 21675201, 22846209, 24064041, 25329929, 26645121, 28010881,
29428489, 30899241, 32424449, 34005441, 35643561, 37340169, 39096641,
40914369, 42794761, 44739241, 46749249, 48826241, 50971689, 53187081,
55473921, 57833729, 60268041, 62778409, 65366401, 68033601, 70781609,
73612041, 76526529, 79526721, 82614281, 85790889, 89058241, 92418049,
95872041, 99421961, 103069569, 106816641, 110664969, 114616361, 118672641,
122835649, 127107241, 131489289, 135983681, 140592321, 145317129, 150160041,
155123009, 160208001, 165417001, 170752009, 176215041, 181808129, 187533321,
193392681, 199388289, 205522241, 211796649, 218213641, 224775361, 231483969,
238341641, 245350569, 252512961, 259831041, 267307049, 274943241, 282741889,
290705281, 298835721, 307135529, 315607041, 324252609, 333074601, 342075401,
351257409, 360623041, 370174729, 379914921, 389846081, 399970689, 410291241,
420810249, 431530241, 442453761, 453583369, 464921641, 476471169, 488234561,
500214441, 512413449, 524834241, 537479489, 550351881, 563454121, 576788929,
590359041, 604167209, 618216201, 632508801,
#if defined(CUSTOM_MODES)
/*...208:*/
647047809, 661836041, 676876329, 692171521, 707724481, 723538089, 739615241,
755958849, 772571841, 789457161, 806617769, 824056641, 841776769, 859781161,
878072841, 896654849, 915530241, 934702089, 954173481, 973947521, 994027329,
1014416041, 1035116809, 1056132801, 1077467201, 1099123209, 1121104041,
1143412929, 1166053121, 1189027881, 1212340489, 1235994241,
#endif
/*N=6, K=6...96:*/
1683, 3653, 7183, 13073, 22363, 36365, 56695, 85305, 124515, 177045, 246047,
335137, 448427, 590557, 766727, 982729, 1244979, 1560549, 1937199, 2383409,
2908411, 3522221, 4235671, 5060441, 6009091, 7095093, 8332863, 9737793,
11326283, 13115773, 15124775, 17372905, 19880915, 22670725, 25765455,
29189457, 32968347, 37129037, 41699767, 46710137, 52191139, 58175189,
64696159, 71789409, 79491819, 87841821, 96879431, 106646281, 117185651,
128542501, 140763503, 153897073, 167993403, 183104493, 199284183, 216588185,
235074115, 254801525, 275831935, 298228865, 322057867, 347386557, 374284647,
402823977, 433078547, 465124549, 499040399, 534906769, 572806619, 612825229,
655050231, 699571641, 746481891, 795875861, 847850911, 902506913, 959946283,
1020274013, 1083597703, 1150027593, 1219676595, 1292660325, 1369097135,
1449108145, 1532817275, 1620351277, 1711839767, 1807415257, 1907213187,
2011371957, 2120032959,
#if defined(CUSTOM_MODES)
/*...109:*/
2233340609U, 2351442379U, 2474488829U, 2602633639U, 2736033641U, 2874848851U,
3019242501U, 3169381071U, 3325434321U, 3487575323U, 3655980493U, 3830829623U,
4012305913U,
#endif
/*N=7, K=7...54*/
8989, 19825, 40081, 75517, 134245, 227305, 369305, 579125, 880685, 1303777,
1884961, 2668525, 3707509, 5064793, 6814249, 9041957, 11847485, 15345233,
19665841, 24957661, 31388293, 39146185, 48442297, 59511829, 72616013,
88043969, 106114625, 127178701, 151620757, 179861305, 212358985, 249612805,
292164445, 340600625, 395555537, 457713341, 527810725, 606639529, 695049433,
793950709, 904317037, 1027188385, 1163673953, 1314955181, 1482288821,
1667010073, 1870535785, 2094367717,
#if defined(CUSTOM_MODES)
/*...60:*/
2340095869U, 2609401873U, 2904062449U, 3225952925U, 3577050821U, 3959439497U,
#endif
/*N=8, K=8...37*/
48639, 108545, 224143, 433905, 795455, 1392065, 2340495, 3800305, 5984767,
9173505, 13726991, 20103025, 28875327, 40754369, 56610575, 77500017,
104692735, 139703809, 184327311, 240673265, 311207743, 398796225, 506750351,
638878193, 799538175, 993696769, 1226990095, 1505789553, 1837271615,
2229491905U,
#if defined(CUSTOM_MODES)
/*...40:*/
2691463695U, 3233240945U, 3866006015U,
#endif
/*N=9, K=9...28:*/
265729, 598417, 1256465, 2485825, 4673345, 8405905, 14546705, 24331777,
39490049, 62390545, 96220561, 145198913, 214828609, 312193553, 446304145,
628496897, 872893441, 1196924561, 1621925137, 2173806145U,
#if defined(CUSTOM_MODES)
/*...29:*/
2883810113U,
#endif
/*N=10, K=10...24:*/
1462563, 3317445, 7059735, 14218905, 27298155, 50250765, 89129247, 152951073,
254831667, 413442773, 654862247, 1014889769, 1541911931, 2300409629U,
3375210671U,
/*N=11, K=11...19:*/
8097453, 18474633, 39753273, 81270333, 158819253, 298199265, 540279585,
948062325, 1616336765,
#if defined(CUSTOM_MODES)
/*...20:*/
2684641785U,
#endif
/*N=12, K=12...18:*/
45046719, 103274625, 224298231, 464387817, 921406335, 1759885185,
3248227095U,
/*N=13, K=13...16:*/
251595969, 579168825, 1267854873, 2653649025U,
/*N=14, K=14:*/
1409933619
};
#if defined(CUSTOM_MODES)
static const opus_uint32 *const CELT_PVQ_U_ROW[15]={
CELT_PVQ_U_DATA+ 0,CELT_PVQ_U_DATA+ 208,CELT_PVQ_U_DATA+ 415,
CELT_PVQ_U_DATA+ 621,CELT_PVQ_U_DATA+ 826,CELT_PVQ_U_DATA+1030,
CELT_PVQ_U_DATA+1233,CELT_PVQ_U_DATA+1336,CELT_PVQ_U_DATA+1389,
CELT_PVQ_U_DATA+1421,CELT_PVQ_U_DATA+1441,CELT_PVQ_U_DATA+1455,
CELT_PVQ_U_DATA+1464,CELT_PVQ_U_DATA+1470,CELT_PVQ_U_DATA+1473
};
#else
static const opus_uint32 *const CELT_PVQ_U_ROW[15]={
CELT_PVQ_U_DATA+ 0,CELT_PVQ_U_DATA+ 176,CELT_PVQ_U_DATA+ 351,
CELT_PVQ_U_DATA+ 525,CELT_PVQ_U_DATA+ 698,CELT_PVQ_U_DATA+ 870,
CELT_PVQ_U_DATA+1041,CELT_PVQ_U_DATA+1131,CELT_PVQ_U_DATA+1178,
CELT_PVQ_U_DATA+1207,CELT_PVQ_U_DATA+1226,CELT_PVQ_U_DATA+1240,
CELT_PVQ_U_DATA+1248,CELT_PVQ_U_DATA+1254,CELT_PVQ_U_DATA+1257
};
#endif
#if defined(CUSTOM_MODES)
void get_required_bits(opus_int16 *_bits,int _n,int _maxk,int _frac){
int k;
/*_maxk==0 => there's nothing to do.*/
celt_assert(_maxk>0);
_bits[0]=0;
for(k=1;k<=_maxk;k++)_bits[k]=log2_frac(CELT_PVQ_V(_n,k),_frac);
}
#endif
static opus_uint32 icwrs(int _n,const int *_y){
opus_uint32 i;
int j;
int k;
celt_assert(_n>=2);
j=_n-1;
i=_y[j]<0;
k=abs(_y[j]);
do{
j--;
i+=CELT_PVQ_U(_n-j,k);
k+=abs(_y[j]);
if(_y[j]<0)i+=CELT_PVQ_U(_n-j,k+1);
}
while(j>0);
return i;
}
/*Compute V(2,_k).*/
static inline opus_uint32 ncwrs2(int _k){
void encode_pulses(const int *_y,int _n,int _k,ec_enc *_enc){
celt_assert(_k>0);
return 4*(opus_uint32)_k;
ec_enc_uint(_enc,icwrs(_n,_y),CELT_PVQ_V(_n,_k));
}
/*Compute U(3,_k).
Note that this may be called with _k=32768 (maxK[3]+1).*/
static inline opus_uint32 ucwrs3(unsigned _k){
static opus_val32 cwrsi(int _n,int _k,opus_uint32 _i,int *_y){
opus_uint32 p;
int s;
int k0;
opus_int16 val;
opus_val32 yy=0;
celt_assert(_k>0);
return (2*(opus_uint32)_k-2)*_k+1;
celt_assert(_n>1);
while(_n>2){
opus_uint32 q;
/*Lots of pulses case:*/
if(_k>=_n){
const opus_uint32 *row;
row=CELT_PVQ_U_ROW[_n];
/*Are the pulses in this dimension negative?*/
p=row[_k+1];
s=-(_i>=p);
_i-=p&s;
/*Count how many pulses were placed in this dimension.*/
k0=_k;
q=row[_n];
if(q>_i){
celt_assert(p>q);
_k=_n;
do p=CELT_PVQ_U_ROW[--_k][_n];
while(p>_i);
}
else for(p=row[_k];p>_i;p=row[_k])_k--;
_i-=p;
val=(k0-_k+s)^s;
*_y++=val;
yy=MAC16_16(yy,val,val);
}
/*Lots of dimensions case:*/
else{
/*Are there any pulses in this dimension at all?*/
p=CELT_PVQ_U_ROW[_k][_n];
q=CELT_PVQ_U_ROW[_k+1][_n];
if(p<=_i&&_i<q){
_i-=p;
*_y++=0;
}
else{
/*Are the pulses in this dimension negative?*/
s=-(_i>=q);
_i-=q&s;
/*Count how many pulses were placed in this dimension.*/
k0=_k;
do p=CELT_PVQ_U_ROW[--_k][_n];
while(p>_i);
_i-=p;
val=(k0-_k+s)^s;
*_y++=val;
yy=MAC16_16(yy,val,val);
}
}
_n--;
}
/*_n==2*/
p=2*_k+1;
s=-(_i>=p);
_i-=p&s;
k0=_k;
_k=(_i+1)>>1;
if(_k)_i-=2*_k-1;
val=(k0-_k+s)^s;
*_y++=val;
yy=MAC16_16(yy,val,val);
/*_n==1*/
s=-(int)_i;
val=(_k+s)^s;
*_y=val;
yy=MAC16_16(yy,val,val);
return yy;
}
/*Compute V(3,_k).*/
static inline opus_uint32 ncwrs3(int _k){
celt_assert(_k>0);
return 2*(2*(unsigned)_k*(opus_uint32)_k+1);
opus_val32 decode_pulses(int *_y,int _n,int _k,ec_dec *_dec){
return cwrsi(_n,_k,ec_dec_uint(_dec,CELT_PVQ_V(_n,_k)),_y);
}
/*Compute U(4,_k).*/
static inline opus_uint32 ucwrs4(int _k){
celt_assert(_k>0);
return imusdiv32odd(2*_k,(2*_k-3)*(opus_uint32)_k+4,3,1);
}
/*Compute V(4,_k).*/
static inline opus_uint32 ncwrs4(int _k){
celt_assert(_k>0);
return ((_k*(opus_uint32)_k+2)*_k)/3<<3;
}
#endif /* SMALL_FOOTPRINT */
#else /* SMALL_FOOTPRINT */
/*Computes the next row/column of any recurrence that obeys the relation
u[i][j]=u[i-1][j]+u[i][j-1]+u[i-1][j-1].
_ui0 is the base case for the new row/column.*/
static inline void unext(opus_uint32 *_ui,unsigned _len,opus_uint32 _ui0){
static OPUS_INLINE void unext(opus_uint32 *_ui,unsigned _len,opus_uint32 _ui0){
opus_uint32 ui1;
unsigned j;
/*This do-while will overrun the array if we don't have storage for at least
@ -308,7 +561,7 @@ static inline void unext(opus_uint32 *_ui,unsigned _len,opus_uint32 _ui0){
/*Computes the previous row/column of any recurrence that obeys the relation
u[i-1][j]=u[i][j]-u[i][j-1]-u[i-1][j-1].
_ui0 is the base case for the new row/column.*/
static inline void uprev(opus_uint32 *_ui,unsigned _n,opus_uint32 _ui0){
static OPUS_INLINE void uprev(opus_uint32 *_ui,unsigned _n,opus_uint32 _ui0){
opus_uint32 ui1;
unsigned j;
/*This do-while will overrun the array if we don't have storage for at least
@ -332,12 +585,6 @@ static opus_uint32 ncwrs_urow(unsigned _n,unsigned _k,opus_uint32 *_u){
celt_assert(len>=3);
_u[0]=0;
_u[1]=um2=1;
#ifndef SMALL_FOOTPRINT
/*_k>52 doesn't work in the false branch due to the limits of INV_TABLE,
but _k isn't tested here because k<=52 for n=7*/
if(_n<=6)
#endif
{
/*If _n==0, _u[0] should be 1 and the rest should be 0.*/
/*If _n==1, _u[i] should be 1 for i>1.*/
celt_assert(_n>=2);
@ -347,117 +594,18 @@ static opus_uint32 ncwrs_urow(unsigned _n,unsigned _k,opus_uint32 *_u){
do _u[k]=(k<<1)-1;
while(++k<len);
for(k=2;k<_n;k++)unext(_u+1,_k+1,1);
}
#ifndef SMALL_FOOTPRINT
else{
opus_uint32 um1;
opus_uint32 n2m1;
_u[2]=n2m1=um1=(_n<<1)-1;
for(k=3;k<len;k++){
/*U(N,K) = ((2*N-1)*U(N,K-1)-U(N,K-2))/(K-1) + U(N,K-2)*/
_u[k]=um2=imusdiv32even(n2m1,um1,um2,k-1)+um2;
if(++k>=len)break;
_u[k]=um1=imusdiv32odd(n2m1,um2,um1,(k-1)>>1)+um1;
}
}
#endif /* SMALL_FOOTPRINT */
return _u[_k]+_u[_k+1];
}
#ifndef SMALL_FOOTPRINT
/*Returns the _i'th combination of _k elements (at most 32767) chosen from a
set of size 1 with associated sign bits.
_y: Returns the vector of pulses.*/
static inline void cwrsi1(int _k,opus_uint32 _i,int *_y){
int s;
s=-(int)_i;
_y[0]=(_k+s)^s;
}
/*Returns the _i'th combination of _k elements (at most 32767) chosen from a
set of size 2 with associated sign bits.
_y: Returns the vector of pulses.*/
static inline void cwrsi2(int _k,opus_uint32 _i,int *_y){
opus_uint32 p;
int s;
int yj;
p=ucwrs2(_k+1U);
s=-(_i>=p);
_i-=p&s;
yj=_k;
_k=(_i+1)>>1;
p=_k?ucwrs2(_k):0;
_i-=p;
yj-=_k;
_y[0]=(yj+s)^s;
cwrsi1(_k,_i,_y+1);
}
/*Returns the _i'th combination of _k elements (at most 32767) chosen from a
set of size 3 with associated sign bits.
_y: Returns the vector of pulses.*/
static void cwrsi3(int _k,opus_uint32 _i,int *_y){
opus_uint32 p;
int s;
int yj;
p=ucwrs3(_k+1U);
s=-(_i>=p);
_i-=p&s;
yj=_k;
/*Finds the maximum _k such that ucwrs3(_k)<=_i (tested for all
_i<2147418113=U(3,32768)).*/
_k=_i>0?(isqrt32(2*_i-1)+1)>>1:0;
p=_k?ucwrs3(_k):0;
_i-=p;
yj-=_k;
_y[0]=(yj+s)^s;
cwrsi2(_k,_i,_y+1);
}
/*Returns the _i'th combination of _k elements (at most 1172) chosen from a set
of size 4 with associated sign bits.
_y: Returns the vector of pulses.*/
static void cwrsi4(int _k,opus_uint32 _i,int *_y){
opus_uint32 p;
int s;
int yj;
int kl;
int kr;
p=ucwrs4(_k+1);
s=-(_i>=p);
_i-=p&s;
yj=_k;
/*We could solve a cubic for k here, but the form of the direct solution does
not lend itself well to exact integer arithmetic.
Instead we do a binary search on U(4,K).*/
kl=0;
kr=_k;
for(;;){
_k=(kl+kr)>>1;
p=_k?ucwrs4(_k):0;
if(p<_i){
if(_k>=kr)break;
kl=_k+1;
}
else if(p>_i)kr=_k-1;
else break;
}
_i-=p;
yj-=_k;
_y[0]=(yj+s)^s;
cwrsi3(_k,_i,_y+1);
}
#endif /* SMALL_FOOTPRINT */
/*Returns the _i'th combination of _k elements chosen from a set of size _n
with associated sign bits.
_y: Returns the vector of pulses.
_u: Must contain entries [0..._k+1] of row _n of U() on input.
Its contents will be destructively modified.*/
static void cwrsi(int _n,int _k,opus_uint32 _i,int *_y,opus_uint32 *_u){
static opus_val32 cwrsi(int _n,int _k,opus_uint32 _i,int *_y,opus_uint32 *_u){
int j;
opus_int16 val;
opus_val32 yy=0;
celt_assert(_n>0);
j=0;
do{
@ -472,75 +620,29 @@ static void cwrsi(int _n,int _k,opus_uint32 _i,int *_y,opus_uint32 *_u){
while(p>_i)p=_u[--_k];
_i-=p;
yj-=_k;
_y[j]=(yj+s)^s;
val=(yj+s)^s;
_y[j]=val;
yy=MAC16_16(yy,val,val);
uprev(_u,_k+2,0);
}
while(++j<_n);
return yy;
}
/*Returns the index of the given combination of K elements chosen from a set
of size 1 with associated sign bits.
_y: The vector of pulses, whose sum of absolute values is K.
_k: Returns K.*/
static inline opus_uint32 icwrs1(const int *_y,int *_k){
static OPUS_INLINE opus_uint32 icwrs1(const int *_y,int *_k){
*_k=abs(_y[0]);
return _y[0]<0;
}
#ifndef SMALL_FOOTPRINT
/*Returns the index of the given combination of K elements chosen from a set
of size 2 with associated sign bits.
_y: The vector of pulses, whose sum of absolute values is K.
_k: Returns K.*/
static inline opus_uint32 icwrs2(const int *_y,int *_k){
opus_uint32 i;
int k;
i=icwrs1(_y+1,&k);
i+=k?ucwrs2(k):0;
k+=abs(_y[0]);
if(_y[0]<0)i+=ucwrs2(k+1U);
*_k=k;
return i;
}
/*Returns the index of the given combination of K elements chosen from a set
of size 3 with associated sign bits.
_y: The vector of pulses, whose sum of absolute values is K.
_k: Returns K.*/
static inline opus_uint32 icwrs3(const int *_y,int *_k){
opus_uint32 i;
int k;
i=icwrs2(_y+1,&k);
i+=k?ucwrs3(k):0;
k+=abs(_y[0]);
if(_y[0]<0)i+=ucwrs3(k+1U);
*_k=k;
return i;
}
/*Returns the index of the given combination of K elements chosen from a set
of size 4 with associated sign bits.
_y: The vector of pulses, whose sum of absolute values is K.
_k: Returns K.*/
static inline opus_uint32 icwrs4(const int *_y,int *_k){
opus_uint32 i;
int k;
i=icwrs3(_y+1,&k);
i+=k?ucwrs4(k):0;
k+=abs(_y[0]);
if(_y[0]<0)i+=ucwrs4(k+1);
*_k=k;
return i;
}
#endif /* SMALL_FOOTPRINT */
/*Returns the index of the given combination of K elements chosen from a set
of size _n with associated sign bits.
_y: The vector of pulses, whose sum of absolute values must be _k.
_nc: Returns V(_n,_k).*/
static inline opus_uint32 icwrs(int _n,int _k,opus_uint32 *_nc,const int *_y,
static OPUS_INLINE opus_uint32 icwrs(int _n,int _k,opus_uint32 *_nc,const int *_y,
opus_uint32 *_u){
opus_uint32 i;
int j;
@ -589,57 +691,25 @@ void get_required_bits(opus_int16 *_bits,int _n,int _maxk,int _frac){
void encode_pulses(const int *_y,int _n,int _k,ec_enc *_enc){
opus_uint32 i;
celt_assert(_k>0);
#ifndef SMALL_FOOTPRINT
switch(_n){
case 2:{
i=icwrs2(_y,&_k);
ec_enc_uint(_enc,i,ncwrs2(_k));
}break;
case 3:{
i=icwrs3(_y,&_k);
ec_enc_uint(_enc,i,ncwrs3(_k));
}break;
case 4:{
i=icwrs4(_y,&_k);
ec_enc_uint(_enc,i,ncwrs4(_k));
}break;
default:
{
#endif
VARDECL(opus_uint32,u);
opus_uint32 nc;
SAVE_STACK;
celt_assert(_k>0);
ALLOC(u,_k+2U,opus_uint32);
i=icwrs(_n,_k,&nc,_y,u);
ec_enc_uint(_enc,i,nc);
RESTORE_STACK;
#ifndef SMALL_FOOTPRINT
}
break;
}
#endif
}
void decode_pulses(int *_y,int _n,int _k,ec_dec *_dec)
{
celt_assert(_k>0);
#ifndef SMALL_FOOTPRINT
switch(_n){
case 2:cwrsi2(_k,ec_dec_uint(_dec,ncwrs2(_k)),_y);break;
case 3:cwrsi3(_k,ec_dec_uint(_dec,ncwrs3(_k)),_y);break;
case 4:cwrsi4(_k,ec_dec_uint(_dec,ncwrs4(_k)),_y);break;
default:
{
#endif
opus_val32 decode_pulses(int *_y,int _n,int _k,ec_dec *_dec){
VARDECL(opus_uint32,u);
int ret;
SAVE_STACK;
celt_assert(_k>0);
ALLOC(u,_k+2U,opus_uint32);
cwrsi(_n,_k,ec_dec_uint(_dec,ncwrs_urow(_n,_k,u)),_y,u);
ret = cwrsi(_n,_k,ec_dec_uint(_dec,ncwrs_urow(_n,_k,u)),_y,u);
RESTORE_STACK;
#ifndef SMALL_FOOTPRINT
}
break;
}
#endif
return ret;
}
#endif /* SMALL_FOOTPRINT */

2
Frameworks/Opus/Opus/opus/celt/cwrs.h
View file

@ -43,6 +43,6 @@ void get_required_bits(opus_int16 *bits, int N, int K, int frac);
void encode_pulses(const int *_y, int N, int K, ec_enc *enc);
void decode_pulses(int *_y, int N, int K, ec_dec *dec);
opus_val32 decode_pulses(int *_y, int N, int K, ec_dec *dec);
#endif /* CWRS_H */

2
Frameworks/Opus/Opus/opus/celt/ecintrin.h
View file

@ -33,7 +33,7 @@
#if !defined(_ecintrin_H)
# define _ecintrin_H (1)
/*Some specific platforms may have optimized intrinsic or inline assembly
/*Some specific platforms may have optimized intrinsic or OPUS_INLINE assembly
versions of these functions which can substantially improve performance.
We define macros for them to allow easy incorporation of these non-ANSI
features.*/

60
Frameworks/Opus/Opus/opus/celt/entcode.c
View file

@ -62,6 +62,27 @@ int ec_ilog(opus_uint32 _v){
}
#endif
#if 1
/* This is a faster version of ec_tell_frac() that takes advantage
of the low (1/8 bit) resolution to use just a linear function
followed by a lookup to determine the exact transition thresholds. */
opus_uint32 ec_tell_frac(ec_ctx *_this){
static const unsigned correction[8] =
{35733, 38967, 42495, 46340,
50535, 55109, 60097, 65535};
opus_uint32 nbits;
opus_uint32 r;
int l;
unsigned b;
nbits=_this->nbits_total<<BITRES;
l=EC_ILOG(_this->rng);
r=_this->rng>>(l-16);
b = (r>>12)-8;
b += r>correction[b];
l = (l<<3)+b;
return nbits-l;
}
#else
opus_uint32 ec_tell_frac(ec_ctx *_this){
opus_uint32 nbits;
opus_uint32 r;
@ -91,3 +112,42 @@ opus_uint32 ec_tell_frac(ec_ctx *_this){
}
return nbits-l;
}
#endif
#ifdef USE_SMALL_DIV_TABLE
/* Result of 2^32/(2*i+1), except for i=0. */
const opus_uint32 SMALL_DIV_TABLE[129] = {
0xFFFFFFFF, 0x55555555, 0x33333333, 0x24924924,
0x1C71C71C, 0x1745D174, 0x13B13B13, 0x11111111,
0x0F0F0F0F, 0x0D79435E, 0x0C30C30C, 0x0B21642C,
0x0A3D70A3, 0x097B425E, 0x08D3DCB0, 0x08421084,
0x07C1F07C, 0x07507507, 0x06EB3E45, 0x06906906,
0x063E7063, 0x05F417D0, 0x05B05B05, 0x0572620A,
0x05397829, 0x05050505, 0x04D4873E, 0x04A7904A,
0x047DC11F, 0x0456C797, 0x04325C53, 0x04104104,
0x03F03F03, 0x03D22635, 0x03B5CC0E, 0x039B0AD1,
0x0381C0E0, 0x0369D036, 0x03531DEC, 0x033D91D2,
0x0329161F, 0x03159721, 0x03030303, 0x02F14990,
0x02E05C0B, 0x02D02D02, 0x02C0B02C, 0x02B1DA46,
0x02A3A0FD, 0x0295FAD4, 0x0288DF0C, 0x027C4597,
0x02702702, 0x02647C69, 0x02593F69, 0x024E6A17,
0x0243F6F0, 0x0239E0D5, 0x02302302, 0x0226B902,
0x021D9EAD, 0x0214D021, 0x020C49BA, 0x02040810,
0x01FC07F0, 0x01F44659, 0x01ECC07B, 0x01E573AC,
0x01DE5D6E, 0x01D77B65, 0x01D0CB58, 0x01CA4B30,
0x01C3F8F0, 0x01BDD2B8, 0x01B7D6C3, 0x01B20364,
0x01AC5701, 0x01A6D01A, 0x01A16D3F, 0x019C2D14,
0x01970E4F, 0x01920FB4, 0x018D3018, 0x01886E5F,
0x0183C977, 0x017F405F, 0x017AD220, 0x01767DCE,
0x01724287, 0x016E1F76, 0x016A13CD, 0x01661EC6,
0x01623FA7, 0x015E75BB, 0x015AC056, 0x01571ED3,
0x01539094, 0x01501501, 0x014CAB88, 0x0149539E,
0x01460CBC, 0x0142D662, 0x013FB013, 0x013C995A,
0x013991C2, 0x013698DF, 0x0133AE45, 0x0130D190,
0x012E025C, 0x012B404A, 0x01288B01, 0x0125E227,
0x01234567, 0x0120B470, 0x011E2EF3, 0x011BB4A4,
0x01194538, 0x0116E068, 0x011485F0, 0x0112358E,
0x010FEF01, 0x010DB20A, 0x010B7E6E, 0x010953F3,
0x01073260, 0x0105197F, 0x0103091B, 0x01010101
};
#endif

44
Frameworks/Opus/Opus/opus/celt/entcode.h
View file

@ -26,6 +26,7 @@
*/
#include "opus_types.h"
#include "opus_defines.h"
#if !defined(_entcode_H)
# define _entcode_H (1)
@ -33,6 +34,12 @@
# include <stddef.h>
# include "ecintrin.h"
extern const opus_uint32 SMALL_DIV_TABLE[129];
#ifdef OPUS_ARM_ASM
#define USE_SMALL_DIV_TABLE
#endif
/*OPT: ec_window must be at least 32 bits, but if you have fast arithmetic on a
larger type, you can speed up the decoder by using it here.*/
typedef opus_uint32 ec_window;
@ -83,15 +90,15 @@ struct ec_ctx{
int error;
};
static inline opus_uint32 ec_range_bytes(ec_ctx *_this){
static OPUS_INLINE opus_uint32 ec_range_bytes(ec_ctx *_this){
return _this->offs;
}
static inline unsigned char *ec_get_buffer(ec_ctx *_this){
static OPUS_INLINE unsigned char *ec_get_buffer(ec_ctx *_this){
return _this->buf;
}
static inline int ec_get_error(ec_ctx *_this){
static OPUS_INLINE int ec_get_error(ec_ctx *_this){
return _this->error;
}
@ -101,7 +108,7 @@ static inline int ec_get_error(ec_ctx *_this){
Return: The number of bits.
This will always be slightly larger than the exact value (e.g., all
rounding error is in the positive direction).*/
static inline int ec_tell(ec_ctx *_this){
static OPUS_INLINE int ec_tell(ec_ctx *_this){
return _this->nbits_total-EC_ILOG(_this->rng);
}
@ -113,4 +120,33 @@ static inline int ec_tell(ec_ctx *_this){
rounding error is in the positive direction).*/
opus_uint32 ec_tell_frac(ec_ctx *_this);
/* Tested exhaustively for all n and for 1<=d<=256 */
static OPUS_INLINE opus_uint32 celt_udiv(opus_uint32 n, opus_uint32 d) {
celt_assert(d>0);
#ifdef USE_SMALL_DIV_TABLE
if (d>256)
return n/d;
else {
opus_uint32 t, q;
t = EC_ILOG(d&-d);
q = (opus_uint64)SMALL_DIV_TABLE[d>>t]*(n>>(t-1))>>32;
return q+(n-q*d >= d);
}
#else
return n/d;
#endif
}
static OPUS_INLINE opus_int32 celt_sudiv(opus_int32 n, opus_int32 d) {
celt_assert(d>0);
#ifdef USE_SMALL_DIV_TABLE
if (n<0)
return -(opus_int32)celt_udiv(-n, d);
else
return celt_udiv(n, d);
#else
return n/d;
#endif
}
#endif

4
Frameworks/Opus/Opus/opus/celt/entdec.c
View file

@ -85,7 +85,7 @@
number=3,
pages="256--294",
month=Jul,
URL="http://www.stanford.edu/class/ee398/handouts/papers/Moffat98ArithmCoding.pdf"
URL="http://www.stanford.edu/class/ee398a/handouts/papers/Moffat98ArithmCoding.pdf"
}*/
static int ec_read_byte(ec_dec *_this){
@ -138,7 +138,7 @@ void ec_dec_init(ec_dec *_this,unsigned char *_buf,opus_uint32 _storage){
unsigned ec_decode(ec_dec *_this,unsigned _ft){
unsigned s;
_this->ext=_this->rng/_ft;
_this->ext=celt_udiv(_this->rng,_ft);
s=(unsigned)(_this->val/_this->ext);
return _ft-EC_MINI(s+1,_ft);
}

4
Frameworks/Opus/Opus/opus/celt/entenc.c
View file

@ -98,7 +98,7 @@ static void ec_enc_carry_out(ec_enc *_this,int _c){
else _this->ext++;
}
static void ec_enc_normalize(ec_enc *_this){
static OPUS_INLINE void ec_enc_normalize(ec_enc *_this){
/*If the range is too small, output some bits and rescale it.*/
while(_this->rng<=EC_CODE_BOT){
ec_enc_carry_out(_this,(int)(_this->val>>EC_CODE_SHIFT));
@ -127,7 +127,7 @@ void ec_enc_init(ec_enc *_this,unsigned char *_buf,opus_uint32 _size){
void ec_encode(ec_enc *_this,unsigned _fl,unsigned _fh,unsigned _ft){
opus_uint32 r;
r=_this->rng/_ft;
r=celt_udiv(_this->rng,_ft);
if(_fl>0){
_this->val+=_this->rng-IMUL32(r,(_ft-_fl));
_this->rng=IMUL32(r,(_fh-_fl));

79
Frameworks/Opus/Opus/opus/celt/fixed_debug.h
View file

@ -33,9 +33,9 @@
#define FIXED_DEBUG_H
#include <stdio.h>
#include "opus_defines.h"
#ifdef CELT_C
#include "opus_defines.h"
OPUS_EXPORT opus_int64 celt_mips=0;
#else
extern opus_int64 celt_mips;
@ -59,7 +59,7 @@ extern opus_int64 celt_mips;
#define SHR(a,b) SHR32(a,b)
#define PSHR(a,b) PSHR32(a,b)
static inline short NEG16(int x)
static OPUS_INLINE short NEG16(int x)
{
int res;
if (!VERIFY_SHORT(x))
@ -80,7 +80,7 @@ static inline short NEG16(int x)
celt_mips++;
return res;
}
static inline int NEG32(opus_int64 x)
static OPUS_INLINE int NEG32(opus_int64 x)
{
opus_int64 res;
if (!VERIFY_INT(x))
@ -103,7 +103,7 @@ static inline int NEG32(opus_int64 x)
}
#define EXTRACT16(x) EXTRACT16_(x, __FILE__, __LINE__)
static inline short EXTRACT16_(int x, char *file, int line)
static OPUS_INLINE short EXTRACT16_(int x, char *file, int line)
{
int res;
if (!VERIFY_SHORT(x))
@ -119,7 +119,7 @@ static inline short EXTRACT16_(int x, char *file, int line)
}
#define EXTEND32(x) EXTEND32_(x, __FILE__, __LINE__)
static inline int EXTEND32_(int x, char *file, int line)
static OPUS_INLINE int EXTEND32_(int x, char *file, int line)
{
int res;
if (!VERIFY_SHORT(x))
@ -135,7 +135,7 @@ static inline int EXTEND32_(int x, char *file, int line)
}
#define SHR16(a, shift) SHR16_(a, shift, __FILE__, __LINE__)
static inline short SHR16_(int a, int shift, char *file, int line)
static OPUS_INLINE short SHR16_(int a, int shift, char *file, int line)
{
int res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(shift))
@ -157,7 +157,7 @@ static inline short SHR16_(int a, int shift, char *file, int line)
return res;
}
#define SHL16(a, shift) SHL16_(a, shift, __FILE__, __LINE__)
static inline short SHL16_(int a, int shift, char *file, int line)
static OPUS_INLINE short SHL16_(int a, int shift, char *file, int line)
{
int res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(shift))
@ -179,7 +179,7 @@ static inline short SHL16_(int a, int shift, char *file, int line)
return res;
}
static inline int SHR32(opus_int64 a, int shift)
static OPUS_INLINE int SHR32(opus_int64 a, int shift)
{
opus_int64 res;
if (!VERIFY_INT(a) || !VERIFY_SHORT(shift))
@ -201,7 +201,7 @@ static inline int SHR32(opus_int64 a, int shift)
return res;
}
#define SHL32(a, shift) SHL32_(a, shift, __FILE__, __LINE__)
static inline int SHL32_(opus_int64 a, int shift, char *file, int line)
static OPUS_INLINE int SHL32_(opus_int64 a, int shift, char *file, int line)
{
opus_int64 res;
if (!VERIFY_INT(a) || !VERIFY_SHORT(shift))
@ -234,7 +234,7 @@ static inline int SHL32_(opus_int64 a, int shift, char *file, int line)
//#define SHL(a,shift) ((a) << (shift))
#define ADD16(a, b) ADD16_(a, b, __FILE__, __LINE__)
static inline short ADD16_(int a, int b, char *file, int line)
static OPUS_INLINE short ADD16_(int a, int b, char *file, int line)
{
int res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -257,7 +257,7 @@ static inline short ADD16_(int a, int b, char *file, int line)
}
#define SUB16(a, b) SUB16_(a, b, __FILE__, __LINE__)
static inline short SUB16_(int a, int b, char *file, int line)
static OPUS_INLINE short SUB16_(int a, int b, char *file, int line)
{
int res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -280,7 +280,7 @@ static inline short SUB16_(int a, int b, char *file, int line)
}
#define ADD32(a, b) ADD32_(a, b, __FILE__, __LINE__)
static inline int ADD32_(opus_int64 a, opus_int64 b, char *file, int line)
static OPUS_INLINE int ADD32_(opus_int64 a, opus_int64 b, char *file, int line)
{
opus_int64 res;
if (!VERIFY_INT(a) || !VERIFY_INT(b))
@ -303,7 +303,7 @@ static inline int ADD32_(opus_int64 a, opus_int64 b, char *file, int line)
}
#define SUB32(a, b) SUB32_(a, b, __FILE__, __LINE__)
static inline int SUB32_(opus_int64 a, opus_int64 b, char *file, int line)
static OPUS_INLINE int SUB32_(opus_int64 a, opus_int64 b, char *file, int line)
{
opus_int64 res;
if (!VERIFY_INT(a) || !VERIFY_INT(b))
@ -327,7 +327,7 @@ static inline int SUB32_(opus_int64 a, opus_int64 b, char *file, int line)
#undef UADD32
#define UADD32(a, b) UADD32_(a, b, __FILE__, __LINE__)
static inline unsigned int UADD32_(opus_uint64 a, opus_uint64 b, char *file, int line)
static OPUS_INLINE unsigned int UADD32_(opus_uint64 a, opus_uint64 b, char *file, int line)
{
opus_uint64 res;
if (!VERIFY_UINT(a) || !VERIFY_UINT(b))
@ -351,7 +351,7 @@ static inline unsigned int UADD32_(opus_uint64 a, opus_uint64 b, char *file, int
#undef USUB32
#define USUB32(a, b) USUB32_(a, b, __FILE__, __LINE__)
static inline unsigned int USUB32_(opus_uint64 a, opus_uint64 b, char *file, int line)
static OPUS_INLINE unsigned int USUB32_(opus_uint64 a, opus_uint64 b, char *file, int line)
{
opus_uint64 res;
if (!VERIFY_UINT(a) || !VERIFY_UINT(b))
@ -381,7 +381,7 @@ static inline unsigned int USUB32_(opus_uint64 a, opus_uint64 b, char *file, int
}
/* result fits in 16 bits */
static inline short MULT16_16_16(int a, int b)
static OPUS_INLINE short MULT16_16_16(int a, int b)
{
int res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -404,7 +404,7 @@ static inline short MULT16_16_16(int a, int b)
}
#define MULT16_16(a, b) MULT16_16_(a, b, __FILE__, __LINE__)
static inline int MULT16_16_(int a, int b, char *file, int line)
static OPUS_INLINE int MULT16_16_(int a, int b, char *file, int line)
{
opus_int64 res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -429,7 +429,7 @@ static inline int MULT16_16_(int a, int b, char *file, int line)
#define MAC16_16(c,a,b) (celt_mips-=2,ADD32((c),MULT16_16((a),(b))))
#define MULT16_32_QX(a, b, Q) MULT16_32_QX_(a, b, Q, __FILE__, __LINE__)
static inline int MULT16_32_QX_(int a, opus_int64 b, int Q, char *file, int line)
static OPUS_INLINE int MULT16_32_QX_(int a, opus_int64 b, int Q, char *file, int line)
{
opus_int64 res;
if (!VERIFY_SHORT(a) || !VERIFY_INT(b))
@ -462,7 +462,7 @@ static inline int MULT16_32_QX_(int a, opus_int64 b, int Q, char *file, int line
}
#define MULT16_32_PX(a, b, Q) MULT16_32_PX_(a, b, Q, __FILE__, __LINE__)
static inline int MULT16_32_PX_(int a, opus_int64 b, int Q, char *file, int line)
static OPUS_INLINE int MULT16_32_PX_(int a, opus_int64 b, int Q, char *file, int line)
{
opus_int64 res;
if (!VERIFY_SHORT(a) || !VERIFY_INT(b))
@ -496,8 +496,9 @@ static inline int MULT16_32_PX_(int a, opus_int64 b, int Q, char *file, int line
#define MULT16_32_Q15(a,b) MULT16_32_QX(a,b,15)
#define MAC16_32_Q15(c,a,b) (celt_mips-=2,ADD32((c),MULT16_32_Q15((a),(b))))
#define MAC16_32_Q16(c,a,b) (celt_mips-=2,ADD32((c),MULT16_32_Q16((a),(b))))
static inline int SATURATE(int a, int b)
static OPUS_INLINE int SATURATE(int a, int b)
{
if (a>b)
a=b;
@ -507,7 +508,17 @@ static inline int SATURATE(int a, int b)
return a;
}
static inline int MULT16_16_Q11_32(int a, int b)
static OPUS_INLINE opus_int16 SATURATE16(opus_int32 a)
{
celt_mips+=3;
if (a>32767)
return 32767;
else if (a<-32768)
return -32768;
else return a;
}
static OPUS_INLINE int MULT16_16_Q11_32(int a, int b)
{
opus_int64 res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -529,7 +540,7 @@ static inline int MULT16_16_Q11_32(int a, int b)
celt_mips+=3;
return res;
}
static inline short MULT16_16_Q13(int a, int b)
static OPUS_INLINE short MULT16_16_Q13(int a, int b)
{
opus_int64 res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -551,7 +562,7 @@ static inline short MULT16_16_Q13(int a, int b)
celt_mips+=3;
return res;
}
static inline short MULT16_16_Q14(int a, int b)
static OPUS_INLINE short MULT16_16_Q14(int a, int b)
{
opus_int64 res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -575,7 +586,7 @@ static inline short MULT16_16_Q14(int a, int b)
}
#define MULT16_16_Q15(a, b) MULT16_16_Q15_(a, b, __FILE__, __LINE__)
static inline short MULT16_16_Q15_(int a, int b, char *file, int line)
static OPUS_INLINE short MULT16_16_Q15_(int a, int b, char *file, int line)
{
opus_int64 res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -598,7 +609,7 @@ static inline short MULT16_16_Q15_(int a, int b, char *file, int line)
return res;
}
static inline short MULT16_16_P13(int a, int b)
static OPUS_INLINE short MULT16_16_P13(int a, int b)
{
opus_int64 res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -628,7 +639,7 @@ static inline short MULT16_16_P13(int a, int b)
celt_mips+=4;
return res;
}
static inline short MULT16_16_P14(int a, int b)
static OPUS_INLINE short MULT16_16_P14(int a, int b)
{
opus_int64 res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -658,7 +669,7 @@ static inline short MULT16_16_P14(int a, int b)
celt_mips+=4;
return res;
}
static inline short MULT16_16_P15(int a, int b)
static OPUS_INLINE short MULT16_16_P15(int a, int b)
{
opus_int64 res;
if (!VERIFY_SHORT(a) || !VERIFY_SHORT(b))
@ -691,7 +702,7 @@ static inline short MULT16_16_P15(int a, int b)
#define DIV32_16(a, b) DIV32_16_(a, b, __FILE__, __LINE__)
static inline int DIV32_16_(opus_int64 a, opus_int64 b, char *file, int line)
static OPUS_INLINE int DIV32_16_(opus_int64 a, opus_int64 b, char *file, int line)
{
opus_int64 res;
if (b==0)
@ -726,7 +737,7 @@ static inline int DIV32_16_(opus_int64 a, opus_int64 b, char *file, int line)
}
#define DIV32(a, b) DIV32_(a, b, __FILE__, __LINE__)
static inline int DIV32_(opus_int64 a, opus_int64 b, char *file, int line)
static OPUS_INLINE int DIV32_(opus_int64 a, opus_int64 b, char *file, int line)
{
opus_int64 res;
if (b==0)
@ -757,6 +768,16 @@ static inline int DIV32_(opus_int64 a, opus_int64 b, char *file, int line)
return res;
}
static OPUS_INLINE opus_val16 SIG2WORD16_generic(celt_sig x)
{
x = PSHR32(x, SIG_SHIFT);
x = MAX32(x, -32768);
x = MIN32(x, 32767);
return EXTRACT16(x);
}
#define SIG2WORD16(x) (SIG2WORD16_generic(x))
#undef PRINT_MIPS
#define PRINT_MIPS(file) do {fprintf (file, "total complexity = %llu MIPS\n", celt_mips);} while (0);

28
Frameworks/Opus/Opus/opus/celt/fixed_generic.h
View file

@ -40,7 +40,7 @@
#define MULT16_32_Q16(a,b) ADD32(MULT16_16((a),SHR((b),16)), SHR(MULT16_16SU((a),((b)&0x0000ffff)),16))
/** 16x32 multiplication, followed by a 16-bit shift right (round-to-nearest). Results fits in 32 bits */
#define MULT16_32_P16(a,b) ADD32(MULT16_16((a),SHR((b),16)), PSHR(MULT16_16((a),((b)&0x0000ffff)),16))
#define MULT16_32_P16(a,b) ADD32(MULT16_16((a),SHR((b),16)), PSHR(MULT16_16SU((a),((b)&0x0000ffff)),16))
/** 16x32 multiplication, followed by a 15-bit shift right. Results fits in 32 bits */
#define MULT16_32_Q15(a,b) ADD32(SHL(MULT16_16((a),SHR((b),16)),1), SHR(MULT16_16SU((a),((b)&0x0000ffff)),15))
@ -84,6 +84,8 @@
#define PSHR(a,shift) (SHR((a)+((EXTEND32(1)<<((shift))>>1)),shift))
#define SATURATE(x,a) (((x)>(a) ? (a) : (x)<-(a) ? -(a) : (x)))
#define SATURATE16(x) (EXTRACT16((x)>32767 ? 32767 : (x)<-32768 ? -32768 : (x)))
/** Shift by a and round-to-neareast 32-bit value. Result is a 16-bit value */
#define ROUND16(x,a) (EXTRACT16(PSHR32((x),(a))))
/** Divide by two */
@ -108,10 +110,17 @@
/** 16x16 multiply-add where the result fits in 32 bits */
#define MAC16_16(c,a,b) (ADD32((c),MULT16_16((a),(b))))
/** 16x32 multiply-add, followed by a 15-bit shift right. Results fits in 32 bits */
#define MAC16_32_Q15(c,a,b) ADD32(c,ADD32(MULT16_16((a),SHR((b),15)), SHR(MULT16_16((a),((b)&0x00007fff)),15)))
/** 16x32 multiply, followed by a 15-bit shift right and 32-bit add.
b must fit in 31 bits.
Result fits in 32 bits. */
#define MAC16_32_Q15(c,a,b) ADD32((c),ADD32(MULT16_16((a),SHR((b),15)), SHR(MULT16_16((a),((b)&0x00007fff)),15)))
/** 16x32 multiplication, followed by a 16-bit shift right and 32-bit add.
Results fits in 32 bits */
#define MAC16_32_Q16(c,a,b) ADD32((c),ADD32(MULT16_16((a),SHR((b),16)), SHR(MULT16_16SU((a),((b)&0x0000ffff)),16)))
#define MULT16_16_Q11_32(a,b) (SHR(MULT16_16((a),(b)),11))
#define MULT16_16_Q11(a,b) (SHR(MULT16_16((a),(b)),11))
#define MULT16_16_Q13(a,b) (SHR(MULT16_16((a),(b)),13))
#define MULT16_16_Q14(a,b) (SHR(MULT16_16((a),(b)),14))
#define MULT16_16_Q15(a,b) (SHR(MULT16_16((a),(b)),15))
@ -126,4 +135,17 @@
/** Divide a 32-bit value by a 32-bit value. Result fits in 32 bits */
#define DIV32(a,b) (((opus_val32)(a))/((opus_val32)(b)))
#if defined(MIPSr1_ASM)
#include "mips/fixed_generic_mipsr1.h"
#endif
static OPUS_INLINE opus_val16 SIG2WORD16_generic(celt_sig x)
{
x = PSHR32(x, SIG_SHIFT);
x = MAX32(x, -32768);
x = MIN32(x, 32767);
return EXTRACT16(x);
}
#define SIG2WORD16(x) (SIG2WORD16_generic(x))
#endif

8
Frameworks/Opus/Opus/opus/celt/float_cast.h
View file

@ -90,18 +90,18 @@
#include <math.h>
#define float2int(x) lrint(x)
#elif (defined(_MSC_VER) && _MSC_VER >= 1400) && (defined (WIN64) || defined (_WIN64))
#elif (defined(_MSC_VER) && _MSC_VER >= 1400) && defined (_M_X64)
#include <xmmintrin.h>
__inline long int float2int(float value)
{
return _mm_cvtss_si32(_mm_load_ss(&value));
}
#elif (defined(_MSC_VER) && _MSC_VER >= 1400) && (defined (WIN32) || defined (_WIN32))
#elif (defined(_MSC_VER) && _MSC_VER >= 1400) && defined (_M_IX86)
#include <math.h>
/* Win32 doesn't seem to have these functions.
** Therefore implement inline versions of these functions here.
** Therefore implement OPUS_INLINE versions of these functions here.
*/
__inline long int
@ -128,7 +128,7 @@
#endif
#ifndef DISABLE_FLOAT_API
static inline opus_int16 FLOAT2INT16(float x)
static OPUS_INLINE opus_int16 FLOAT2INT16(float x)
{
x = x*CELT_SIG_SCALE;
x = MAX32(x, -32768);

464
Frameworks/Opus/Opus/opus/celt/kiss_fft.c
View file

@ -40,7 +40,6 @@
#include "os_support.h"
#include "mathops.h"
#include "stack_alloc.h"
#include "os_support.h"
/* The guts header contains all the multiplication and addition macros that are defined for
complex numbers. It also delares the kf_ internal functions.
@ -48,64 +47,56 @@
static void kf_bfly2(
kiss_fft_cpx * Fout,
const size_t fstride,
const kiss_fft_state *st,
int m,
int N,
int mm
int N
)
{
kiss_fft_cpx * Fout2;
const kiss_twiddle_cpx * tw1;
int i,j;
kiss_fft_cpx * Fout_beg = Fout;
int i;
(void)m;
#ifdef CUSTOM_MODES
if (m==1)
{
celt_assert(m==1);
for (i=0;i<N;i++)
{
Fout = Fout_beg + i*mm;
Fout2 = Fout + m;
tw1 = st->twiddles;
for(j=0;j<m;j++)
{
kiss_fft_cpx t;
Fout->r = SHR32(Fout->r, 1);Fout->i = SHR32(Fout->i, 1);
Fout2->r = SHR32(Fout2->r, 1);Fout2->i = SHR32(Fout2->i, 1);
C_MUL (t, *Fout2 , *tw1);
tw1 += fstride;
Fout2 = Fout + 1;
t = *Fout2;
C_SUB( *Fout2 , *Fout , t );
C_ADDTO( *Fout , t );
++Fout2;
++Fout;
Fout += 2;
}
}
}
} else
#endif
{
opus_val16 tw;
tw = QCONST16(0.7071067812f, 15);
/* We know that m==4 here because the radix-2 is just after a radix-4 */
celt_assert(m==4);
for (i=0;i<N;i++)
{
kiss_fft_cpx t;
Fout2 = Fout + 4;
t = Fout2[0];
C_SUB( Fout2[0] , Fout[0] , t );
C_ADDTO( Fout[0] , t );
static void ki_bfly2(
kiss_fft_cpx * Fout,
const size_t fstride,
const kiss_fft_state *st,
int m,
int N,
int mm
)
{
kiss_fft_cpx * Fout2;
const kiss_twiddle_cpx * tw1;
kiss_fft_cpx t;
int i,j;
kiss_fft_cpx * Fout_beg = Fout;
for (i=0;i<N;i++)
{
Fout = Fout_beg + i*mm;
Fout2 = Fout + m;
tw1 = st->twiddles;
for(j=0;j<m;j++)
{
C_MULC (t, *Fout2 , *tw1);
tw1 += fstride;
C_SUB( *Fout2 , *Fout , t );
C_ADDTO( *Fout , t );
++Fout2;
++Fout;
t.r = S_MUL(Fout2[1].r+Fout2[1].i, tw);
t.i = S_MUL(Fout2[1].i-Fout2[1].r, tw);
C_SUB( Fout2[1] , Fout[1] , t );
C_ADDTO( Fout[1] , t );
t.r = Fout2[2].i;
t.i = -Fout2[2].r;
C_SUB( Fout2[2] , Fout[2] , t );
C_ADDTO( Fout[2] , t );
t.r = S_MUL(Fout2[3].i-Fout2[3].r, tw);
t.i = S_MUL(-Fout2[3].i-Fout2[3].r, tw);
C_SUB( Fout2[3] , Fout[3] , t );
C_ADDTO( Fout[3] , t );
Fout += 8;
}
}
}
@ -119,31 +110,50 @@ static void kf_bfly4(
int mm
)
{
const kiss_twiddle_cpx *tw1,*tw2,*tw3;
kiss_fft_cpx scratch[6];
const size_t m2=2*m;
const size_t m3=3*m;
int i, j;
int i;
if (m==1)
{
/* Degenerate case where all the twiddles are 1. */
for (i=0;i<N;i++)
{
kiss_fft_cpx scratch0, scratch1;
C_SUB( scratch0 , *Fout, Fout[2] );
C_ADDTO(*Fout, Fout[2]);
C_ADD( scratch1 , Fout[1] , Fout[3] );
C_SUB( Fout[2], *Fout, scratch1 );
C_ADDTO( *Fout , scratch1 );
C_SUB( scratch1 , Fout[1] , Fout[3] );
Fout[1].r = scratch0.r + scratch1.i;
Fout[1].i = scratch0.i - scratch1.r;
Fout[3].r = scratch0.r - scratch1.i;
Fout[3].i = scratch0.i + scratch1.r;
Fout+=4;
}
} else {
int j;
kiss_fft_cpx scratch[6];
const kiss_twiddle_cpx *tw1,*tw2,*tw3;
const int m2=2*m;
const int m3=3*m;
kiss_fft_cpx * Fout_beg = Fout;
for (i=0;i<N;i++)
{
Fout = Fout_beg + i*mm;
tw3 = tw2 = tw1 = st->twiddles;
/* m is guaranteed to be a multiple of 4. */
for (j=0;j<m;j++)
{
C_MUL4(scratch[0],Fout[m] , *tw1 );
C_MUL4(scratch[1],Fout[m2] , *tw2 );
C_MUL4(scratch[2],Fout[m3] , *tw3 );
C_MUL(scratch[0],Fout[m] , *tw1 );
C_MUL(scratch[1],Fout[m2] , *tw2 );
C_MUL(scratch[2],Fout[m3] , *tw3 );
Fout->r = PSHR32(Fout->r, 2);
Fout->i = PSHR32(Fout->i, 2);
C_SUB( scratch[5] , *Fout, scratch[1] );
C_ADDTO(*Fout, scratch[1]);
C_ADD( scratch[3] , scratch[0] , scratch[2] );
C_SUB( scratch[4] , scratch[0] , scratch[2] );
Fout[m2].r = PSHR32(Fout[m2].r, 2);
Fout[m2].i = PSHR32(Fout[m2].i, 2);
C_SUB( Fout[m2], *Fout, scratch[3] );
tw1 += fstride;
tw2 += fstride*2;
@ -157,53 +167,10 @@ static void kf_bfly4(
++Fout;
}
}
}
static void ki_bfly4(
kiss_fft_cpx * Fout,
const size_t fstride,
const kiss_fft_state *st,
int m,
int N,
int mm
)
{
const kiss_twiddle_cpx *tw1,*tw2,*tw3;
kiss_fft_cpx scratch[6];
const size_t m2=2*m;
const size_t m3=3*m;
int i, j;
kiss_fft_cpx * Fout_beg = Fout;
for (i=0;i<N;i++)
{
Fout = Fout_beg + i*mm;
tw3 = tw2 = tw1 = st->twiddles;
for (j=0;j<m;j++)
{
C_MULC(scratch[0],Fout[m] , *tw1 );
C_MULC(scratch[1],Fout[m2] , *tw2 );
C_MULC(scratch[2],Fout[m3] , *tw3 );
C_SUB( scratch[5] , *Fout, scratch[1] );
C_ADDTO(*Fout, scratch[1]);
C_ADD( scratch[3] , scratch[0] , scratch[2] );
C_SUB( scratch[4] , scratch[0] , scratch[2] );
C_SUB( Fout[m2], *Fout, scratch[3] );
tw1 += fstride;
tw2 += fstride*2;
tw3 += fstride*3;
C_ADDTO( *Fout , scratch[3] );
Fout[m].r = scratch[5].r - scratch[4].i;
Fout[m].i = scratch[5].i + scratch[4].r;
Fout[m3].r = scratch[5].r + scratch[4].i;
Fout[m3].i = scratch[5].i - scratch[4].r;
++Fout;
}
}
}
#ifndef RADIX_TWO_ONLY
static void kf_bfly3(
@ -223,14 +190,19 @@ static void kf_bfly3(
kiss_twiddle_cpx epi3;
kiss_fft_cpx * Fout_beg = Fout;
#ifdef FIXED_POINT
epi3.r = -16384;
epi3.i = -28378;
#else
epi3 = st->twiddles[fstride*m];
#endif
for (i=0;i<N;i++)
{
Fout = Fout_beg + i*mm;
tw1=tw2=st->twiddles;
/* For non-custom modes, m is guaranteed to be a multiple of 4. */
k=m;
do {
C_FIXDIV(*Fout,3); C_FIXDIV(Fout[m],3); C_FIXDIV(Fout[m2],3);
C_MUL(scratch[1],Fout[m] , *tw1);
C_MUL(scratch[2],Fout[m2] , *tw2);
@ -258,56 +230,8 @@ static void kf_bfly3(
}
}
static void ki_bfly3(
kiss_fft_cpx * Fout,
const size_t fstride,
const kiss_fft_state *st,
int m,
int N,
int mm
)
{
int i, k;
const size_t m2 = 2*m;
const kiss_twiddle_cpx *tw1,*tw2;
kiss_fft_cpx scratch[5];
kiss_twiddle_cpx epi3;
kiss_fft_cpx * Fout_beg = Fout;
epi3 = st->twiddles[fstride*m];
for (i=0;i<N;i++)
{
Fout = Fout_beg + i*mm;
tw1=tw2=st->twiddles;
k=m;
do{
C_MULC(scratch[1],Fout[m] , *tw1);
C_MULC(scratch[2],Fout[m2] , *tw2);
C_ADD(scratch[3],scratch[1],scratch[2]);
C_SUB(scratch[0],scratch[1],scratch[2]);
tw1 += fstride;
tw2 += fstride*2;
Fout[m].r = Fout->r - HALF_OF(scratch[3].r);
Fout[m].i = Fout->i - HALF_OF(scratch[3].i);
C_MULBYSCALAR( scratch[0] , -epi3.i );
C_ADDTO(*Fout,scratch[3]);
Fout[m2].r = Fout[m].r + scratch[0].i;
Fout[m2].i = Fout[m].i - scratch[0].r;
Fout[m].r -= scratch[0].i;
Fout[m].i += scratch[0].r;
++Fout;
}while(--k);
}
}
#ifndef OVERRIDE_kf_bfly5
static void kf_bfly5(
kiss_fft_cpx * Fout,
const size_t fstride,
@ -320,13 +244,19 @@ static void kf_bfly5(
kiss_fft_cpx *Fout0,*Fout1,*Fout2,*Fout3,*Fout4;
int i, u;
kiss_fft_cpx scratch[13];
const kiss_twiddle_cpx * twiddles = st->twiddles;
const kiss_twiddle_cpx *tw;
kiss_twiddle_cpx ya,yb;
kiss_fft_cpx * Fout_beg = Fout;
ya = twiddles[fstride*m];
yb = twiddles[fstride*2*m];
#ifdef FIXED_POINT
ya.r = 10126;
ya.i = -31164;
yb.r = -26510;
yb.i = -19261;
#else
ya = st->twiddles[fstride*m];
yb = st->twiddles[fstride*2*m];
#endif
tw=st->twiddles;
for (i=0;i<N;i++)
@ -338,8 +268,8 @@ static void kf_bfly5(
Fout3=Fout0+3*m;
Fout4=Fout0+4*m;
/* For non-custom modes, m is guaranteed to be a multiple of 4. */
for ( u=0; u<m; ++u ) {
C_FIXDIV( *Fout0,5); C_FIXDIV( *Fout1,5); C_FIXDIV( *Fout2,5); C_FIXDIV( *Fout3,5); C_FIXDIV( *Fout4,5);
scratch[0] = *Fout0;
C_MUL(scratch[1] ,*Fout1, tw[u*fstride]);
@ -376,74 +306,8 @@ static void kf_bfly5(
}
}
}
#endif /* OVERRIDE_kf_bfly5 */
static void ki_bfly5(
kiss_fft_cpx * Fout,
const size_t fstride,
const kiss_fft_state *st,
int m,
int N,
int mm
)
{
kiss_fft_cpx *Fout0,*Fout1,*Fout2,*Fout3,*Fout4;
int i, u;
kiss_fft_cpx scratch[13];
const kiss_twiddle_cpx * twiddles = st->twiddles;
const kiss_twiddle_cpx *tw;
kiss_twiddle_cpx ya,yb;
kiss_fft_cpx * Fout_beg = Fout;
ya = twiddles[fstride*m];
yb = twiddles[fstride*2*m];
tw=st->twiddles;
for (i=0;i<N;i++)
{
Fout = Fout_beg + i*mm;
Fout0=Fout;
Fout1=Fout0+m;
Fout2=Fout0+2*m;
Fout3=Fout0+3*m;
Fout4=Fout0+4*m;
for ( u=0; u<m; ++u ) {
scratch[0] = *Fout0;
C_MULC(scratch[1] ,*Fout1, tw[u*fstride]);
C_MULC(scratch[2] ,*Fout2, tw[2*u*fstride]);
C_MULC(scratch[3] ,*Fout3, tw[3*u*fstride]);
C_MULC(scratch[4] ,*Fout4, tw[4*u*fstride]);
C_ADD( scratch[7],scratch[1],scratch[4]);
C_SUB( scratch[10],scratch[1],scratch[4]);
C_ADD( scratch[8],scratch[2],scratch[3]);
C_SUB( scratch[9],scratch[2],scratch[3]);
Fout0->r += scratch[7].r + scratch[8].r;
Fout0->i += scratch[7].i + scratch[8].i;
scratch[5].r = scratch[0].r + S_MUL(scratch[7].r,ya.r) + S_MUL(scratch[8].r,yb.r);
scratch[5].i = scratch[0].i + S_MUL(scratch[7].i,ya.r) + S_MUL(scratch[8].i,yb.r);
scratch[6].r = -S_MUL(scratch[10].i,ya.i) - S_MUL(scratch[9].i,yb.i);
scratch[6].i = S_MUL(scratch[10].r,ya.i) + S_MUL(scratch[9].r,yb.i);
C_SUB(*Fout1,scratch[5],scratch[6]);
C_ADD(*Fout4,scratch[5],scratch[6]);
scratch[11].r = scratch[0].r + S_MUL(scratch[7].r,yb.r) + S_MUL(scratch[8].r,ya.r);
scratch[11].i = scratch[0].i + S_MUL(scratch[7].i,yb.r) + S_MUL(scratch[8].i,ya.r);
scratch[12].r = S_MUL(scratch[10].i,yb.i) - S_MUL(scratch[9].i,ya.i);
scratch[12].i = -S_MUL(scratch[10].r,yb.i) + S_MUL(scratch[9].r,ya.i);
C_ADD(*Fout2,scratch[11],scratch[12]);
C_SUB(*Fout3,scratch[11],scratch[12]);
++Fout0;++Fout1;++Fout2;++Fout3;++Fout4;
}
}
}
#endif
@ -491,6 +355,9 @@ static
int kf_factor(int n,opus_int16 * facbuf)
{
int p=4;
int i;
int stages=0;
int nbak = n;
/*factor out powers of 4, powers of 2, then any remaining primes */
do {
@ -512,9 +379,30 @@ int kf_factor(int n,opus_int16 * facbuf)
{
return 0;
}
*facbuf++ = p;
*facbuf++ = n;
facbuf[2*stages] = p;
if (p==2 && stages > 1)
{
facbuf[2*stages] = 4;
facbuf[2] = 2;
}
stages++;
} while (n > 1);
n = nbak;
/* Reverse the order to get the radix 4 at the end, so we can use the
fast degenerate case. It turns out that reversing the order also
improves the noise behaviour. */
for (i=0;i<stages/2;i++)
{
int tmp;
tmp = facbuf[2*i];
facbuf[2*i] = facbuf[2*(stages-i-1)];
facbuf[2*(stages-i-1)] = tmp;
}
for (i=0;i<stages;i++)
{
n /= facbuf[2*i];
facbuf[2*i+1] = n;
}
return 1;
}
@ -535,13 +423,19 @@ static void compute_twiddles(kiss_twiddle_cpx *twiddles, int nfft)
#endif
}
int opus_fft_alloc_arch_c(kiss_fft_state *st) {
(void)st;
return 0;
}
/*
*
* Allocates all necessary storage space for the fft and ifft.
* The return value is a contiguous block of memory. As such,
* It can be freed with free().
* */
kiss_fft_state *opus_fft_alloc_twiddles(int nfft,void * mem,size_t * lenmem, const kiss_fft_state *base)
kiss_fft_state *opus_fft_alloc_twiddles(int nfft,void * mem,size_t * lenmem,
const kiss_fft_state *base, int arch)
{
kiss_fft_state *st=NULL;
size_t memneeded = sizeof(struct kiss_fft_state); /* twiddle factors*/
@ -558,14 +452,20 @@ kiss_fft_state *opus_fft_alloc_twiddles(int nfft,void * mem,size_t * lenmem, co
kiss_twiddle_cpx *twiddles;
st->nfft=nfft;
#ifndef FIXED_POINT
#ifdef FIXED_POINT
st->scale_shift = celt_ilog2(st->nfft);
if (st->nfft == 1<<st->scale_shift)
st->scale = Q15ONE;
else
st->scale = (1073741824+st->nfft/2)/st->nfft>>(15-st->scale_shift);
#else
st->scale = 1.f/nfft;
#endif
if (base != NULL)
{
st->twiddles = base->twiddles;
st->shift = 0;
while (nfft<<st->shift != base->nfft && st->shift < 32)
while (st->shift < 32 && nfft<<st->shift != base->nfft)
st->shift++;
if (st->shift>=32)
goto fail;
@ -584,22 +484,31 @@ kiss_fft_state *opus_fft_alloc_twiddles(int nfft,void * mem,size_t * lenmem, co
if (st->bitrev==NULL)
goto fail;
compute_bitrev_table(0, bitrev, 1,1, st->factors,st);
/* Initialize architecture specific fft parameters */
if (opus_fft_alloc_arch(st, arch))
goto fail;
}
return st;
fail:
opus_fft_free(st);
opus_fft_free(st, arch);
return NULL;
}
kiss_fft_state *opus_fft_alloc(int nfft,void * mem,size_t * lenmem )
kiss_fft_state *opus_fft_alloc(int nfft,void * mem,size_t * lenmem, int arch)
{
return opus_fft_alloc_twiddles(nfft, mem, lenmem, NULL);
return opus_fft_alloc_twiddles(nfft, mem, lenmem, NULL, arch);
}
void opus_fft_free(const kiss_fft_state *cfg)
void opus_fft_free_arch_c(kiss_fft_state *st) {
(void)st;
}
void opus_fft_free(const kiss_fft_state *cfg, int arch)
{
if (cfg)
{
opus_fft_free_arch((kiss_fft_state *)cfg, arch);
opus_free((opus_int16*)cfg->bitrev);
if (cfg->shift < 0)
opus_free((kiss_twiddle_cpx*)cfg->twiddles);
@ -609,7 +518,7 @@ void opus_fft_free(const kiss_fft_state *cfg)
#endif /* CUSTOM_MODES */
void opus_fft(const kiss_fft_state *st,const kiss_fft_cpx *fin,kiss_fft_cpx *fout)
void opus_fft_impl(const kiss_fft_state *st,kiss_fft_cpx *fout)
{
int m2, m;
int p;
@ -621,17 +530,6 @@ void opus_fft(const kiss_fft_state *st,const kiss_fft_cpx *fin,kiss_fft_cpx *fou
/* st->shift can be -1 */
shift = st->shift>0 ? st->shift : 0;
celt_assert2 (fin != fout, "In-place FFT not supported");
/* Bit-reverse the input */
for (i=0;i<st->nfft;i++)
{
fout[st->bitrev[i]] = fin[i];
#ifndef FIXED_POINT
fout[st->bitrev[i]].r *= st->scale;
fout[st->bitrev[i]].i *= st->scale;
#endif
}
fstride[0] = 1;
L=0;
do {
@ -650,7 +548,7 @@ void opus_fft(const kiss_fft_state *st,const kiss_fft_cpx *fin,kiss_fft_cpx *fou
switch (st->factors[2*i])
{
case 2:
kf_bfly2(fout,fstride[i]<<shift,st,m, fstride[i], m2);
kf_bfly2(fout, m, fstride[i]);
break;
case 4:
kf_bfly4(fout,fstride[i]<<shift,st,m, fstride[i], m2);
@ -668,55 +566,39 @@ void opus_fft(const kiss_fft_state *st,const kiss_fft_cpx *fin,kiss_fft_cpx *fou
}
}
void opus_ifft(const kiss_fft_state *st,const kiss_fft_cpx *fin,kiss_fft_cpx *fout)
void opus_fft_c(const kiss_fft_state *st,const kiss_fft_cpx *fin,kiss_fft_cpx *fout)
{
int m2, m;
int p;
int L;
int fstride[MAXFACTORS];
int i;
int shift;
opus_val16 scale;
#ifdef FIXED_POINT
/* Allows us to scale with MULT16_32_Q16(), which is faster than
MULT16_32_Q15() on ARM. */
int scale_shift = st->scale_shift-1;
#endif
scale = st->scale;
/* st->shift can be -1 */
shift = st->shift>0 ? st->shift : 0;
celt_assert2 (fin != fout, "In-place FFT not supported");
/* Bit-reverse the input */
for (i=0;i<st->nfft;i++)
{
kiss_fft_cpx x = fin[i];
fout[st->bitrev[i]].r = SHR32(MULT16_32_Q16(scale, x.r), scale_shift);
fout[st->bitrev[i]].i = SHR32(MULT16_32_Q16(scale, x.i), scale_shift);
}
opus_fft_impl(st, fout);
}
void opus_ifft_c(const kiss_fft_state *st,const kiss_fft_cpx *fin,kiss_fft_cpx *fout)
{
int i;
celt_assert2 (fin != fout, "In-place FFT not supported");
/* Bit-reverse the input */
for (i=0;i<st->nfft;i++)
fout[st->bitrev[i]] = fin[i];
fstride[0] = 1;
L=0;
do {
p = st->factors[2*L];
m = st->factors[2*L+1];
fstride[L+1] = fstride[L]*p;
L++;
} while(m!=1);
m = st->factors[2*L-1];
for (i=L-1;i>=0;i--)
{
if (i!=0)
m2 = st->factors[2*i-1];
else
m2 = 1;
switch (st->factors[2*i])
{
case 2:
ki_bfly2(fout,fstride[i]<<shift,st,m, fstride[i], m2);
break;
case 4:
ki_bfly4(fout,fstride[i]<<shift,st,m, fstride[i], m2);
break;
#ifndef RADIX_TWO_ONLY
case 3:
ki_bfly3(fout,fstride[i]<<shift,st,m, fstride[i], m2);
break;
case 5:
ki_bfly5(fout,fstride[i]<<shift,st,m, fstride[i], m2);
break;
#endif
}
m = m2;
}
for (i=0;i<st->nfft;i++)
fout[i].i = -fout[i].i;
opus_fft_impl(st, fout);
for (i=0;i<st->nfft;i++)
fout[i].i = -fout[i].i;
}

75
Frameworks/Opus/Opus/opus/celt/kiss_fft.h
View file

@ -32,6 +32,7 @@
#include <stdlib.h>
#include <math.h>
#include "arch.h"
#include "cpu_support.h"
#ifdef __cplusplus
extern "C" {
@ -77,17 +78,28 @@ typedef struct {
4*4*4*2
*/
typedef struct arch_fft_state{
int is_supported;
void *priv;
} arch_fft_state;
typedef struct kiss_fft_state{
int nfft;
#ifndef FIXED_POINT
kiss_fft_scalar scale;
opus_val16 scale;
#ifdef FIXED_POINT
int scale_shift;
#endif
int shift;
opus_int16 factors[2*MAXFACTORS];
const opus_int16 *bitrev;
const kiss_twiddle_cpx *twiddles;
arch_fft_state *arch_fft;
} kiss_fft_state;
#if defined(HAVE_ARM_NE10)
#include "arm/fft_arm.h"
#endif
/*typedef struct kiss_fft_state* kiss_fft_cfg;*/
/**
@ -113,9 +125,9 @@ typedef struct kiss_fft_state{
* buffer size in *lenmem.
* */
kiss_fft_state *opus_fft_alloc_twiddles(int nfft,void * mem,size_t * lenmem, const kiss_fft_state *base);
kiss_fft_state *opus_fft_alloc_twiddles(int nfft,void * mem,size_t * lenmem, const kiss_fft_state *base, int arch);
kiss_fft_state *opus_fft_alloc(int nfft,void * mem,size_t * lenmem);
kiss_fft_state *opus_fft_alloc(int nfft,void * mem,size_t * lenmem, int arch);
/**
* opus_fft(cfg,in_out_buf)
@ -127,10 +139,59 @@ kiss_fft_state *opus_fft_alloc(int nfft,void * mem,size_t * lenmem);
* Note that each element is complex and can be accessed like
f[k].r and f[k].i
* */
void opus_fft(const kiss_fft_state *cfg,const kiss_fft_cpx *fin,kiss_fft_cpx *fout);
void opus_ifft(const kiss_fft_state *cfg,const kiss_fft_cpx *fin,kiss_fft_cpx *fout);
void opus_fft_c(const kiss_fft_state *cfg,const kiss_fft_cpx *fin,kiss_fft_cpx *fout);
void opus_ifft_c(const kiss_fft_state *cfg,const kiss_fft_cpx *fin,kiss_fft_cpx *fout);
void opus_fft_free(const kiss_fft_state *cfg);
void opus_fft_impl(const kiss_fft_state *st,kiss_fft_cpx *fout);
void opus_ifft_impl(const kiss_fft_state *st,kiss_fft_cpx *fout);
void opus_fft_free(const kiss_fft_state *cfg, int arch);
void opus_fft_free_arch_c(kiss_fft_state *st);
int opus_fft_alloc_arch_c(kiss_fft_state *st);
#if !defined(OVERRIDE_OPUS_FFT)
/* Is run-time CPU detection enabled on this platform? */
#if defined(OPUS_HAVE_RTCD) && (defined(HAVE_ARM_NE10))
extern int (*const OPUS_FFT_ALLOC_ARCH_IMPL[OPUS_ARCHMASK+1])(
kiss_fft_state *st);
#define opus_fft_alloc_arch(_st, arch) \
((*OPUS_FFT_ALLOC_ARCH_IMPL[(arch)&OPUS_ARCHMASK])(_st))
extern void (*const OPUS_FFT_FREE_ARCH_IMPL[OPUS_ARCHMASK+1])(
kiss_fft_state *st);
#define opus_fft_free_arch(_st, arch) \
((*OPUS_FFT_FREE_ARCH_IMPL[(arch)&OPUS_ARCHMASK])(_st))
extern void (*const OPUS_FFT[OPUS_ARCHMASK+1])(const kiss_fft_state *cfg,
const kiss_fft_cpx *fin, kiss_fft_cpx *fout);
#define opus_fft(_cfg, _fin, _fout, arch) \
((*OPUS_FFT[(arch)&OPUS_ARCHMASK])(_cfg, _fin, _fout))
extern void (*const OPUS_IFFT[OPUS_ARCHMASK+1])(const kiss_fft_state *cfg,
const kiss_fft_cpx *fin, kiss_fft_cpx *fout);
#define opus_ifft(_cfg, _fin, _fout, arch) \
((*OPUS_IFFT[(arch)&OPUS_ARCHMASK])(_cfg, _fin, _fout))
#else /* else for if defined(OPUS_HAVE_RTCD) && (defined(HAVE_ARM_NE10)) */
#define opus_fft_alloc_arch(_st, arch) \
((void)(arch), opus_fft_alloc_arch_c(_st))
#define opus_fft_free_arch(_st, arch) \
((void)(arch), opus_fft_free_arch_c(_st))
#define opus_fft(_cfg, _fin, _fout, arch) \
((void)(arch), opus_fft_c(_cfg, _fin, _fout))
#define opus_ifft(_cfg, _fin, _fout, arch) \
((void)(arch), opus_ifft_c(_cfg, _fin, _fout))
#endif /* end if defined(OPUS_HAVE_RTCD) && (defined(HAVE_ARM_NE10)) */
#endif /* end if !defined(OVERRIDE_OPUS_FFT) */
#ifdef __cplusplus
}

4
Frameworks/Opus/Opus/opus/celt/mathops.c
View file

@ -123,6 +123,8 @@ opus_val32 celt_sqrt(opus_val32 x)
static const opus_val16 C[5] = {23175, 11561, -3011, 1699, -664};
if (x==0)
return 0;
else if (x>=1073741824)
return 32767;
k = (celt_ilog2(x)>>1)-7;
x = VSHR32(x, 2*k);
n = x-32768;
@ -137,7 +139,7 @@ opus_val32 celt_sqrt(opus_val32 x)
#define L3 8277
#define L4 -626
static inline opus_val16 _celt_cos_pi_2(opus_val16 x)
static OPUS_INLINE opus_val16 _celt_cos_pi_2(opus_val16 x)
{
opus_val16 x2;

87
Frameworks/Opus/Opus/opus/celt/mathops.h
View file

@ -43,6 +43,41 @@
unsigned isqrt32(opus_uint32 _val);
#ifndef OVERRIDE_CELT_MAXABS16
static OPUS_INLINE opus_val32 celt_maxabs16(const opus_val16 *x, int len)
{
int i;
opus_val16 maxval = 0;
opus_val16 minval = 0;
for (i=0;i<len;i++)
{
maxval = MAX16(maxval, x[i]);
minval = MIN16(minval, x[i]);
}
return MAX32(EXTEND32(maxval),-EXTEND32(minval));
}
#endif
#ifndef OVERRIDE_CELT_MAXABS32
#ifdef FIXED_POINT
static OPUS_INLINE opus_val32 celt_maxabs32(const opus_val32 *x, int len)
{
int i;
opus_val32 maxval = 0;
opus_val32 minval = 0;
for (i=0;i<len;i++)
{
maxval = MAX32(maxval, x[i]);
minval = MIN32(minval, x[i]);
}
return MAX32(maxval, -minval);
}
#else
#define celt_maxabs32(x,len) celt_maxabs16(x,len)
#endif
#endif
#ifndef FIXED_POINT
#define PI 3.141592653f
@ -60,7 +95,7 @@ unsigned isqrt32(opus_uint32 _val);
denorm, +/- inf and NaN are *not* handled */
/** Base-2 log approximation (log2(x)). */
static inline float celt_log2(float x)
static OPUS_INLINE float celt_log2(float x)
{
int integer;
float frac;
@ -78,7 +113,7 @@ static inline float celt_log2(float x)
}
/** Base-2 exponential approximation (2^x). */
static inline float celt_exp2(float x)
static OPUS_INLINE float celt_exp2(float x)
{
int integer;
float frac;
@ -110,41 +145,16 @@ static inline float celt_exp2(float x)
#ifndef OVERRIDE_CELT_ILOG2
/** Integer log in base2. Undefined for zero and negative numbers */
static inline opus_int16 celt_ilog2(opus_int32 x)
static OPUS_INLINE opus_int16 celt_ilog2(opus_int32 x)
{
celt_assert2(x>0, "celt_ilog2() only defined for strictly positive numbers");
return EC_ILOG(x)-1;
}
#endif
#ifndef OVERRIDE_CELT_MAXABS16
static inline opus_val32 celt_maxabs16(const opus_val16 *x, int len)
{
int i;
opus_val16 maxval = 0;
opus_val16 minval = 0;
for (i=0;i<len;i++)
{
maxval = MAX16(maxval, x[i]);
minval = MIN16(minval, x[i]);
}
return MAX32(EXTEND32(maxval),-EXTEND32(minval));
}
#endif
#ifndef OVERRIDE_CELT_MAXABS32
static inline opus_val32 celt_maxabs32(opus_val32 *x, int len)
{
int i;
opus_val32 maxval = 0;
for (i=0;i<len;i++)
maxval = MAX32(maxval, ABS32(x[i]));
return maxval;
}
#endif
/** Integer log in base2. Defined for zero, but not for negative numbers */
static inline opus_int16 celt_zlog2(opus_val32 x)
static OPUS_INLINE opus_int16 celt_zlog2(opus_val32 x)
{
return x <= 0 ? 0 : celt_ilog2(x);
}
@ -155,7 +165,8 @@ opus_val32 celt_sqrt(opus_val32 x);
opus_val16 celt_cos_norm(opus_val32 x);
static inline opus_val16 celt_log2(opus_val32 x)
/** Base-2 logarithm approximation (log2(x)). (Q14 input, Q10 output) */
static OPUS_INLINE opus_val16 celt_log2(opus_val32 x)
{
int i;
opus_val16 n, frac;
@ -180,8 +191,15 @@ static inline opus_val16 celt_log2(opus_val32 x)
#define D1 22804
#define D2 14819
#define D3 10204
static OPUS_INLINE opus_val32 celt_exp2_frac(opus_val16 x)
{
opus_val16 frac;
frac = SHL16(x, 4);
return ADD16(D0, MULT16_16_Q15(frac, ADD16(D1, MULT16_16_Q15(frac, ADD16(D2 , MULT16_16_Q15(D3,frac))))));
}
/** Base-2 exponential approximation (2^x). (Q10 input, Q16 output) */
static inline opus_val32 celt_exp2(opus_val16 x)
static OPUS_INLINE opus_val32 celt_exp2(opus_val16 x)
{
int integer;
opus_val16 frac;
@ -190,8 +208,7 @@ static inline opus_val32 celt_exp2(opus_val16 x)
return 0x7f000000;
else if (integer < -15)
return 0;
frac = SHL16(x-SHL16(integer,10),4);
frac = ADD16(D0, MULT16_16_Q15(frac, ADD16(D1, MULT16_16_Q15(frac, ADD16(D2 , MULT16_16_Q15(D3,frac))))));
frac = celt_exp2_frac(x-SHL16(integer,10));
return VSHR32(EXTEND32(frac), -integer-2);
}
@ -208,7 +225,7 @@ opus_val32 frac_div32(opus_val32 a, opus_val32 b);
/* Atan approximation using a 4th order polynomial. Input is in Q15 format
and normalized by pi/4. Output is in Q15 format */
static inline opus_val16 celt_atan01(opus_val16 x)
static OPUS_INLINE opus_val16 celt_atan01(opus_val16 x)
{
return MULT16_16_P15(x, ADD32(M1, MULT16_16_P15(x, ADD32(M2, MULT16_16_P15(x, ADD32(M3, MULT16_16_P15(M4, x)))))));
}
@ -219,7 +236,7 @@ static inline opus_val16 celt_atan01(opus_val16 x)
#undef M4
/* atan2() approximation valid for positive input values */
static inline opus_val16 celt_atan2p(opus_val16 y, opus_val16 x)
static OPUS_INLINE opus_val16 celt_atan2p(opus_val16 y, opus_val16 x)
{
if (y < x)
{

273
Frameworks/Opus/Opus/opus/celt/mdct.c
View file

@ -53,74 +53,100 @@
#include "mathops.h"
#include "stack_alloc.h"
#if defined(MIPSr1_ASM)
#include "mips/mdct_mipsr1.h"
#endif
#ifdef CUSTOM_MODES
int clt_mdct_init(mdct_lookup *l,int N, int maxshift)
int clt_mdct_init(mdct_lookup *l,int N, int maxshift, int arch)
{
int i;
int N4;
kiss_twiddle_scalar *trig;
#if defined(FIXED_POINT)
int shift;
int N2=N>>1;
#endif
l->n = N;
N4 = N>>2;
l->maxshift = maxshift;
for (i=0;i<=maxshift;i++)
{
if (i==0)
l->kfft[i] = opus_fft_alloc(N>>2>>i, 0, 0);
l->kfft[i] = opus_fft_alloc(N>>2>>i, 0, 0, arch);
else
l->kfft[i] = opus_fft_alloc_twiddles(N>>2>>i, 0, 0, l->kfft[0]);
l->kfft[i] = opus_fft_alloc_twiddles(N>>2>>i, 0, 0, l->kfft[0], arch);
#ifndef ENABLE_TI_DSPLIB55
if (l->kfft[i]==NULL)
return 0;
#endif
}
l->trig = trig = (kiss_twiddle_scalar*)opus_alloc((N4+1)*sizeof(kiss_twiddle_scalar));
l->trig = trig = (kiss_twiddle_scalar*)opus_alloc((N-(N2>>maxshift))*sizeof(kiss_twiddle_scalar));
if (l->trig==NULL)
return 0;
for (shift=0;shift<=maxshift;shift++)
{
/* We have enough points that sine isn't necessary */
#if defined(FIXED_POINT)
for (i=0;i<=N4;i++)
trig[i] = TRIG_UPSCALE*celt_cos_norm(DIV32(ADD32(SHL32(EXTEND32(i),17),N2),N));
#if 1
for (i=0;i<N2;i++)
trig[i] = TRIG_UPSCALE*celt_cos_norm(DIV32(ADD32(SHL32(EXTEND32(i),17),N2+16384),N));
#else
for (i=0;i<=N4;i++)
trig[i] = (kiss_twiddle_scalar)cos(2*PI*i/N);
for (i=0;i<N2;i++)
trig[i] = (kiss_twiddle_scalar)MAX32(-32767,MIN32(32767,floor(.5+32768*cos(2*M_PI*(i+.125)/N))));
#endif
#else
for (i=0;i<N2;i++)
trig[i] = (kiss_twiddle_scalar)cos(2*PI*(i+.125)/N);
#endif
trig += N2;
N2 >>= 1;
N >>= 1;
}
return 1;
}
void clt_mdct_clear(mdct_lookup *l)
void clt_mdct_clear(mdct_lookup *l, int arch)
{
int i;
for (i=0;i<=l->maxshift;i++)
opus_fft_free(l->kfft[i]);
opus_fft_free(l->kfft[i], arch);
opus_free((kiss_twiddle_scalar*)l->trig);
}
#endif /* CUSTOM_MODES */
/* Forward MDCT trashes the input array */
void clt_mdct_forward(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 *window, int overlap, int shift, int stride)
#ifndef OVERRIDE_clt_mdct_forward
void clt_mdct_forward_c(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 *window, int overlap, int shift, int stride, int arch)
{
int i;
int N, N2, N4;
kiss_twiddle_scalar sine;
VARDECL(kiss_fft_scalar, f);
VARDECL(kiss_fft_cpx, f2);
const kiss_fft_state *st = l->kfft[shift];
const kiss_twiddle_scalar *trig;
opus_val16 scale;
#ifdef FIXED_POINT
/* Allows us to scale with MULT16_32_Q16(), which is faster than
MULT16_32_Q15() on ARM. */
int scale_shift = st->scale_shift-1;
#endif
SAVE_STACK;
(void)arch;
scale = st->scale;
N = l->n;
N >>= shift;
trig = l->trig;
for (i=0;i<shift;i++)
{
N >>= 1;
trig += N;
}
N2 = N>>1;
N4 = N>>2;
ALLOC(f, N2, kiss_fft_scalar);
/* sin(x) ~= x here */
#ifdef FIXED_POINT
sine = TRIG_UPSCALE*(QCONST16(0.7853981f, 15)+N2)/N;
#else
sine = (kiss_twiddle_scalar)2*PI*(.125f)/N;
#endif
ALLOC(f2, N4, kiss_fft_cpx);
/* Consider the input to be composed of four blocks: [a, b, c, d] */
/* Window, shuffle, fold */
@ -131,7 +157,7 @@ void clt_mdct_forward(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar
kiss_fft_scalar * OPUS_RESTRICT yp = f;
const opus_val16 * OPUS_RESTRICT wp1 = window+(overlap>>1);
const opus_val16 * OPUS_RESTRICT wp2 = window+(overlap>>1)-1;
for(i=0;i<(overlap>>2);i++)
for(i=0;i<((overlap+3)>>2);i++)
{
/* Real part arranged as -d-cR, Imag part arranged as -b+aR*/
*yp++ = MULT16_32_Q15(*wp2, xp1[N2]) + MULT16_32_Q15(*wp1,*xp2);
@ -143,7 +169,7 @@ void clt_mdct_forward(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar
}
wp1 = window;
wp2 = window+overlap-1;
for(;i<N4-(overlap>>2);i++)
for(;i<N4-((overlap+3)>>2);i++)
{
/* Real part arranged as a-bR, Imag part arranged as -c-dR */
*yp++ = *xp2;
@ -165,168 +191,153 @@ void clt_mdct_forward(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar
/* Pre-rotation */
{
kiss_fft_scalar * OPUS_RESTRICT yp = f;
const kiss_twiddle_scalar *t = &l->trig[0];
const kiss_twiddle_scalar *t = &trig[0];
for(i=0;i<N4;i++)
{
kiss_fft_cpx yc;
kiss_twiddle_scalar t0, t1;
kiss_fft_scalar re, im, yr, yi;
re = yp[0];
im = yp[1];
yr = -S_MUL(re,t[i<<shift]) - S_MUL(im,t[(N4-i)<<shift]);
yi = -S_MUL(im,t[i<<shift]) + S_MUL(re,t[(N4-i)<<shift]);
/* works because the cos is nearly one */
*yp++ = yr + S_MUL(yi,sine);
*yp++ = yi - S_MUL(yr,sine);
t0 = t[i];
t1 = t[N4+i];
re = *yp++;
im = *yp++;
yr = S_MUL(re,t0) - S_MUL(im,t1);
yi = S_MUL(im,t0) + S_MUL(re,t1);
yc.r = yr;
yc.i = yi;
yc.r = PSHR32(MULT16_32_Q16(scale, yc.r), scale_shift);
yc.i = PSHR32(MULT16_32_Q16(scale, yc.i), scale_shift);
f2[st->bitrev[i]] = yc;
}
}
/* N/4 complex FFT, down-scales by 4/N */
opus_fft(l->kfft[shift], (kiss_fft_cpx *)f, (kiss_fft_cpx *)in);
/* N/4 complex FFT, does not downscale anymore */
opus_fft_impl(st, f2);
/* Post-rotate */
{
/* Temp pointers to make it really clear to the compiler what we're doing */
const kiss_fft_scalar * OPUS_RESTRICT fp = in;
const kiss_fft_cpx * OPUS_RESTRICT fp = f2;
kiss_fft_scalar * OPUS_RESTRICT yp1 = out;
kiss_fft_scalar * OPUS_RESTRICT yp2 = out+stride*(N2-1);
const kiss_twiddle_scalar *t = &l->trig[0];
const kiss_twiddle_scalar *t = &trig[0];
/* Temp pointers to make it really clear to the compiler what we're doing */
for(i=0;i<N4;i++)
{
kiss_fft_scalar yr, yi;
yr = S_MUL(fp[1],t[(N4-i)<<shift]) + S_MUL(fp[0],t[i<<shift]);
yi = S_MUL(fp[0],t[(N4-i)<<shift]) - S_MUL(fp[1],t[i<<shift]);
/* works because the cos is nearly one */
*yp1 = yr - S_MUL(yi,sine);
*yp2 = yi + S_MUL(yr,sine);;
fp += 2;
yr = S_MUL(fp->i,t[N4+i]) - S_MUL(fp->r,t[i]);
yi = S_MUL(fp->r,t[N4+i]) + S_MUL(fp->i,t[i]);
*yp1 = yr;
*yp2 = yi;
fp++;
yp1 += 2*stride;
yp2 -= 2*stride;
}
}
RESTORE_STACK;
}
#endif /* OVERRIDE_clt_mdct_forward */
void clt_mdct_backward(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 * OPUS_RESTRICT window, int overlap, int shift, int stride)
#ifndef OVERRIDE_clt_mdct_backward
void clt_mdct_backward_c(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 * OPUS_RESTRICT window, int overlap, int shift, int stride, int arch)
{
int i;
int N, N2, N4;
kiss_twiddle_scalar sine;
VARDECL(kiss_fft_scalar, f);
VARDECL(kiss_fft_scalar, f2);
SAVE_STACK;
const kiss_twiddle_scalar *trig;
(void) arch;
N = l->n;
N >>= shift;
trig = l->trig;
for (i=0;i<shift;i++)
{
N >>= 1;
trig += N;
}
N2 = N>>1;
N4 = N>>2;
ALLOC(f, N2, kiss_fft_scalar);
ALLOC(f2, N2, kiss_fft_scalar);
/* sin(x) ~= x here */
#ifdef FIXED_POINT
sine = TRIG_UPSCALE*(QCONST16(0.7853981f, 15)+N2)/N;
#else
sine = (kiss_twiddle_scalar)2*PI*(.125f)/N;
#endif
/* Pre-rotate */
{
/* Temp pointers to make it really clear to the compiler what we're doing */
const kiss_fft_scalar * OPUS_RESTRICT xp1 = in;
const kiss_fft_scalar * OPUS_RESTRICT xp2 = in+stride*(N2-1);
kiss_fft_scalar * OPUS_RESTRICT yp = f2;
const kiss_twiddle_scalar *t = &l->trig[0];
kiss_fft_scalar * OPUS_RESTRICT yp = out+(overlap>>1);
const kiss_twiddle_scalar * OPUS_RESTRICT t = &trig[0];
const opus_int16 * OPUS_RESTRICT bitrev = l->kfft[shift]->bitrev;
for(i=0;i<N4;i++)
{
int rev;
kiss_fft_scalar yr, yi;
yr = -S_MUL(*xp2, t[i<<shift]) + S_MUL(*xp1,t[(N4-i)<<shift]);
yi = -S_MUL(*xp2, t[(N4-i)<<shift]) - S_MUL(*xp1,t[i<<shift]);
/* works because the cos is nearly one */
*yp++ = yr - S_MUL(yi,sine);
*yp++ = yi + S_MUL(yr,sine);
rev = *bitrev++;
yr = S_MUL(*xp2, t[i]) + S_MUL(*xp1, t[N4+i]);
yi = S_MUL(*xp1, t[i]) - S_MUL(*xp2, t[N4+i]);
/* We swap real and imag because we use an FFT instead of an IFFT. */
yp[2*rev+1] = yr;
yp[2*rev] = yi;
/* Storing the pre-rotation directly in the bitrev order. */
xp1+=2*stride;
xp2-=2*stride;
}
}
/* Inverse N/4 complex FFT. This one should *not* downscale even in fixed-point */
opus_ifft(l->kfft[shift], (kiss_fft_cpx *)f2, (kiss_fft_cpx *)f);
opus_fft_impl(l->kfft[shift], (kiss_fft_cpx*)(out+(overlap>>1)));
/* Post-rotate */
/* Post-rotate and de-shuffle from both ends of the buffer at once to make
it in-place. */
{
kiss_fft_scalar * OPUS_RESTRICT fp = f;
const kiss_twiddle_scalar *t = &l->trig[0];
for(i=0;i<N4;i++)
kiss_fft_scalar * yp0 = out+(overlap>>1);
kiss_fft_scalar * yp1 = out+(overlap>>1)+N2-2;
const kiss_twiddle_scalar *t = &trig[0];
/* Loop to (N4+1)>>1 to handle odd N4. When N4 is odd, the
middle pair will be computed twice. */
for(i=0;i<(N4+1)>>1;i++)
{
kiss_fft_scalar re, im, yr, yi;
re = fp[0];
im = fp[1];
kiss_twiddle_scalar t0, t1;
/* We swap real and imag because we're using an FFT instead of an IFFT. */
re = yp0[1];
im = yp0[0];
t0 = t[i];
t1 = t[N4+i];
/* We'd scale up by 2 here, but instead it's done when mixing the windows */
yr = S_MUL(re,t[i<<shift]) - S_MUL(im,t[(N4-i)<<shift]);
yi = S_MUL(im,t[i<<shift]) + S_MUL(re,t[(N4-i)<<shift]);
/* works because the cos is nearly one */
*fp++ = yr - S_MUL(yi,sine);
*fp++ = yi + S_MUL(yr,sine);
yr = S_MUL(re,t0) + S_MUL(im,t1);
yi = S_MUL(re,t1) - S_MUL(im,t0);
/* We swap real and imag because we're using an FFT instead of an IFFT. */
re = yp1[1];
im = yp1[0];
yp0[0] = yr;
yp1[1] = yi;
t0 = t[(N4-i-1)];
t1 = t[(N2-i-1)];
/* We'd scale up by 2 here, but instead it's done when mixing the windows */
yr = S_MUL(re,t0) + S_MUL(im,t1);
yi = S_MUL(re,t1) - S_MUL(im,t0);
yp1[0] = yr;
yp0[1] = yi;
yp0 += 2;
yp1 -= 2;
}
}
/* De-shuffle the components for the middle of the window only */
{
const kiss_fft_scalar * OPUS_RESTRICT fp1 = f;
const kiss_fft_scalar * OPUS_RESTRICT fp2 = f+N2-1;
kiss_fft_scalar * OPUS_RESTRICT yp = f2;
for(i = 0; i < N4; i++)
{
*yp++ =-*fp1;
*yp++ = *fp2;
fp1 += 2;
fp2 -= 2;
}
}
out -= (N2-overlap)>>1;
/* Mirror on both sides for TDAC */
{
kiss_fft_scalar * OPUS_RESTRICT fp1 = f2+N4-1;
kiss_fft_scalar * OPUS_RESTRICT xp1 = out+N2-1;
kiss_fft_scalar * OPUS_RESTRICT yp1 = out+N4-overlap/2;
kiss_fft_scalar * OPUS_RESTRICT xp1 = out+overlap-1;
kiss_fft_scalar * OPUS_RESTRICT yp1 = out;
const opus_val16 * OPUS_RESTRICT wp1 = window;
const opus_val16 * OPUS_RESTRICT wp2 = window+overlap-1;
for(i = 0; i< N4-overlap/2; i++)
for(i = 0; i < overlap/2; i++)
{
*xp1 = *fp1;
xp1--;
fp1--;
}
for(; i < N4; i++)
{
kiss_fft_scalar x1;
x1 = *fp1--;
*yp1++ +=-MULT16_32_Q15(*wp1, x1);
*xp1-- += MULT16_32_Q15(*wp2, x1);
kiss_fft_scalar x1, x2;
x1 = *xp1;
x2 = *yp1;
*yp1++ = MULT16_32_Q15(*wp2, x2) - MULT16_32_Q15(*wp1, x1);
*xp1-- = MULT16_32_Q15(*wp1, x2) + MULT16_32_Q15(*wp2, x1);
wp1++;
wp2--;
}
}
{
kiss_fft_scalar * OPUS_RESTRICT fp2 = f2+N4;
kiss_fft_scalar * OPUS_RESTRICT xp2 = out+N2;
kiss_fft_scalar * OPUS_RESTRICT yp2 = out+N-1-(N4-overlap/2);
const opus_val16 * OPUS_RESTRICT wp1 = window;
const opus_val16 * OPUS_RESTRICT wp2 = window+overlap-1;
for(i = 0; i< N4-overlap/2; i++)
{
*xp2 = *fp2;
xp2++;
fp2++;
}
for(; i < N4; i++)
{
kiss_fft_scalar x2;
x2 = *fp2++;
*yp2-- = MULT16_32_Q15(*wp1, x2);
*xp2++ = MULT16_32_Q15(*wp2, x2);
wp1++;
wp2--;
}
}
RESTORE_STACK;
}
#endif /* OVERRIDE_clt_mdct_backward */

54
Frameworks/Opus/Opus/opus/celt/mdct.h
View file

@ -53,18 +53,60 @@ typedef struct {
const kiss_twiddle_scalar * OPUS_RESTRICT trig;
} mdct_lookup;
int clt_mdct_init(mdct_lookup *l,int N, int maxshift);
void clt_mdct_clear(mdct_lookup *l);
#if defined(HAVE_ARM_NE10)
#include "arm/mdct_arm.h"
#endif
int clt_mdct_init(mdct_lookup *l,int N, int maxshift, int arch);
void clt_mdct_clear(mdct_lookup *l, int arch);
/** Compute a forward MDCT and scale by 4/N, trashes the input array */
void clt_mdct_forward(const mdct_lookup *l, kiss_fft_scalar *in,
void clt_mdct_forward_c(const mdct_lookup *l, kiss_fft_scalar *in,
kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 *window, int overlap, int shift, int stride);
const opus_val16 *window, int overlap,
int shift, int stride, int arch);
/** Compute a backward MDCT (no scaling) and performs weighted overlap-add
(scales implicitly by 1/2) */
void clt_mdct_backward(const mdct_lookup *l, kiss_fft_scalar *in,
void clt_mdct_backward_c(const mdct_lookup *l, kiss_fft_scalar *in,
kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 * OPUS_RESTRICT window, int overlap, int shift, int stride);
const opus_val16 * OPUS_RESTRICT window,
int overlap, int shift, int stride, int arch);
#if !defined(OVERRIDE_OPUS_MDCT)
/* Is run-time CPU detection enabled on this platform? */
#if defined(OPUS_HAVE_RTCD) && defined(HAVE_ARM_NE10)
extern void (*const CLT_MDCT_FORWARD_IMPL[OPUS_ARCHMASK+1])(
const mdct_lookup *l, kiss_fft_scalar *in,
kiss_fft_scalar * OPUS_RESTRICT out, const opus_val16 *window,
int overlap, int shift, int stride, int arch);
#define clt_mdct_forward(_l, _in, _out, _window, _overlap, _shift, _stride, _arch) \
((*CLT_MDCT_FORWARD_IMPL[(arch)&OPUS_ARCHMASK])(_l, _in, _out, \
_window, _overlap, _shift, \
_stride, _arch))
extern void (*const CLT_MDCT_BACKWARD_IMPL[OPUS_ARCHMASK+1])(
const mdct_lookup *l, kiss_fft_scalar *in,
kiss_fft_scalar * OPUS_RESTRICT out, const opus_val16 *window,
int overlap, int shift, int stride, int arch);
#define clt_mdct_backward(_l, _in, _out, _window, _overlap, _shift, _stride, _arch) \
(*CLT_MDCT_BACKWARD_IMPL[(arch)&OPUS_ARCHMASK])(_l, _in, _out, \
_window, _overlap, _shift, \
_stride, _arch)
#else /* if defined(OPUS_HAVE_RTCD) && defined(HAVE_ARM_NE10) */
#define clt_mdct_forward(_l, _in, _out, _window, _overlap, _shift, _stride, _arch) \
clt_mdct_forward_c(_l, _in, _out, _window, _overlap, _shift, _stride, _arch)
#define clt_mdct_backward(_l, _in, _out, _window, _overlap, _shift, _stride, _arch) \
clt_mdct_backward_c(_l, _in, _out, _window, _overlap, _shift, _stride, _arch)
#endif /* end if defined(OPUS_HAVE_RTCD) && defined(HAVE_ARM_NE10) && !defined(FIXED_POINT) */
#endif /* end if !defined(OVERRIDE_OPUS_MDCT) */
#endif

151
Frameworks/Opus/Opus/opus/celt/mips/celt_mipsr1.h Normal file
View file

@ -0,0 +1,151 @@
/* Copyright (c) 2007-2008 CSIRO
Copyright (c) 2007-2010 Xiph.Org Foundation
Copyright (c) 2008 Gregory Maxwell
Written by Jean-Marc Valin and Gregory Maxwell */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef __CELT_MIPSR1_H__
#define __CELT_MIPSR1_H__
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#define CELT_C
#include "os_support.h"
#include "mdct.h"
#include <math.h>
#include "celt.h"
#include "pitch.h"
#include "bands.h"
#include "modes.h"
#include "entcode.h"
#include "quant_bands.h"
#include "rate.h"
#include "stack_alloc.h"
#include "mathops.h"
#include "float_cast.h"
#include <stdarg.h>
#include "celt_lpc.h"
#include "vq.h"
#define OVERRIDE_comb_filter
void comb_filter(opus_val32 *y, opus_val32 *x, int T0, int T1, int N,
opus_val16 g0, opus_val16 g1, int tapset0, int tapset1,
const opus_val16 *window, int overlap, int arch)
{
int i;
opus_val32 x0, x1, x2, x3, x4;
(void)arch;
/* printf ("%d %d %f %f\n", T0, T1, g0, g1); */
opus_val16 g00, g01, g02, g10, g11, g12;
static const opus_val16 gains[3][3] = {
{QCONST16(0.3066406250f, 15), QCONST16(0.2170410156f, 15), QCONST16(0.1296386719f, 15)},
{QCONST16(0.4638671875f, 15), QCONST16(0.2680664062f, 15), QCONST16(0.f, 15)},
{QCONST16(0.7998046875f, 15), QCONST16(0.1000976562f, 15), QCONST16(0.f, 15)}};
if (g0==0 && g1==0)
{
/* OPT: Happens to work without the OPUS_MOVE(), but only because the current encoder already copies x to y */
if (x!=y)
OPUS_MOVE(y, x, N);
return;
}
g00 = MULT16_16_P15(g0, gains[tapset0][0]);
g01 = MULT16_16_P15(g0, gains[tapset0][1]);
g02 = MULT16_16_P15(g0, gains[tapset0][2]);
g10 = MULT16_16_P15(g1, gains[tapset1][0]);
g11 = MULT16_16_P15(g1, gains[tapset1][1]);
g12 = MULT16_16_P15(g1, gains[tapset1][2]);
x1 = x[-T1+1];
x2 = x[-T1 ];
x3 = x[-T1-1];
x4 = x[-T1-2];
/* If the filter didn't change, we don't need the overlap */
if (g0==g1 && T0==T1 && tapset0==tapset1)
overlap=0;
for (i=0;i<overlap;i++)
{
opus_val16 f;
opus_val32 res;
f = MULT16_16_Q15(window[i],window[i]);
x0= x[i-T1+2];
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)MULT16_16_Q15((Q15ONE-f),g00)), "r" ((int)x[i-T0]));
asm volatile("MADD $ac1, %0, %1" : : "r" ((int)MULT16_16_Q15((Q15ONE-f),g01)), "r" ((int)ADD32(x[i-T0-1],x[i-T0+1])));
asm volatile("MADD $ac1, %0, %1" : : "r" ((int)MULT16_16_Q15((Q15ONE-f),g02)), "r" ((int)ADD32(x[i-T0-2],x[i-T0+2])));
asm volatile("MADD $ac1, %0, %1" : : "r" ((int)MULT16_16_Q15(f,g10)), "r" ((int)x2));
asm volatile("MADD $ac1, %0, %1" : : "r" ((int)MULT16_16_Q15(f,g11)), "r" ((int)ADD32(x3,x1)));
asm volatile("MADD $ac1, %0, %1" : : "r" ((int)MULT16_16_Q15(f,g12)), "r" ((int)ADD32(x4,x0)));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (res): "i" (15));
y[i] = x[i] + res;
x4=x3;
x3=x2;
x2=x1;
x1=x0;
}
x4 = x[i-T1-2];
x3 = x[i-T1-1];
x2 = x[i-T1];
x1 = x[i-T1+1];
if (g1==0)
{
/* OPT: Happens to work without the OPUS_MOVE(), but only because the current encoder already copies x to y */
if (x!=y)
OPUS_MOVE(y+overlap, x+overlap, N-overlap);
return;
}
for (i=overlap;i<N;i++)
{
opus_val32 res;
x0=x[i-T1+2];
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)g10), "r" ((int)x2));
asm volatile("MADD $ac1, %0, %1" : : "r" ((int)g11), "r" ((int)ADD32(x3,x1)));
asm volatile("MADD $ac1, %0, %1" : : "r" ((int)g12), "r" ((int)ADD32(x4,x0)));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (res): "i" (15));
y[i] = x[i] + res;
x4=x3;
x3=x2;
x2=x1;
x1=x0;
}
}
#endif /* __CELT_MIPSR1_H__ */

126
Frameworks/Opus/Opus/opus/celt/mips/fixed_generic_mipsr1.h Normal file
View file

@ -0,0 +1,126 @@
/* Copyright (C) 2007-2009 Xiph.Org Foundation
Copyright (C) 2003-2008 Jean-Marc Valin
Copyright (C) 2007-2008 CSIRO */
/**
@file fixed_generic.h
@brief Generic fixed-point operations
*/
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef CELT_FIXED_GENERIC_MIPSR1_H
#define CELT_FIXED_GENERIC_MIPSR1_H
#undef MULT16_32_Q15_ADD
static inline int MULT16_32_Q15_ADD(int a, int b, int c, int d) {
int m;
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)a), "r" ((int)b));
asm volatile("madd $ac1, %0, %1" : : "r" ((int)c), "r" ((int)d));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (m): "i" (15));
return m;
}
#undef MULT16_32_Q15_SUB
static inline int MULT16_32_Q15_SUB(int a, int b, int c, int d) {
int m;
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)a), "r" ((int)b));
asm volatile("msub $ac1, %0, %1" : : "r" ((int)c), "r" ((int)d));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (m): "i" (15));
return m;
}
#undef MULT16_16_Q15_ADD
static inline int MULT16_16_Q15_ADD(int a, int b, int c, int d) {
int m;
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)a), "r" ((int)b));
asm volatile("madd $ac1, %0, %1" : : "r" ((int)c), "r" ((int)d));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (m): "i" (15));
return m;
}
#undef MULT16_16_Q15_SUB
static inline int MULT16_16_Q15_SUB(int a, int b, int c, int d) {
int m;
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)a), "r" ((int)b));
asm volatile("msub $ac1, %0, %1" : : "r" ((int)c), "r" ((int)d));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (m): "i" (15));
return m;
}
#undef MULT16_32_Q16
static inline int MULT16_32_Q16(int a, int b)
{
int c;
asm volatile("MULT $ac1,%0, %1" : : "r" (a), "r" (b));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (c): "i" (16));
return c;
}
#undef MULT16_32_P16
static inline int MULT16_32_P16(int a, int b)
{
int c;
asm volatile("MULT $ac1, %0, %1" : : "r" (a), "r" (b));
asm volatile("EXTR_R.W %0,$ac1, %1" : "=r" (c): "i" (16));
return c;
}
#undef MULT16_32_Q15
static inline int MULT16_32_Q15(int a, int b)
{
int c;
asm volatile("MULT $ac1, %0, %1" : : "r" (a), "r" (b));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (c): "i" (15));
return c;
}
#undef MULT32_32_Q31
static inline int MULT32_32_Q31(int a, int b)
{
int r;
asm volatile("MULT $ac1, %0, %1" : : "r" (a), "r" (b));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (r): "i" (31));
return r;
}
#undef PSHR32
static inline int PSHR32(int a, int shift)
{
int r;
asm volatile ("SHRAV_R.W %0, %1, %2" :"=r" (r): "r" (a), "r" (shift));
return r;
}
#undef MULT16_16_P15
static inline int MULT16_16_P15(int a, int b)
{
int r;
asm volatile ("mul %0, %1, %2" :"=r" (r): "r" (a), "r" (b));
asm volatile ("SHRA_R.W %0, %1, %2" : "+r" (r): "0" (r), "i"(15));
return r;
}
#endif /* CELT_FIXED_GENERIC_MIPSR1_H */

167
Frameworks/Opus/Opus/opus/celt/mips/kiss_fft_mipsr1.h Normal file
View file

@ -0,0 +1,167 @@
/*Copyright (c) 2013, Xiph.Org Foundation and contributors.
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.*/
#ifndef KISS_FFT_MIPSR1_H
#define KISS_FFT_MIPSR1_H
#if !defined(KISS_FFT_GUTS_H)
#error "This file should only be included from _kiss_fft_guts.h"
#endif
#ifdef FIXED_POINT
#define S_MUL_ADD(a, b, c, d) (S_MUL(a,b)+S_MUL(c,d))
#define S_MUL_SUB(a, b, c, d) (S_MUL(a,b)-S_MUL(c,d))
#undef S_MUL_ADD
static inline int S_MUL_ADD(int a, int b, int c, int d) {
int m;
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)a), "r" ((int)b));
asm volatile("madd $ac1, %0, %1" : : "r" ((int)c), "r" ((int)d));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (m): "i" (15));
return m;
}
#undef S_MUL_SUB
static inline int S_MUL_SUB(int a, int b, int c, int d) {
int m;
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)a), "r" ((int)b));
asm volatile("msub $ac1, %0, %1" : : "r" ((int)c), "r" ((int)d));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (m): "i" (15));
return m;
}
#undef C_MUL
# define C_MUL(m,a,b) (m=C_MUL_fun(a,b))
static inline kiss_fft_cpx C_MUL_fun(kiss_fft_cpx a, kiss_twiddle_cpx b) {
kiss_fft_cpx m;
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)a.r), "r" ((int)b.r));
asm volatile("msub $ac1, %0, %1" : : "r" ((int)a.i), "r" ((int)b.i));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (m.r): "i" (15));
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)a.r), "r" ((int)b.i));
asm volatile("madd $ac1, %0, %1" : : "r" ((int)a.i), "r" ((int)b.r));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (m.i): "i" (15));
return m;
}
#undef C_MULC
# define C_MULC(m,a,b) (m=C_MULC_fun(a,b))
static inline kiss_fft_cpx C_MULC_fun(kiss_fft_cpx a, kiss_twiddle_cpx b) {
kiss_fft_cpx m;
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)a.r), "r" ((int)b.r));
asm volatile("madd $ac1, %0, %1" : : "r" ((int)a.i), "r" ((int)b.i));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (m.r): "i" (15));
asm volatile("MULT $ac1, %0, %1" : : "r" ((int)a.i), "r" ((int)b.r));
asm volatile("msub $ac1, %0, %1" : : "r" ((int)a.r), "r" ((int)b.i));
asm volatile("EXTR.W %0,$ac1, %1" : "=r" (m.i): "i" (15));
return m;
}
#endif /* FIXED_POINT */
#define OVERRIDE_kf_bfly5
static void kf_bfly5(
kiss_fft_cpx * Fout,
const size_t fstride,
const kiss_fft_state *st,
int m,
int N,
int mm
)
{
kiss_fft_cpx *Fout0,*Fout1,*Fout2,*Fout3,*Fout4;
int i, u;
kiss_fft_cpx scratch[13];
const kiss_twiddle_cpx *tw;
kiss_twiddle_cpx ya,yb;
kiss_fft_cpx * Fout_beg = Fout;
#ifdef FIXED_POINT
ya.r = 10126;
ya.i = -31164;
yb.r = -26510;
yb.i = -19261;
#else
ya = st->twiddles[fstride*m];
yb = st->twiddles[fstride*2*m];
#endif
tw=st->twiddles;
for (i=0;i<N;i++)
{
Fout = Fout_beg + i*mm;
Fout0=Fout;
Fout1=Fout0+m;
Fout2=Fout0+2*m;
Fout3=Fout0+3*m;
Fout4=Fout0+4*m;
/* For non-custom modes, m is guaranteed to be a multiple of 4. */
for ( u=0; u<m; ++u ) {
scratch[0] = *Fout0;
C_MUL(scratch[1] ,*Fout1, tw[u*fstride]);
C_MUL(scratch[2] ,*Fout2, tw[2*u*fstride]);
C_MUL(scratch[3] ,*Fout3, tw[3*u*fstride]);
C_MUL(scratch[4] ,*Fout4, tw[4*u*fstride]);
C_ADD( scratch[7],scratch[1],scratch[4]);
C_SUB( scratch[10],scratch[1],scratch[4]);
C_ADD( scratch[8],scratch[2],scratch[3]);
C_SUB( scratch[9],scratch[2],scratch[3]);
Fout0->r += scratch[7].r + scratch[8].r;
Fout0->i += scratch[7].i + scratch[8].i;
scratch[5].r = scratch[0].r + S_MUL_ADD(scratch[7].r,ya.r,scratch[8].r,yb.r);
scratch[5].i = scratch[0].i + S_MUL_ADD(scratch[7].i,ya.r,scratch[8].i,yb.r);
scratch[6].r = S_MUL_ADD(scratch[10].i,ya.i,scratch[9].i,yb.i);
scratch[6].i = -S_MUL_ADD(scratch[10].r,ya.i,scratch[9].r,yb.i);
C_SUB(*Fout1,scratch[5],scratch[6]);
C_ADD(*Fout4,scratch[5],scratch[6]);
scratch[11].r = scratch[0].r + S_MUL_ADD(scratch[7].r,yb.r,scratch[8].r,ya.r);
scratch[11].i = scratch[0].i + S_MUL_ADD(scratch[7].i,yb.r,scratch[8].i,ya.r);
scratch[12].r = S_MUL_SUB(scratch[9].i,ya.i,scratch[10].i,yb.i);
scratch[12].i = S_MUL_SUB(scratch[10].r,yb.i,scratch[9].r,ya.i);
C_ADD(*Fout2,scratch[11],scratch[12]);
C_SUB(*Fout3,scratch[11],scratch[12]);
++Fout0;++Fout1;++Fout2;++Fout3;++Fout4;
}
}
}
#endif /* KISS_FFT_MIPSR1_H */

288
Frameworks/Opus/Opus/opus/celt/mips/mdct_mipsr1.h Normal file
View file

@ -0,0 +1,288 @@
/* Copyright (c) 2007-2008 CSIRO
Copyright (c) 2007-2008 Xiph.Org Foundation
Written by Jean-Marc Valin */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
/* This is a simple MDCT implementation that uses a N/4 complex FFT
to do most of the work. It should be relatively straightforward to
plug in pretty much and FFT here.
This replaces the Vorbis FFT (and uses the exact same API), which
was a bit too messy and that was ending up duplicating code
(might as well use the same FFT everywhere).
The algorithm is similar to (and inspired from) Fabrice Bellard's
MDCT implementation in FFMPEG, but has differences in signs, ordering
and scaling in many places.
*/
#ifndef __MDCT_MIPSR1_H__
#define __MDCT_MIPSR1_H__
#ifndef SKIP_CONFIG_H
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#endif
#include "mdct.h"
#include "kiss_fft.h"
#include "_kiss_fft_guts.h"
#include <math.h>
#include "os_support.h"
#include "mathops.h"
#include "stack_alloc.h"
/* Forward MDCT trashes the input array */
#define OVERRIDE_clt_mdct_forward
void clt_mdct_forward(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 *window, int overlap, int shift, int stride, int arch)
{
int i;
int N, N2, N4;
VARDECL(kiss_fft_scalar, f);
VARDECL(kiss_fft_cpx, f2);
const kiss_fft_state *st = l->kfft[shift];
const kiss_twiddle_scalar *trig;
opus_val16 scale;
#ifdef FIXED_POINT
/* Allows us to scale with MULT16_32_Q16(), which is faster than
MULT16_32_Q15() on ARM. */
int scale_shift = st->scale_shift-1;
#endif
(void)arch;
SAVE_STACK;
scale = st->scale;
N = l->n;
trig = l->trig;
for (i=0;i<shift;i++)
{
N >>= 1;
trig += N;
}
N2 = N>>1;
N4 = N>>2;
ALLOC(f, N2, kiss_fft_scalar);
ALLOC(f2, N4, kiss_fft_cpx);
/* Consider the input to be composed of four blocks: [a, b, c, d] */
/* Window, shuffle, fold */
{
/* Temp pointers to make it really clear to the compiler what we're doing */
const kiss_fft_scalar * OPUS_RESTRICT xp1 = in+(overlap>>1);
const kiss_fft_scalar * OPUS_RESTRICT xp2 = in+N2-1+(overlap>>1);
kiss_fft_scalar * OPUS_RESTRICT yp = f;
const opus_val16 * OPUS_RESTRICT wp1 = window+(overlap>>1);
const opus_val16 * OPUS_RESTRICT wp2 = window+(overlap>>1)-1;
for(i=0;i<((overlap+3)>>2);i++)
{
/* Real part arranged as -d-cR, Imag part arranged as -b+aR*/
*yp++ = S_MUL_ADD(*wp2, xp1[N2],*wp1,*xp2);
*yp++ = S_MUL_SUB(*wp1, *xp1,*wp2, xp2[-N2]);
xp1+=2;
xp2-=2;
wp1+=2;
wp2-=2;
}
wp1 = window;
wp2 = window+overlap-1;
for(;i<N4-((overlap+3)>>2);i++)
{
/* Real part arranged as a-bR, Imag part arranged as -c-dR */
*yp++ = *xp2;
*yp++ = *xp1;
xp1+=2;
xp2-=2;
}
for(;i<N4;i++)
{
/* Real part arranged as a-bR, Imag part arranged as -c-dR */
*yp++ = S_MUL_SUB(*wp2, *xp2, *wp1, xp1[-N2]);
*yp++ = S_MUL_ADD(*wp2, *xp1, *wp1, xp2[N2]);
xp1+=2;
xp2-=2;
wp1+=2;
wp2-=2;
}
}
/* Pre-rotation */
{
kiss_fft_scalar * OPUS_RESTRICT yp = f;
const kiss_twiddle_scalar *t = &trig[0];
for(i=0;i<N4;i++)
{
kiss_fft_cpx yc;
kiss_twiddle_scalar t0, t1;
kiss_fft_scalar re, im, yr, yi;
t0 = t[i];
t1 = t[N4+i];
re = *yp++;
im = *yp++;
yr = S_MUL_SUB(re,t0,im,t1);
yi = S_MUL_ADD(im,t0,re,t1);
yc.r = yr;
yc.i = yi;
yc.r = PSHR32(MULT16_32_Q16(scale, yc.r), scale_shift);
yc.i = PSHR32(MULT16_32_Q16(scale, yc.i), scale_shift);
f2[st->bitrev[i]] = yc;
}
}
/* N/4 complex FFT, does not downscale anymore */
opus_fft_impl(st, f2);
/* Post-rotate */
{
/* Temp pointers to make it really clear to the compiler what we're doing */
const kiss_fft_cpx * OPUS_RESTRICT fp = f2;
kiss_fft_scalar * OPUS_RESTRICT yp1 = out;
kiss_fft_scalar * OPUS_RESTRICT yp2 = out+stride*(N2-1);
const kiss_twiddle_scalar *t = &trig[0];
/* Temp pointers to make it really clear to the compiler what we're doing */
for(i=0;i<N4;i++)
{
kiss_fft_scalar yr, yi;
yr = S_MUL_SUB(fp->i,t[N4+i] , fp->r,t[i]);
yi = S_MUL_ADD(fp->r,t[N4+i] ,fp->i,t[i]);
*yp1 = yr;
*yp2 = yi;
fp++;
yp1 += 2*stride;
yp2 -= 2*stride;
}
}
RESTORE_STACK;
}
#define OVERRIDE_clt_mdct_backward
void clt_mdct_backward(const mdct_lookup *l, kiss_fft_scalar *in, kiss_fft_scalar * OPUS_RESTRICT out,
const opus_val16 * OPUS_RESTRICT window, int overlap, int shift, int stride, int arch)
{
int i;
int N, N2, N4;
const kiss_twiddle_scalar *trig;
(void)arch;
N = l->n;
trig = l->trig;
for (i=0;i<shift;i++)
{
N >>= 1;
trig += N;
}
N2 = N>>1;
N4 = N>>2;
/* Pre-rotate */
{
/* Temp pointers to make it really clear to the compiler what we're doing */
const kiss_fft_scalar * OPUS_RESTRICT xp1 = in;
const kiss_fft_scalar * OPUS_RESTRICT xp2 = in+stride*(N2-1);
kiss_fft_scalar * OPUS_RESTRICT yp = out+(overlap>>1);
const kiss_twiddle_scalar * OPUS_RESTRICT t = &trig[0];
const opus_int16 * OPUS_RESTRICT bitrev = l->kfft[shift]->bitrev;
for(i=0;i<N4;i++)
{
int rev;
kiss_fft_scalar yr, yi;
rev = *bitrev++;
yr = S_MUL_ADD(*xp2, t[i] , *xp1, t[N4+i]);
yi = S_MUL_SUB(*xp1, t[i] , *xp2, t[N4+i]);
/* We swap real and imag because we use an FFT instead of an IFFT. */
yp[2*rev+1] = yr;
yp[2*rev] = yi;
/* Storing the pre-rotation directly in the bitrev order. */
xp1+=2*stride;
xp2-=2*stride;
}
}
opus_fft_impl(l->kfft[shift], (kiss_fft_cpx*)(out+(overlap>>1)));
/* Post-rotate and de-shuffle from both ends of the buffer at once to make
it in-place. */
{
kiss_fft_scalar * OPUS_RESTRICT yp0 = out+(overlap>>1);
kiss_fft_scalar * OPUS_RESTRICT yp1 = out+(overlap>>1)+N2-2;
const kiss_twiddle_scalar *t = &trig[0];
/* Loop to (N4+1)>>1 to handle odd N4. When N4 is odd, the
middle pair will be computed twice. */
for(i=0;i<(N4+1)>>1;i++)
{
kiss_fft_scalar re, im, yr, yi;
kiss_twiddle_scalar t0, t1;
/* We swap real and imag because we're using an FFT instead of an IFFT. */
re = yp0[1];
im = yp0[0];
t0 = t[i];
t1 = t[N4+i];
/* We'd scale up by 2 here, but instead it's done when mixing the windows */
yr = S_MUL_ADD(re,t0 , im,t1);
yi = S_MUL_SUB(re,t1 , im,t0);
/* We swap real and imag because we're using an FFT instead of an IFFT. */
re = yp1[1];
im = yp1[0];
yp0[0] = yr;
yp1[1] = yi;
t0 = t[(N4-i-1)];
t1 = t[(N2-i-1)];
/* We'd scale up by 2 here, but instead it's done when mixing the windows */
yr = S_MUL_ADD(re,t0,im,t1);
yi = S_MUL_SUB(re,t1,im,t0);
yp1[0] = yr;
yp0[1] = yi;
yp0 += 2;
yp1 -= 2;
}
}
/* Mirror on both sides for TDAC */
{
kiss_fft_scalar * OPUS_RESTRICT xp1 = out+overlap-1;
kiss_fft_scalar * OPUS_RESTRICT yp1 = out;
const opus_val16 * OPUS_RESTRICT wp1 = window;
const opus_val16 * OPUS_RESTRICT wp2 = window+overlap-1;
for(i = 0; i < overlap/2; i++)
{
kiss_fft_scalar x1, x2;
x1 = *xp1;
x2 = *yp1;
*yp1++ = MULT16_32_Q15(*wp2, x2) - MULT16_32_Q15(*wp1, x1);
*xp1-- = MULT16_32_Q15(*wp1, x2) + MULT16_32_Q15(*wp2, x1);
wp1++;
wp2--;
}
}
}
#endif /* __MDCT_MIPSR1_H__ */

161
Frameworks/Opus/Opus/opus/celt/mips/pitch_mipsr1.h Normal file
View file

@ -0,0 +1,161 @@
/* Copyright (c) 2007-2008 CSIRO
Copyright (c) 2007-2009 Xiph.Org Foundation
Written by Jean-Marc Valin */
/**
@file pitch.h
@brief Pitch analysis
*/
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef PITCH_MIPSR1_H
#define PITCH_MIPSR1_H
#define OVERRIDE_DUAL_INNER_PROD
static inline void dual_inner_prod(const opus_val16 *x, const opus_val16 *y01, const opus_val16 *y02,
int N, opus_val32 *xy1, opus_val32 *xy2, int arch)
{
int j;
opus_val32 xy01=0;
opus_val32 xy02=0;
(void)arch;
asm volatile("MULT $ac1, $0, $0");
asm volatile("MULT $ac2, $0, $0");
/* Compute the norm of X+Y and X-Y as |X|^2 + |Y|^2 +/- sum(xy) */
for (j=0;j<N;j++)
{
asm volatile("MADD $ac1, %0, %1" : : "r" ((int)x[j]), "r" ((int)y01[j]));
asm volatile("MADD $ac2, %0, %1" : : "r" ((int)x[j]), "r" ((int)y02[j]));
++j;
asm volatile("MADD $ac1, %0, %1" : : "r" ((int)x[j]), "r" ((int)y01[j]));
asm volatile("MADD $ac2, %0, %1" : : "r" ((int)x[j]), "r" ((int)y02[j]));
}
asm volatile ("mflo %0, $ac1": "=r"(xy01));
asm volatile ("mflo %0, $ac2": "=r"(xy02));
*xy1 = xy01;
*xy2 = xy02;
}
static inline void xcorr_kernel_mips(const opus_val16 * x,
const opus_val16 * y, opus_val32 sum[4], int len)
{
int j;
opus_val16 y_0, y_1, y_2, y_3;
opus_int64 sum_0, sum_1, sum_2, sum_3;
sum_0 = (opus_int64)sum[0];
sum_1 = (opus_int64)sum[1];
sum_2 = (opus_int64)sum[2];
sum_3 = (opus_int64)sum[3];
y_3=0; /* gcc doesn't realize that y_3 can't be used uninitialized */
y_0=*y++;
y_1=*y++;
y_2=*y++;
for (j=0;j<len-3;j+=4)
{
opus_val16 tmp;
tmp = *x++;
y_3=*y++;
sum_0 = __builtin_mips_madd( sum_0, tmp, y_0);
sum_1 = __builtin_mips_madd( sum_1, tmp, y_1);
sum_2 = __builtin_mips_madd( sum_2, tmp, y_2);
sum_3 = __builtin_mips_madd( sum_3, tmp, y_3);
tmp=*x++;
y_0=*y++;
sum_0 = __builtin_mips_madd( sum_0, tmp, y_1 );
sum_1 = __builtin_mips_madd( sum_1, tmp, y_2 );
sum_2 = __builtin_mips_madd( sum_2, tmp, y_3);
sum_3 = __builtin_mips_madd( sum_3, tmp, y_0);
tmp=*x++;
y_1=*y++;
sum_0 = __builtin_mips_madd( sum_0, tmp, y_2 );
sum_1 = __builtin_mips_madd( sum_1, tmp, y_3 );
sum_2 = __builtin_mips_madd( sum_2, tmp, y_0);
sum_3 = __builtin_mips_madd( sum_3, tmp, y_1);
tmp=*x++;
y_2=*y++;
sum_0 = __builtin_mips_madd( sum_0, tmp, y_3 );
sum_1 = __builtin_mips_madd( sum_1, tmp, y_0 );
sum_2 = __builtin_mips_madd( sum_2, tmp, y_1);
sum_3 = __builtin_mips_madd( sum_3, tmp, y_2);
}
if (j++<len)
{
opus_val16 tmp = *x++;
y_3=*y++;
sum_0 = __builtin_mips_madd( sum_0, tmp, y_0 );
sum_1 = __builtin_mips_madd( sum_1, tmp, y_1 );
sum_2 = __builtin_mips_madd( sum_2, tmp, y_2);
sum_3 = __builtin_mips_madd( sum_3, tmp, y_3);
}
if (j++<len)
{
opus_val16 tmp=*x++;
y_0=*y++;
sum_0 = __builtin_mips_madd( sum_0, tmp, y_1 );
sum_1 = __builtin_mips_madd( sum_1, tmp, y_2 );
sum_2 = __builtin_mips_madd( sum_2, tmp, y_3);
sum_3 = __builtin_mips_madd( sum_3, tmp, y_0);
}
if (j<len)
{
opus_val16 tmp=*x++;
y_1=*y++;
sum_0 = __builtin_mips_madd( sum_0, tmp, y_2 );
sum_1 = __builtin_mips_madd( sum_1, tmp, y_3 );
sum_2 = __builtin_mips_madd( sum_2, tmp, y_0);
sum_3 = __builtin_mips_madd( sum_3, tmp, y_1);
}
sum[0] = (opus_val32)sum_0;
sum[1] = (opus_val32)sum_1;
sum[2] = (opus_val32)sum_2;
sum[3] = (opus_val32)sum_3;
}
#define OVERRIDE_XCORR_KERNEL
#define xcorr_kernel(x, y, sum, len, arch) \
((void)(arch), xcorr_kernel_mips(x, y, sum, len))
#endif /* PITCH_MIPSR1_H */

125
Frameworks/Opus/Opus/opus/celt/mips/vq_mipsr1.h Normal file
View file

@ -0,0 +1,125 @@
/* Copyright (c) 2007-2008 CSIRO
Copyright (c) 2007-2009 Xiph.Org Foundation
Written by Jean-Marc Valin */
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef __VQ_MIPSR1_H__
#define __VQ_MIPSR1_H__
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "mathops.h"
#include "arch.h"
static unsigned extract_collapse_mask(int *iy, int N, int B);
static void normalise_residual(int * OPUS_RESTRICT iy, celt_norm * OPUS_RESTRICT X, int N, opus_val32 Ryy, opus_val16 gain);
static void exp_rotation(celt_norm *X, int len, int dir, int stride, int K, int spread);
static void renormalise_vector_mips(celt_norm *X, int N, opus_val16 gain, int arch);
#define OVERRIDE_vq_exp_rotation1
static void exp_rotation1(celt_norm *X, int len, int stride, opus_val16 c, opus_val16 s)
{
int i;
opus_val16 ms;
celt_norm *Xptr;
Xptr = X;
ms = NEG16(s);
for (i=0;i<len-stride;i++)
{
celt_norm x1, x2;
x1 = Xptr[0];
x2 = Xptr[stride];
Xptr[stride] = EXTRACT16(PSHR32(MAC16_16(MULT16_16(c, x2), s, x1), 15));
*Xptr++ = EXTRACT16(PSHR32(MAC16_16(MULT16_16(c, x1), ms, x2), 15));
}
Xptr = &X[len-2*stride-1];
for (i=len-2*stride-1;i>=0;i--)
{
celt_norm x1, x2;
x1 = Xptr[0];
x2 = Xptr[stride];
Xptr[stride] = EXTRACT16(PSHR32(MAC16_16(MULT16_16(c, x2), s, x1), 15));
*Xptr-- = EXTRACT16(PSHR32(MAC16_16(MULT16_16(c, x1), ms, x2), 15));
}
}
#define OVERRIDE_renormalise_vector
#define renormalise_vector(X, N, gain, arch) \
(renormalise_vector_mips(X, N, gain, arch))
void renormalise_vector_mips(celt_norm *X, int N, opus_val16 gain, int arch)
{
int i;
#ifdef FIXED_POINT
int k;
#endif
opus_val32 E = EPSILON;
opus_val16 g;
opus_val32 t;
celt_norm *xptr = X;
int X0, X1;
(void)arch;
asm volatile("mult $ac1, $0, $0");
asm volatile("MTLO %0, $ac1" : :"r" (E));
/*if(N %4)
printf("error");*/
for (i=0;i<N-2;i+=2)
{
X0 = (int)*xptr++;
asm volatile("MADD $ac1, %0, %1" : : "r" (X0), "r" (X0));
X1 = (int)*xptr++;
asm volatile("MADD $ac1, %0, %1" : : "r" (X1), "r" (X1));
}
for (;i<N;i++)
{
X0 = (int)*xptr++;
asm volatile("MADD $ac1, %0, %1" : : "r" (X0), "r" (X0));
}
asm volatile("MFLO %0, $ac1" : "=r" (E));
#ifdef FIXED_POINT
k = celt_ilog2(E)>>1;
#endif
t = VSHR32(E, 2*(k-7));
g = MULT16_16_P15(celt_rsqrt_norm(t),gain);
xptr = X;
for (i=0;i<N;i++)
{
*xptr = EXTRACT16(PSHR32(MULT16_16(g, *xptr), k+1));
xptr++;
}
/*return celt_sqrt(E);*/
}
#endif /* __VQ_MIPSR1_H__ */

16
Frameworks/Opus/Opus/opus/celt/modes.c
View file

@ -37,6 +37,7 @@
#include "os_support.h"
#include "stack_alloc.h"
#include "quant_bands.h"
#include "cpu_support.h"
static const opus_int16 eband5ms[] = {
/*0 200 400 600 800 1k 1.2 1.4 1.6 2k 2.4 2.8 3.2 4k 4.8 5.6 6.8 8k 9.6 12k 15.6 */
@ -229,6 +230,7 @@ CELTMode *opus_custom_mode_create(opus_int32 Fs, int frame_size, int *error)
opus_val16 *window;
opus_int16 *logN;
int LM;
int arch = opus_select_arch();
ALLOC_STACK;
#if !defined(VAR_ARRAYS) && !defined(USE_ALLOCA)
if (global_stack==NULL)
@ -345,6 +347,14 @@ CELTMode *opus_custom_mode_create(opus_int32 Fs, int frame_size, int *error)
mode->eBands = compute_ebands(Fs, mode->shortMdctSize, res, &mode->nbEBands);
if (mode->eBands==NULL)
goto failure;
#if !defined(SMALL_FOOTPRINT)
/* Make sure we don't allocate a band larger than our PVQ table.
208 should be enough, but let's be paranoid. */
if ((mode->eBands[mode->nbEBands] - mode->eBands[mode->nbEBands-1])<<LM >
208) {
goto failure;
}
#endif
mode->effEBands = mode->nbEBands;
while (mode->eBands[mode->effEBands] > mode->shortMdctSize)
@ -381,7 +391,7 @@ CELTMode *opus_custom_mode_create(opus_int32 Fs, int frame_size, int *error)
compute_pulse_cache(mode, mode->maxLM);
if (clt_mdct_init(&mode->mdct, 2*mode->shortMdctSize*mode->nbShortMdcts,
mode->maxLM) == 0)
mode->maxLM, arch) == 0)
goto failure;
if (error)
@ -400,6 +410,8 @@ failure:
#ifdef CUSTOM_MODES
void opus_custom_mode_destroy(CELTMode *mode)
{
int arch = opus_select_arch();
if (mode == NULL)
return;
#ifndef CUSTOM_MODES_ONLY
@ -423,7 +435,7 @@ void opus_custom_mode_destroy(CELTMode *mode)
opus_free((opus_int16*)mode->cache.index);
opus_free((unsigned char*)mode->cache.bits);
opus_free((unsigned char*)mode->cache.caps);
clt_mdct_clear(&mode->mdct);
clt_mdct_clear(&mode->mdct, arch);
opus_free((CELTMode *)mode);
}

8
Frameworks/Opus/Opus/opus/celt/modes.h
View file

@ -39,14 +39,6 @@
#define MAX_PERIOD 1024
#ifndef OVERLAP
#define OVERLAP(mode) ((mode)->overlap)
#endif
#ifndef FRAMESIZE
#define FRAMESIZE(mode) ((mode)->mdctSize)
#endif
typedef struct {
int size;
const opus_int16 *index;

15
Frameworks/Opus/Opus/opus/celt/os_support.h
View file

@ -35,13 +35,16 @@
# include "custom_support.h"
#endif
#include "opus_types.h"
#include "opus_defines.h"
#include <string.h>
#include <stdio.h>
#include <stdlib.h>
/** Opus wrapper for malloc(). To do your own dynamic allocation, all you need to do is replace this function and opus_free */
#ifndef OVERRIDE_OPUS_ALLOC
static inline void *opus_alloc (size_t size)
static OPUS_INLINE void *opus_alloc (size_t size)
{
return malloc(size);
}
@ -49,7 +52,7 @@ static inline void *opus_alloc (size_t size)
/** Same as celt_alloc(), except that the area is only needed inside a CELT call (might cause problem with wideband though) */
#ifndef OVERRIDE_OPUS_ALLOC_SCRATCH
static inline void *opus_alloc_scratch (size_t size)
static OPUS_INLINE void *opus_alloc_scratch (size_t size)
{
/* Scratch space doesn't need to be cleared */
return opus_alloc(size);
@ -58,24 +61,24 @@ static inline void *opus_alloc_scratch (size_t size)
/** Opus wrapper for free(). To do your own dynamic allocation, all you need to do is replace this function and opus_alloc */
#ifndef OVERRIDE_OPUS_FREE
static inline void opus_free (void *ptr)
static OPUS_INLINE void opus_free (void *ptr)
{
free(ptr);
}
#endif
/** Copy n bytes of memory from src to dst. The 0* term provides compile-time type checking */
/** Copy n elements from src to dst. The 0* term provides compile-time type checking */
#ifndef OVERRIDE_OPUS_COPY
#define OPUS_COPY(dst, src, n) (memcpy((dst), (src), (n)*sizeof(*(dst)) + 0*((dst)-(src)) ))
#endif
/** Copy n bytes of memory from src to dst, allowing overlapping regions. The 0* term
/** Copy n elements from src to dst, allowing overlapping regions. The 0* term
provides compile-time type checking */
#ifndef OVERRIDE_OPUS_MOVE
#define OPUS_MOVE(dst, src, n) (memmove((dst), (src), (n)*sizeof(*(dst)) + 0*((dst)-(src)) ))
#endif
/** Set n elements of dst to zero, starting at address s */
/** Set n elements of dst to zero */
#ifndef OVERRIDE_OPUS_CLEAR
#define OPUS_CLEAR(dst, n) (memset((dst), 0, (n)*sizeof(*(dst))))
#endif

222
Frameworks/Opus/Opus/opus/celt/pitch.c
View file

@ -102,13 +102,57 @@ static void find_best_pitch(opus_val32 *xcorr, opus_val16 *y, int len,
}
}
static void celt_fir5(const opus_val16 *x,
const opus_val16 *num,
opus_val16 *y,
int N,
opus_val16 *mem)
{
int i;
opus_val16 num0, num1, num2, num3, num4;
opus_val32 mem0, mem1, mem2, mem3, mem4;
num0=num[0];
num1=num[1];
num2=num[2];
num3=num[3];
num4=num[4];
mem0=mem[0];
mem1=mem[1];
mem2=mem[2];
mem3=mem[3];
mem4=mem[4];
for (i=0;i<N;i++)
{
opus_val32 sum = SHL32(EXTEND32(x[i]), SIG_SHIFT);
sum = MAC16_16(sum,num0,mem0);
sum = MAC16_16(sum,num1,mem1);
sum = MAC16_16(sum,num2,mem2);
sum = MAC16_16(sum,num3,mem3);
sum = MAC16_16(sum,num4,mem4);
mem4 = mem3;
mem3 = mem2;
mem2 = mem1;
mem1 = mem0;
mem0 = x[i];
y[i] = ROUND16(sum, SIG_SHIFT);
}
mem[0]=mem0;
mem[1]=mem1;
mem[2]=mem2;
mem[3]=mem3;
mem[4]=mem4;
}
void pitch_downsample(celt_sig * OPUS_RESTRICT x[], opus_val16 * OPUS_RESTRICT x_lp,
int len, int C)
int len, int C, int arch)
{
int i;
opus_val32 ac[5];
opus_val16 tmp=Q15ONE;
opus_val16 lpc[4], mem[4]={0,0,0,0};
opus_val16 lpc[4], mem[5]={0,0,0,0,0};
opus_val16 lpc2[5];
opus_val16 c1 = QCONST16(.8f,15);
#ifdef FIXED_POINT
int shift;
opus_val32 maxabs = celt_maxabs32(x[0], len);
@ -136,7 +180,7 @@ void pitch_downsample(celt_sig * OPUS_RESTRICT x[], opus_val16 * OPUS_RESTRICT x
}
_celt_autocorr(x_lp, ac, NULL, 0,
4, len>>1);
4, len>>1, arch);
/* Noise floor -40 dB */
#ifdef FIXED_POINT
@ -161,16 +205,104 @@ void pitch_downsample(celt_sig * OPUS_RESTRICT x[], opus_val16 * OPUS_RESTRICT x
tmp = MULT16_16_Q15(QCONST16(.9f,15), tmp);
lpc[i] = MULT16_16_Q15(lpc[i], tmp);
}
celt_fir(x_lp, lpc, x_lp, len>>1, 4, mem);
/* Add a zero */
lpc2[0] = lpc[0] + QCONST16(.8f,SIG_SHIFT);
lpc2[1] = lpc[1] + MULT16_16_Q15(c1,lpc[0]);
lpc2[2] = lpc[2] + MULT16_16_Q15(c1,lpc[1]);
lpc2[3] = lpc[3] + MULT16_16_Q15(c1,lpc[2]);
lpc2[4] = MULT16_16_Q15(c1,lpc[3]);
celt_fir5(x_lp, lpc2, x_lp, len>>1, mem);
}
mem[0]=0;
lpc[0]=QCONST16(.8f,12);
celt_fir(x_lp, lpc, x_lp, len>>1, 1, mem);
/* Pure C implementation. */
#ifdef FIXED_POINT
opus_val32
#else
void
#endif
#if defined(OVERRIDE_PITCH_XCORR)
celt_pitch_xcorr_c(const opus_val16 *_x, const opus_val16 *_y,
opus_val32 *xcorr, int len, int max_pitch)
#else
celt_pitch_xcorr(const opus_val16 *_x, const opus_val16 *_y,
opus_val32 *xcorr, int len, int max_pitch, int arch)
#endif
{
#if 0 /* This is a simple version of the pitch correlation that should work
well on DSPs like Blackfin and TI C5x/C6x */
int i, j;
#ifdef FIXED_POINT
opus_val32 maxcorr=1;
#endif
#if !defined(OVERRIDE_PITCH_XCORR)
(void)arch;
#endif
for (i=0;i<max_pitch;i++)
{
opus_val32 sum = 0;
for (j=0;j<len;j++)
sum = MAC16_16(sum, _x[j], _y[i+j]);
xcorr[i] = sum;
#ifdef FIXED_POINT
maxcorr = MAX32(maxcorr, sum);
#endif
}
#ifdef FIXED_POINT
return maxcorr;
#endif
#else /* Unrolled version of the pitch correlation -- runs faster on x86 and ARM */
int i;
/*The EDSP version requires that max_pitch is at least 1, and that _x is
32-bit aligned.
Since it's hard to put asserts in assembly, put them here.*/
#ifdef FIXED_POINT
opus_val32 maxcorr=1;
#endif
celt_assert(max_pitch>0);
celt_assert((((unsigned char *)_x-(unsigned char *)NULL)&3)==0);
for (i=0;i<max_pitch-3;i+=4)
{
opus_val32 sum[4]={0,0,0,0};
#if defined(OVERRIDE_PITCH_XCORR)
xcorr_kernel_c(_x, _y+i, sum, len);
#else
xcorr_kernel(_x, _y+i, sum, len, arch);
#endif
xcorr[i]=sum[0];
xcorr[i+1]=sum[1];
xcorr[i+2]=sum[2];
xcorr[i+3]=sum[3];
#ifdef FIXED_POINT
sum[0] = MAX32(sum[0], sum[1]);
sum[2] = MAX32(sum[2], sum[3]);
sum[0] = MAX32(sum[0], sum[2]);
maxcorr = MAX32(maxcorr, sum[0]);
#endif
}
/* In case max_pitch isn't a multiple of 4, do non-unrolled version. */
for (;i<max_pitch;i++)
{
opus_val32 sum;
#if defined(OVERRIDE_PITCH_XCORR)
sum = celt_inner_prod_c(_x, _y+i, len);
#else
sum = celt_inner_prod(_x, _y+i, len, arch);
#endif
xcorr[i] = sum;
#ifdef FIXED_POINT
maxcorr = MAX32(maxcorr, sum);
#endif
}
#ifdef FIXED_POINT
return maxcorr;
#endif
#endif
}
void pitch_search(const opus_val16 * OPUS_RESTRICT x_lp, opus_val16 * OPUS_RESTRICT y,
int len, int max_pitch, int *pitch)
int len, int max_pitch, int *pitch, int arch)
{
int i, j;
int lag;
@ -179,7 +311,7 @@ void pitch_search(const opus_val16 * OPUS_RESTRICT x_lp, opus_val16 * OPUS_RESTR
VARDECL(opus_val16, y_lp4);
VARDECL(opus_val32, xcorr);
#ifdef FIXED_POINT
opus_val32 maxcorr=1;
opus_val32 maxcorr;
opus_val32 xmax, ymax;
int shift=0;
#endif
@ -220,16 +352,11 @@ void pitch_search(const opus_val16 * OPUS_RESTRICT x_lp, opus_val16 * OPUS_RESTR
/* Coarse search with 4x decimation */
for (i=0;i<max_pitch>>2;i++)
{
opus_val32 sum = 0;
for (j=0;j<len>>2;j++)
sum = MAC16_16(sum, x_lp4[j],y_lp4[i+j]);
xcorr[i] = MAX32(-1, sum);
#ifdef FIXED_POINT
maxcorr = MAX32(maxcorr, sum);
maxcorr =
#endif
}
celt_pitch_xcorr(x_lp4, y_lp4, xcorr, len>>2, max_pitch>>2, arch);
find_best_pitch(xcorr, y_lp4, len>>2, max_pitch>>2, best_pitch
#ifdef FIXED_POINT
, 0, maxcorr
@ -242,12 +369,17 @@ void pitch_search(const opus_val16 * OPUS_RESTRICT x_lp, opus_val16 * OPUS_RESTR
#endif
for (i=0;i<max_pitch>>1;i++)
{
opus_val32 sum=0;
opus_val32 sum;
xcorr[i] = 0;
if (abs(i-2*best_pitch[0])>2 && abs(i-2*best_pitch[1])>2)
continue;
#ifdef FIXED_POINT
sum = 0;
for (j=0;j<len>>1;j++)
sum += SHR32(MULT16_16(x_lp[j],y[i+j]), shift);
#else
sum = celt_inner_prod_c(x_lp, y+i, len>>1);
#endif
xcorr[i] = MAX32(-1, sum);
#ifdef FIXED_POINT
maxcorr = MAX32(maxcorr, sum);
@ -282,16 +414,18 @@ void pitch_search(const opus_val16 * OPUS_RESTRICT x_lp, opus_val16 * OPUS_RESTR
static const int second_check[16] = {0, 0, 3, 2, 3, 2, 5, 2, 3, 2, 3, 2, 5, 2, 3, 2};
opus_val16 remove_doubling(opus_val16 *x, int maxperiod, int minperiod,
int N, int *T0_, int prev_period, opus_val16 prev_gain)
int N, int *T0_, int prev_period, opus_val16 prev_gain, int arch)
{
int k, i, T, T0;
opus_val16 g, g0;
opus_val16 pg;
opus_val32 xy,xx,yy;
opus_val32 xy,xx,yy,xy2;
opus_val32 xcorr[3];
opus_val32 best_xy, best_yy;
int offset;
int minperiod0;
VARDECL(opus_val32, yy_lookup);
SAVE_STACK;
minperiod0 = minperiod;
maxperiod /= 2;
@ -304,13 +438,16 @@ opus_val16 remove_doubling(opus_val16 *x, int maxperiod, int minperiod,
*T0_=maxperiod-1;
T = T0 = *T0_;
xx=xy=yy=0;
for (i=0;i<N;i++)
ALLOC(yy_lookup, maxperiod+1, opus_val32);
dual_inner_prod(x, x, x-T0, N, &xx, &xy, arch);
yy_lookup[0] = xx;
yy=xx;
for (i=1;i<=maxperiod;i++)
{
xy = MAC16_16(xy, x[i], x[i-T0]);
xx = MAC16_16(xx, x[i], x[i]);
yy = MAC16_16(yy, x[i-T0],x[i-T0]);
yy = yy+MULT16_16(x[-i],x[-i])-MULT16_16(x[N-i],x[N-i]);
yy_lookup[i] = MAX32(0, yy);
}
yy = yy_lookup[T0];
best_xy = xy;
best_yy = yy;
#ifdef FIXED_POINT
@ -331,7 +468,8 @@ opus_val16 remove_doubling(opus_val16 *x, int maxperiod, int minperiod,
int T1, T1b;
opus_val16 g1;
opus_val16 cont=0;
T1 = (2*T0+k)/(2*k);
opus_val16 thresh;
T1 = celt_udiv(2*T0+k, 2*k);
if (T1 < minperiod)
break;
/* Look for another strong correlation at T1b */
@ -343,17 +481,11 @@ opus_val16 remove_doubling(opus_val16 *x, int maxperiod, int minperiod,
T1b = T0+T1;
} else
{
T1b = (2*second_check[k]*T0+k)/(2*k);
}
xy=yy=0;
for (i=0;i<N;i++)
{
xy = MAC16_16(xy, x[i], x[i-T1]);
yy = MAC16_16(yy, x[i-T1], x[i-T1]);
xy = MAC16_16(xy, x[i], x[i-T1b]);
yy = MAC16_16(yy, x[i-T1b], x[i-T1b]);
T1b = celt_udiv(2*second_check[k]*T0+k, 2*k);
}
dual_inner_prod(x, &x[-T1], &x[-T1b], N, &xy, &xy2, arch);
xy += xy2;
yy = yy_lookup[T1] + yy_lookup[T1b];
#ifdef FIXED_POINT
{
opus_val32 x2y2;
@ -372,7 +504,14 @@ opus_val16 remove_doubling(opus_val16 *x, int maxperiod, int minperiod,
cont = HALF32(prev_gain);
else
cont = 0;
if (g1 > QCONST16(.3f,15) + MULT16_16_Q15(QCONST16(.4f,15),g0)-cont)
thresh = MAX16(QCONST16(.3f,15), MULT16_16_Q15(QCONST16(.7f,15),g0)-cont);
/* Bias against very high pitch (very short period) to avoid false-positives
due to short-term correlation */
if (T1<3*minperiod)
thresh = MAX16(QCONST16(.4f,15), MULT16_16_Q15(QCONST16(.85f,15),g0)-cont);
else if (T1<2*minperiod)
thresh = MAX16(QCONST16(.5f,15), MULT16_16_Q15(QCONST16(.9f,15),g0)-cont);
if (g1 > thresh)
{
best_xy = xy;
best_yy = yy;
@ -387,13 +526,7 @@ opus_val16 remove_doubling(opus_val16 *x, int maxperiod, int minperiod,
pg = SHR32(frac_div32(best_xy,best_yy+1),16);
for (k=0;k<3;k++)
{
int T1 = T+k-1;
xy = 0;
for (i=0;i<N;i++)
xy = MAC16_16(xy, x[i], x[i-T1]);
xcorr[k] = xy;
}
xcorr[k] = celt_inner_prod(x, x-(T+k-1), N, arch);
if ((xcorr[2]-xcorr[0]) > MULT16_32_Q15(QCONST16(.7f,15),xcorr[1]-xcorr[0]))
offset = 1;
else if ((xcorr[0]-xcorr[2]) > MULT16_32_Q15(QCONST16(.7f,15),xcorr[1]-xcorr[2]))
@ -406,5 +539,6 @@ opus_val16 remove_doubling(opus_val16 *x, int maxperiod, int minperiod,
if (*T0_<minperiod0)
*T0_=minperiod0;
RESTORE_STACK;
return pg;
}

178
Frameworks/Opus/Opus/opus/celt/pitch.h
View file

@ -35,14 +35,186 @@
#define PITCH_H
#include "modes.h"
#include "cpu_support.h"
#if (defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT)) \
|| ((defined(OPUS_X86_MAY_HAVE_SSE4_1) || defined(OPUS_X86_MAY_HAVE_SSE2)) && defined(FIXED_POINT))
#include "x86/pitch_sse.h"
#endif
#if defined(MIPSr1_ASM)
#include "mips/pitch_mipsr1.h"
#endif
#if ((defined(OPUS_ARM_ASM) && defined(FIXED_POINT)) \
|| defined(OPUS_ARM_MAY_HAVE_NEON_INTR))
# include "arm/pitch_arm.h"
#endif
void pitch_downsample(celt_sig * OPUS_RESTRICT x[], opus_val16 * OPUS_RESTRICT x_lp,
int len, int C);
int len, int C, int arch);
void pitch_search(const opus_val16 * OPUS_RESTRICT x_lp, opus_val16 * OPUS_RESTRICT y,
int len, int max_pitch, int *pitch);
int len, int max_pitch, int *pitch, int arch);
opus_val16 remove_doubling(opus_val16 *x, int maxperiod, int minperiod,
int N, int *T0, int prev_period, opus_val16 prev_gain);
int N, int *T0, int prev_period, opus_val16 prev_gain, int arch);
/* OPT: This is the kernel you really want to optimize. It gets used a lot
by the prefilter and by the PLC. */
static OPUS_INLINE void xcorr_kernel_c(const opus_val16 * x, const opus_val16 * y, opus_val32 sum[4], int len)
{
int j;
opus_val16 y_0, y_1, y_2, y_3;
celt_assert(len>=3);
y_3=0; /* gcc doesn't realize that y_3 can't be used uninitialized */
y_0=*y++;
y_1=*y++;
y_2=*y++;
for (j=0;j<len-3;j+=4)
{
opus_val16 tmp;
tmp = *x++;
y_3=*y++;
sum[0] = MAC16_16(sum[0],tmp,y_0);
sum[1] = MAC16_16(sum[1],tmp,y_1);
sum[2] = MAC16_16(sum[2],tmp,y_2);
sum[3] = MAC16_16(sum[3],tmp,y_3);
tmp=*x++;
y_0=*y++;
sum[0] = MAC16_16(sum[0],tmp,y_1);
sum[1] = MAC16_16(sum[1],tmp,y_2);
sum[2] = MAC16_16(sum[2],tmp,y_3);
sum[3] = MAC16_16(sum[3],tmp,y_0);
tmp=*x++;
y_1=*y++;
sum[0] = MAC16_16(sum[0],tmp,y_2);
sum[1] = MAC16_16(sum[1],tmp,y_3);
sum[2] = MAC16_16(sum[2],tmp,y_0);
sum[3] = MAC16_16(sum[3],tmp,y_1);
tmp=*x++;
y_2=*y++;
sum[0] = MAC16_16(sum[0],tmp,y_3);
sum[1] = MAC16_16(sum[1],tmp,y_0);
sum[2] = MAC16_16(sum[2],tmp,y_1);
sum[3] = MAC16_16(sum[3],tmp,y_2);
}
if (j++<len)
{
opus_val16 tmp = *x++;
y_3=*y++;
sum[0] = MAC16_16(sum[0],tmp,y_0);
sum[1] = MAC16_16(sum[1],tmp,y_1);
sum[2] = MAC16_16(sum[2],tmp,y_2);
sum[3] = MAC16_16(sum[3],tmp,y_3);
}
if (j++<len)
{
opus_val16 tmp=*x++;
y_0=*y++;
sum[0] = MAC16_16(sum[0],tmp,y_1);
sum[1] = MAC16_16(sum[1],tmp,y_2);
sum[2] = MAC16_16(sum[2],tmp,y_3);
sum[3] = MAC16_16(sum[3],tmp,y_0);
}
if (j<len)
{
opus_val16 tmp=*x++;
y_1=*y++;
sum[0] = MAC16_16(sum[0],tmp,y_2);
sum[1] = MAC16_16(sum[1],tmp,y_3);
sum[2] = MAC16_16(sum[2],tmp,y_0);
sum[3] = MAC16_16(sum[3],tmp,y_1);
}
}
#ifndef OVERRIDE_XCORR_KERNEL
#define xcorr_kernel(x, y, sum, len, arch) \
((void)(arch),xcorr_kernel_c(x, y, sum, len))
#endif /* OVERRIDE_XCORR_KERNEL */
static OPUS_INLINE void dual_inner_prod_c(const opus_val16 *x, const opus_val16 *y01, const opus_val16 *y02,
int N, opus_val32 *xy1, opus_val32 *xy2)
{
int i;
opus_val32 xy01=0;
opus_val32 xy02=0;
for (i=0;i<N;i++)
{
xy01 = MAC16_16(xy01, x[i], y01[i]);
xy02 = MAC16_16(xy02, x[i], y02[i]);
}
*xy1 = xy01;
*xy2 = xy02;
}
#ifndef OVERRIDE_DUAL_INNER_PROD
# define dual_inner_prod(x, y01, y02, N, xy1, xy2, arch) \
((void)(arch),dual_inner_prod_c(x, y01, y02, N, xy1, xy2))
#endif
/*We make sure a C version is always available for cases where the overhead of
vectorization and passing around an arch flag aren't worth it.*/
static OPUS_INLINE opus_val32 celt_inner_prod_c(const opus_val16 *x,
const opus_val16 *y, int N)
{
int i;
opus_val32 xy=0;
for (i=0;i<N;i++)
xy = MAC16_16(xy, x[i], y[i]);
return xy;
}
#if !defined(OVERRIDE_CELT_INNER_PROD)
# define celt_inner_prod(x, y, N, arch) \
((void)(arch),celt_inner_prod_c(x, y, N))
#endif
#ifdef NON_STATIC_COMB_FILTER_CONST_C
void comb_filter_const_c(opus_val32 *y, opus_val32 *x, int T, int N,
opus_val16 g10, opus_val16 g11, opus_val16 g12);
#endif
#ifdef FIXED_POINT
opus_val32
#else
void
#endif
celt_pitch_xcorr_c(const opus_val16 *_x, const opus_val16 *_y,
opus_val32 *xcorr, int len, int max_pitch);
#if !defined(OVERRIDE_PITCH_XCORR)
/*Is run-time CPU detection enabled on this platform?*/
# if defined(OPUS_HAVE_RTCD) && (defined(OPUS_ARM_ASM) \
|| (defined(OPUS_ARM_MAY_HAVE_NEON_INTR) \
&& !defined(OPUS_ARM_PRESUME_NEON_INTR)))
extern
# if defined(FIXED_POINT)
opus_val32
# else
void
# endif
(*const CELT_PITCH_XCORR_IMPL[OPUS_ARCHMASK+1])(const opus_val16 *,
const opus_val16 *, opus_val32 *, int, int);
# define OVERRIDE_PITCH_XCORR
# define celt_pitch_xcorr(_x, _y, xcorr, len, max_pitch, arch) \
((*CELT_PITCH_XCORR_IMPL[(arch)&OPUS_ARCHMASK])(_x, _y, \
xcorr, len, max_pitch))
# else
#ifdef FIXED_POINT
opus_val32
#else
void
#endif
celt_pitch_xcorr(const opus_val16 *_x, const opus_val16 *_y,
opus_val32 *xcorr, int len, int max_pitch, int arch);
# endif
#endif
#endif

50
Frameworks/Opus/Opus/opus/celt/quant_bands.c
View file

@ -40,8 +40,8 @@
#include "rate.h"
#ifdef FIXED_POINT
/* Mean energy in each band quantized in Q6 */
static const signed char eMeans[25] = {
/* Mean energy in each band quantized in Q4 */
const signed char eMeans[25] = {
103,100, 92, 85, 81,
77, 72, 70, 78, 75,
73, 71, 78, 74, 69,
@ -49,8 +49,8 @@ static const signed char eMeans[25] = {
60, 60, 60, 60, 60
};
#else
/* Mean energy in each band quantized in Q6 and converted back to float */
static const opus_val16 eMeans[25] = {
/* Mean energy in each band quantized in Q4 and converted back to float */
const opus_val16 eMeans[25] = {
6.437500f, 6.250000f, 5.750000f, 5.312500f, 5.062500f,
4.812500f, 4.500000f, 4.375000f, 4.875000f, 4.687500f,
4.562500f, 4.437500f, 4.875000f, 4.625000f, 4.312500f,
@ -157,7 +157,7 @@ static int quant_coarse_energy_impl(const CELTMode *m, int start, int end,
const opus_val16 *eBands, opus_val16 *oldEBands,
opus_int32 budget, opus_int32 tell,
const unsigned char *prob_model, opus_val16 *error, ec_enc *enc,
int C, int LM, int intra, opus_val16 max_decay)
int C, int LM, int intra, opus_val16 max_decay, int lfe)
{
int i, c;
int badness = 0;
@ -222,6 +222,8 @@ static int quant_coarse_energy_impl(const CELTMode *m, int start, int end,
if (bits_left < 16)
qi = IMAX(-1, qi);
}
if (lfe && i>=2)
qi = IMIN(qi, 0);
if (budget-tell >= 15)
{
int pi;
@ -253,13 +255,13 @@ static int quant_coarse_energy_impl(const CELTMode *m, int start, int end,
prev[c] = prev[c] + SHL32(q,7) - MULT16_16(beta,PSHR32(q,8));
} while (++c < C);
}
return badness;
return lfe ? 0 : badness;
}
void quant_coarse_energy(const CELTMode *m, int start, int end, int effEnd,
const opus_val16 *eBands, opus_val16 *oldEBands, opus_uint32 budget,
opus_val16 *error, ec_enc *enc, int C, int LM, int nbAvailableBytes,
int force_intra, opus_val32 *delayedIntra, int two_pass, int loss_rate)
int force_intra, opus_val32 *delayedIntra, int two_pass, int loss_rate, int lfe)
{
int intra;
opus_val16 max_decay;
@ -280,9 +282,6 @@ void quant_coarse_energy(const CELTMode *m, int start, int end, int effEnd,
if (tell+3 > budget)
two_pass = intra = 0;
/* Encode the global flags using a simple probability model
(first symbols in the stream) */
max_decay = QCONST16(16.f,DB_SHIFT);
if (end-start>10)
{
@ -292,6 +291,8 @@ void quant_coarse_energy(const CELTMode *m, int start, int end, int effEnd,
max_decay = MIN32(max_decay, .125f*nbAvailableBytes);
#endif
}
if (lfe)
max_decay=3;
enc_start_state = *enc;
ALLOC(oldEBands_intra, C*m->nbEBands, opus_val16);
@ -301,7 +302,7 @@ void quant_coarse_energy(const CELTMode *m, int start, int end, int effEnd,
if (two_pass || intra)
{
badness1 = quant_coarse_energy_impl(m, start, end, eBands, oldEBands_intra, budget,
tell, e_prob_model[LM][1], error_intra, enc, C, LM, 1, max_decay);
tell, e_prob_model[LM][1], error_intra, enc, C, LM, 1, max_decay, lfe);
}
if (!intra)
@ -311,6 +312,7 @@ void quant_coarse_energy(const CELTMode *m, int start, int end, int effEnd,
opus_int32 tell_intra;
opus_uint32 nstart_bytes;
opus_uint32 nintra_bytes;
opus_uint32 save_bytes;
int badness2;
VARDECL(unsigned char, intra_bits);
@ -321,14 +323,17 @@ void quant_coarse_energy(const CELTMode *m, int start, int end, int effEnd,
nstart_bytes = ec_range_bytes(&enc_start_state);
nintra_bytes = ec_range_bytes(&enc_intra_state);
intra_buf = ec_get_buffer(&enc_intra_state) + nstart_bytes;
ALLOC(intra_bits, nintra_bytes-nstart_bytes, unsigned char);
save_bytes = nintra_bytes-nstart_bytes;
if (save_bytes == 0)
save_bytes = ALLOC_NONE;
ALLOC(intra_bits, save_bytes, unsigned char);
/* Copy bits from intra bit-stream */
OPUS_COPY(intra_bits, intra_buf, nintra_bytes - nstart_bytes);
*enc = enc_start_state;
badness2 = quant_coarse_energy_impl(m, start, end, eBands, oldEBands, budget,
tell, e_prob_model[LM][intra], error, enc, C, LM, 0, max_decay);
tell, e_prob_model[LM][intra], error, enc, C, LM, 0, max_decay, lfe);
if (two_pass && (badness1 < badness2 || (badness1 == badness2 && ((opus_int32)ec_tell_frac(enc))+intra_bias > tell_intra)))
{
@ -535,25 +540,6 @@ void unquant_energy_finalise(const CELTMode *m, int start, int end, opus_val16 *
}
}
void log2Amp(const CELTMode *m, int start, int end,
celt_ener *eBands, const opus_val16 *oldEBands, int C)
{
int c, i;
c=0;
do {
for (i=0;i<start;i++)
eBands[i+c*m->nbEBands] = 0;
for (;i<end;i++)
{
opus_val16 lg = ADD16(oldEBands[i+c*m->nbEBands],
SHL16((opus_val16)eMeans[i],6));
eBands[i+c*m->nbEBands] = PSHR32(celt_exp2(lg),4);
}
for (;i<m->nbEBands;i++)
eBands[i+c*m->nbEBands] = 0;
} while (++c < C);
}
void amp2Log2(const CELTMode *m, int effEnd, int end,
celt_ener *bandE, opus_val16 *bandLogE, int C)
{

8
Frameworks/Opus/Opus/opus/celt/quant_bands.h
View file

@ -35,6 +35,12 @@
#include "entdec.h"
#include "mathops.h"
#ifdef FIXED_POINT
extern const signed char eMeans[25];
#else
extern const opus_val16 eMeans[25];
#endif
void amp2Log2(const CELTMode *m, int effEnd, int end,
celt_ener *bandE, opus_val16 *bandLogE, int C);
@ -45,7 +51,7 @@ void quant_coarse_energy(const CELTMode *m, int start, int end, int effEnd,
const opus_val16 *eBands, opus_val16 *oldEBands, opus_uint32 budget,
opus_val16 *error, ec_enc *enc, int C, int LM,
int nbAvailableBytes, int force_intra, opus_val32 *delayedIntra,
int two_pass, int loss_rate);
int two_pass, int loss_rate, int lfe);
void quant_fine_energy(const CELTMode *m, int start, int end, opus_val16 *oldEBands, opus_val16 *error, int *fine_quant, ec_enc *enc, int C);

17
Frameworks/Opus/Opus/opus/celt/rate.c
View file

@ -245,10 +245,10 @@ void compute_pulse_cache(CELTMode *m, int LM)
#define ALLOC_STEPS 6
static inline int interp_bits2pulses(const CELTMode *m, int start, int end, int skip_start,
static OPUS_INLINE int interp_bits2pulses(const CELTMode *m, int start, int end, int skip_start,
const int *bits1, const int *bits2, const int *thresh, const int *cap, opus_int32 total, opus_int32 *_balance,
int skip_rsv, int *intensity, int intensity_rsv, int *dual_stereo, int dual_stereo_rsv, int *bits,
int *ebits, int *fine_priority, int C, int LM, ec_ctx *ec, int encode, int prev)
int *ebits, int *fine_priority, int C, int LM, ec_ctx *ec, int encode, int prev, int signalBandwidth)
{
opus_int32 psum;
int lo, hi;
@ -333,7 +333,7 @@ static inline int interp_bits2pulses(const CELTMode *m, int start, int end, int
/*Figure out how many left-over bits we would be adding to this band.
This can include bits we've stolen back from higher, skipped bands.*/
left = total-psum;
percoeff = left/(m->eBands[codedBands]-m->eBands[start]);
percoeff = celt_udiv(left, m->eBands[codedBands]-m->eBands[start]);
left -= (m->eBands[codedBands]-m->eBands[start])*percoeff;
rem = IMAX(left-(m->eBands[j]-m->eBands[start]),0);
band_width = m->eBands[codedBands]-m->eBands[j];
@ -353,7 +353,7 @@ static inline int interp_bits2pulses(const CELTMode *m, int start, int end, int
#ifdef FUZZING
if ((rand()&0x1) == 0)
#else
if (codedBands<=start+2 || band_bits > ((j<prev?7:9)*band_width<<LM<<BITRES)>>4)
if (codedBands<=start+2 || (band_bits > ((j<prev?7:9)*band_width<<LM<<BITRES)>>4 && j<=signalBandwidth))
#endif
{
ec_enc_bit_logp(ec, 1, 1);
@ -414,7 +414,7 @@ static inline int interp_bits2pulses(const CELTMode *m, int start, int end, int
/* Allocate the remaining bits */
left = total-psum;
percoeff = left/(m->eBands[codedBands]-m->eBands[start]);
percoeff = celt_udiv(left, m->eBands[codedBands]-m->eBands[start]);
left -= (m->eBands[codedBands]-m->eBands[start])*percoeff;
for (j=start;j<codedBands;j++)
bits[j] += ((int)percoeff*(m->eBands[j+1]-m->eBands[j]));
@ -465,7 +465,8 @@ static inline int interp_bits2pulses(const CELTMode *m, int start, int end, int
offset += NClogN>>3;
/* Divide with rounding */
ebits[j] = IMAX(0, (bits[j] + offset + (den<<(BITRES-1))) / (den<<BITRES));
ebits[j] = IMAX(0, (bits[j] + offset + (den<<(BITRES-1))));
ebits[j] = celt_udiv(ebits[j], den)>>BITRES;
/* Make sure not to bust */
if (C*ebits[j] > (bits[j]>>BITRES))
@ -524,7 +525,7 @@ static inline int interp_bits2pulses(const CELTMode *m, int start, int end, int
}
int compute_allocation(const CELTMode *m, int start, int end, const int *offsets, const int *cap, int alloc_trim, int *intensity, int *dual_stereo,
opus_int32 total, opus_int32 *balance, int *pulses, int *ebits, int *fine_priority, int C, int LM, ec_ctx *ec, int encode, int prev)
opus_int32 total, opus_int32 *balance, int *pulses, int *ebits, int *fine_priority, int C, int LM, ec_ctx *ec, int encode, int prev, int signalBandwidth)
{
int lo, hi, len, j;
int codedBands;
@ -631,7 +632,7 @@ int compute_allocation(const CELTMode *m, int start, int end, const int *offsets
}
codedBands = interp_bits2pulses(m, start, end, skip_start, bits1, bits2, thresh, cap,
total, balance, skip_rsv, intensity, intensity_rsv, dual_stereo, dual_stereo_rsv,
pulses, ebits, fine_priority, C, LM, ec, encode, prev);
pulses, ebits, fine_priority, C, LM, ec, encode, prev, signalBandwidth);
RESTORE_STACK;
return codedBands;
}

8
Frameworks/Opus/Opus/opus/celt/rate.h
View file

@ -45,12 +45,12 @@
void compute_pulse_cache(CELTMode *m, int LM);
static inline int get_pulses(int i)
static OPUS_INLINE int get_pulses(int i)
{
return i<8 ? i : (8 + (i&7)) << ((i>>3)-1);
}
static inline int bits2pulses(const CELTMode *m, int band, int LM, int bits)
static OPUS_INLINE int bits2pulses(const CELTMode *m, int band, int LM, int bits)
{
int i;
int lo, hi;
@ -77,7 +77,7 @@ static inline int bits2pulses(const CELTMode *m, int band, int LM, int bits)
return hi;
}
static inline int pulses2bits(const CELTMode *m, int band, int LM, int pulses)
static OPUS_INLINE int pulses2bits(const CELTMode *m, int band, int LM, int pulses)
{
const unsigned char *cache;
@ -96,6 +96,6 @@ static inline int pulses2bits(const CELTMode *m, int band, int LM, int pulses)
@return Total number of bits allocated
*/
int compute_allocation(const CELTMode *m, int start, int end, const int *offsets, const int *cap, int alloc_trim, int *intensity, int *dual_stero,
opus_int32 total, opus_int32 *balance, int *pulses, int *ebits, int *fine_priority, int C, int LM, ec_ctx *ec, int encode, int prev);
opus_int32 total, opus_int32 *balance, int *pulses, int *ebits, int *fine_priority, int C, int LM, ec_ctx *ec, int encode, int prev, int signalBandwidth);
#endif

37
Frameworks/Opus/Opus/opus/celt/stack_alloc.h
View file

@ -32,6 +32,9 @@
#ifndef STACK_ALLOC_H
#define STACK_ALLOC_H
#include "opus_types.h"
#include "opus_defines.h"
#if (!defined (VAR_ARRAYS) && !defined (USE_ALLOCA) && !defined (NONTHREADSAFE_PSEUDOSTACK))
#error "Opus requires one of VAR_ARRAYS, USE_ALLOCA, or NONTHREADSAFE_PSEUDOSTACK be defined to select the temporary allocation mode."
#endif
@ -92,6 +95,8 @@
#define SAVE_STACK
#define RESTORE_STACK
#define ALLOC_STACK
/* C99 does not allow VLAs of size zero */
#define ALLOC_NONE 1
#elif defined(USE_ALLOCA)
@ -106,13 +111,16 @@
#define SAVE_STACK
#define RESTORE_STACK
#define ALLOC_STACK
#define ALLOC_NONE 0
#else
#ifdef CELT_C
char *scratch_ptr=0;
char *global_stack=0;
#else
extern char *global_stack;
extern char *scratch_ptr;
#endif /* CELT_C */
#ifdef ENABLE_VALGRIND
@ -134,8 +142,12 @@ extern char *global_stack_top;
#define ALIGN(stack, size) ((stack) += ((size) - (long)(stack)) & ((size) - 1))
#define PUSH(stack, size, type) (ALIGN((stack),sizeof(type)/sizeof(char)),(stack)+=(size)*(sizeof(type)/sizeof(char)),(type*)((stack)-(size)*(sizeof(type)/sizeof(char))))
#if 0 /* Set this to 1 to instrument pseudostack usage */
#define RESTORE_STACK (printf("%ld %s:%d\n", global_stack-scratch_ptr, __FILE__, __LINE__),global_stack = _saved_stack)
#else
#define RESTORE_STACK (global_stack = _saved_stack)
#define ALLOC_STACK char *_saved_stack; (global_stack = (global_stack==0) ? opus_alloc_scratch(GLOBAL_STACK_SIZE) : global_stack); _saved_stack = global_stack;
#endif
#define ALLOC_STACK char *_saved_stack; (global_stack = (global_stack==0) ? (scratch_ptr=opus_alloc_scratch(GLOBAL_STACK_SIZE)) : global_stack); _saved_stack = global_stack;
#endif /* ENABLE_VALGRIND */
@ -143,7 +155,30 @@ extern char *global_stack_top;
#define VARDECL(type, var) type *var
#define ALLOC(var, size, type) var = PUSH(global_stack, size, type)
#define SAVE_STACK char *_saved_stack = global_stack;
#define ALLOC_NONE 0
#endif /* VAR_ARRAYS */
#ifdef ENABLE_VALGRIND
#include <valgrind/memcheck.h>
#define OPUS_CHECK_ARRAY(ptr, len) VALGRIND_CHECK_MEM_IS_DEFINED(ptr, len*sizeof(*ptr))
#define OPUS_CHECK_VALUE(value) VALGRIND_CHECK_VALUE_IS_DEFINED(value)
#define OPUS_CHECK_ARRAY_COND(ptr, len) VALGRIND_CHECK_MEM_IS_DEFINED(ptr, len*sizeof(*ptr))
#define OPUS_CHECK_VALUE_COND(value) VALGRIND_CHECK_VALUE_IS_DEFINED(value)
#define OPUS_PRINT_INT(value) do {fprintf(stderr, #value " = %d at %s:%d\n", value, __FILE__, __LINE__);}while(0)
#define OPUS_FPRINTF fprintf
#else
static OPUS_INLINE int _opus_false(void) {return 0;}
#define OPUS_CHECK_ARRAY(ptr, len) _opus_false()
#define OPUS_CHECK_VALUE(value) _opus_false()
#define OPUS_PRINT_INT(value) do{}while(0)
#define OPUS_FPRINTF (void)
#endif
#endif /* STACK_ALLOC_H */

621
Frameworks/Opus/Opus/opus/celt/static_modes_fixed.h
View file

@ -4,6 +4,11 @@
#include "modes.h"
#include "rate.h"
#ifdef HAVE_ARM_NE10
#define OVERRIDE_FFT 1
#include "static_modes_fixed_arm_ne10.h"
#endif
#ifndef DEF_WINDOW120
#define DEF_WINDOW120
static const opus_val16 window120[120] = {
@ -341,84 +346,84 @@ static const kiss_twiddle_cpx fft_twiddles48000_960[480] = {
#ifndef FFT_BITREV480
#define FFT_BITREV480
static const opus_int16 fft_bitrev480[480] = {
0, 120, 240, 360, 30, 150, 270, 390, 60, 180, 300, 420, 90, 210, 330,
450, 15, 135, 255, 375, 45, 165, 285, 405, 75, 195, 315, 435, 105, 225,
345, 465, 5, 125, 245, 365, 35, 155, 275, 395, 65, 185, 305, 425, 95,
215, 335, 455, 20, 140, 260, 380, 50, 170, 290, 410, 80, 200, 320, 440,
110, 230, 350, 470, 10, 130, 250, 370, 40, 160, 280, 400, 70, 190, 310,
430, 100, 220, 340, 460, 25, 145, 265, 385, 55, 175, 295, 415, 85, 205,
325, 445, 115, 235, 355, 475, 1, 121, 241, 361, 31, 151, 271, 391, 61,
181, 301, 421, 91, 211, 331, 451, 16, 136, 256, 376, 46, 166, 286, 406,
76, 196, 316, 436, 106, 226, 346, 466, 6, 126, 246, 366, 36, 156, 276,
396, 66, 186, 306, 426, 96, 216, 336, 456, 21, 141, 261, 381, 51, 171,
291, 411, 81, 201, 321, 441, 111, 231, 351, 471, 11, 131, 251, 371, 41,
161, 281, 401, 71, 191, 311, 431, 101, 221, 341, 461, 26, 146, 266, 386,
56, 176, 296, 416, 86, 206, 326, 446, 116, 236, 356, 476, 2, 122, 242,
362, 32, 152, 272, 392, 62, 182, 302, 422, 92, 212, 332, 452, 17, 137,
257, 377, 47, 167, 287, 407, 77, 197, 317, 437, 107, 227, 347, 467, 7,
127, 247, 367, 37, 157, 277, 397, 67, 187, 307, 427, 97, 217, 337, 457,
22, 142, 262, 382, 52, 172, 292, 412, 82, 202, 322, 442, 112, 232, 352,
472, 12, 132, 252, 372, 42, 162, 282, 402, 72, 192, 312, 432, 102, 222,
342, 462, 27, 147, 267, 387, 57, 177, 297, 417, 87, 207, 327, 447, 117,
237, 357, 477, 3, 123, 243, 363, 33, 153, 273, 393, 63, 183, 303, 423,
93, 213, 333, 453, 18, 138, 258, 378, 48, 168, 288, 408, 78, 198, 318,
438, 108, 228, 348, 468, 8, 128, 248, 368, 38, 158, 278, 398, 68, 188,
308, 428, 98, 218, 338, 458, 23, 143, 263, 383, 53, 173, 293, 413, 83,
203, 323, 443, 113, 233, 353, 473, 13, 133, 253, 373, 43, 163, 283, 403,
73, 193, 313, 433, 103, 223, 343, 463, 28, 148, 268, 388, 58, 178, 298,
418, 88, 208, 328, 448, 118, 238, 358, 478, 4, 124, 244, 364, 34, 154,
274, 394, 64, 184, 304, 424, 94, 214, 334, 454, 19, 139, 259, 379, 49,
169, 289, 409, 79, 199, 319, 439, 109, 229, 349, 469, 9, 129, 249, 369,
39, 159, 279, 399, 69, 189, 309, 429, 99, 219, 339, 459, 24, 144, 264,
384, 54, 174, 294, 414, 84, 204, 324, 444, 114, 234, 354, 474, 14, 134,
254, 374, 44, 164, 284, 404, 74, 194, 314, 434, 104, 224, 344, 464, 29,
149, 269, 389, 59, 179, 299, 419, 89, 209, 329, 449, 119, 239, 359, 479,
0, 96, 192, 288, 384, 32, 128, 224, 320, 416, 64, 160, 256, 352, 448,
8, 104, 200, 296, 392, 40, 136, 232, 328, 424, 72, 168, 264, 360, 456,
16, 112, 208, 304, 400, 48, 144, 240, 336, 432, 80, 176, 272, 368, 464,
24, 120, 216, 312, 408, 56, 152, 248, 344, 440, 88, 184, 280, 376, 472,
4, 100, 196, 292, 388, 36, 132, 228, 324, 420, 68, 164, 260, 356, 452,
12, 108, 204, 300, 396, 44, 140, 236, 332, 428, 76, 172, 268, 364, 460,
20, 116, 212, 308, 404, 52, 148, 244, 340, 436, 84, 180, 276, 372, 468,
28, 124, 220, 316, 412, 60, 156, 252, 348, 444, 92, 188, 284, 380, 476,
1, 97, 193, 289, 385, 33, 129, 225, 321, 417, 65, 161, 257, 353, 449,
9, 105, 201, 297, 393, 41, 137, 233, 329, 425, 73, 169, 265, 361, 457,
17, 113, 209, 305, 401, 49, 145, 241, 337, 433, 81, 177, 273, 369, 465,
25, 121, 217, 313, 409, 57, 153, 249, 345, 441, 89, 185, 281, 377, 473,
5, 101, 197, 293, 389, 37, 133, 229, 325, 421, 69, 165, 261, 357, 453,
13, 109, 205, 301, 397, 45, 141, 237, 333, 429, 77, 173, 269, 365, 461,
21, 117, 213, 309, 405, 53, 149, 245, 341, 437, 85, 181, 277, 373, 469,
29, 125, 221, 317, 413, 61, 157, 253, 349, 445, 93, 189, 285, 381, 477,
2, 98, 194, 290, 386, 34, 130, 226, 322, 418, 66, 162, 258, 354, 450,
10, 106, 202, 298, 394, 42, 138, 234, 330, 426, 74, 170, 266, 362, 458,
18, 114, 210, 306, 402, 50, 146, 242, 338, 434, 82, 178, 274, 370, 466,
26, 122, 218, 314, 410, 58, 154, 250, 346, 442, 90, 186, 282, 378, 474,
6, 102, 198, 294, 390, 38, 134, 230, 326, 422, 70, 166, 262, 358, 454,
14, 110, 206, 302, 398, 46, 142, 238, 334, 430, 78, 174, 270, 366, 462,
22, 118, 214, 310, 406, 54, 150, 246, 342, 438, 86, 182, 278, 374, 470,
30, 126, 222, 318, 414, 62, 158, 254, 350, 446, 94, 190, 286, 382, 478,
3, 99, 195, 291, 387, 35, 131, 227, 323, 419, 67, 163, 259, 355, 451,
11, 107, 203, 299, 395, 43, 139, 235, 331, 427, 75, 171, 267, 363, 459,
19, 115, 211, 307, 403, 51, 147, 243, 339, 435, 83, 179, 275, 371, 467,
27, 123, 219, 315, 411, 59, 155, 251, 347, 443, 91, 187, 283, 379, 475,
7, 103, 199, 295, 391, 39, 135, 231, 327, 423, 71, 167, 263, 359, 455,
15, 111, 207, 303, 399, 47, 143, 239, 335, 431, 79, 175, 271, 367, 463,
23, 119, 215, 311, 407, 55, 151, 247, 343, 439, 87, 183, 279, 375, 471,
31, 127, 223, 319, 415, 63, 159, 255, 351, 447, 95, 191, 287, 383, 479,
};
#endif
#ifndef FFT_BITREV240
#define FFT_BITREV240
static const opus_int16 fft_bitrev240[240] = {
0, 60, 120, 180, 15, 75, 135, 195, 30, 90, 150, 210, 45, 105, 165,
225, 5, 65, 125, 185, 20, 80, 140, 200, 35, 95, 155, 215, 50, 110,
170, 230, 10, 70, 130, 190, 25, 85, 145, 205, 40, 100, 160, 220, 55,
115, 175, 235, 1, 61, 121, 181, 16, 76, 136, 196, 31, 91, 151, 211,
46, 106, 166, 226, 6, 66, 126, 186, 21, 81, 141, 201, 36, 96, 156,
216, 51, 111, 171, 231, 11, 71, 131, 191, 26, 86, 146, 206, 41, 101,
161, 221, 56, 116, 176, 236, 2, 62, 122, 182, 17, 77, 137, 197, 32,
92, 152, 212, 47, 107, 167, 227, 7, 67, 127, 187, 22, 82, 142, 202,
37, 97, 157, 217, 52, 112, 172, 232, 12, 72, 132, 192, 27, 87, 147,
207, 42, 102, 162, 222, 57, 117, 177, 237, 3, 63, 123, 183, 18, 78,
138, 198, 33, 93, 153, 213, 48, 108, 168, 228, 8, 68, 128, 188, 23,
83, 143, 203, 38, 98, 158, 218, 53, 113, 173, 233, 13, 73, 133, 193,
28, 88, 148, 208, 43, 103, 163, 223, 58, 118, 178, 238, 4, 64, 124,
184, 19, 79, 139, 199, 34, 94, 154, 214, 49, 109, 169, 229, 9, 69,
129, 189, 24, 84, 144, 204, 39, 99, 159, 219, 54, 114, 174, 234, 14,
74, 134, 194, 29, 89, 149, 209, 44, 104, 164, 224, 59, 119, 179, 239,
0, 48, 96, 144, 192, 16, 64, 112, 160, 208, 32, 80, 128, 176, 224,
4, 52, 100, 148, 196, 20, 68, 116, 164, 212, 36, 84, 132, 180, 228,
8, 56, 104, 152, 200, 24, 72, 120, 168, 216, 40, 88, 136, 184, 232,
12, 60, 108, 156, 204, 28, 76, 124, 172, 220, 44, 92, 140, 188, 236,
1, 49, 97, 145, 193, 17, 65, 113, 161, 209, 33, 81, 129, 177, 225,
5, 53, 101, 149, 197, 21, 69, 117, 165, 213, 37, 85, 133, 181, 229,
9, 57, 105, 153, 201, 25, 73, 121, 169, 217, 41, 89, 137, 185, 233,
13, 61, 109, 157, 205, 29, 77, 125, 173, 221, 45, 93, 141, 189, 237,
2, 50, 98, 146, 194, 18, 66, 114, 162, 210, 34, 82, 130, 178, 226,
6, 54, 102, 150, 198, 22, 70, 118, 166, 214, 38, 86, 134, 182, 230,
10, 58, 106, 154, 202, 26, 74, 122, 170, 218, 42, 90, 138, 186, 234,
14, 62, 110, 158, 206, 30, 78, 126, 174, 222, 46, 94, 142, 190, 238,
3, 51, 99, 147, 195, 19, 67, 115, 163, 211, 35, 83, 131, 179, 227,
7, 55, 103, 151, 199, 23, 71, 119, 167, 215, 39, 87, 135, 183, 231,
11, 59, 107, 155, 203, 27, 75, 123, 171, 219, 43, 91, 139, 187, 235,
15, 63, 111, 159, 207, 31, 79, 127, 175, 223, 47, 95, 143, 191, 239,
};
#endif
#ifndef FFT_BITREV120
#define FFT_BITREV120
static const opus_int16 fft_bitrev120[120] = {
0, 30, 60, 90, 15, 45, 75, 105, 5, 35, 65, 95, 20, 50, 80,
110, 10, 40, 70, 100, 25, 55, 85, 115, 1, 31, 61, 91, 16, 46,
76, 106, 6, 36, 66, 96, 21, 51, 81, 111, 11, 41, 71, 101, 26,
56, 86, 116, 2, 32, 62, 92, 17, 47, 77, 107, 7, 37, 67, 97,
22, 52, 82, 112, 12, 42, 72, 102, 27, 57, 87, 117, 3, 33, 63,
93, 18, 48, 78, 108, 8, 38, 68, 98, 23, 53, 83, 113, 13, 43,
73, 103, 28, 58, 88, 118, 4, 34, 64, 94, 19, 49, 79, 109, 9,
39, 69, 99, 24, 54, 84, 114, 14, 44, 74, 104, 29, 59, 89, 119,
0, 24, 48, 72, 96, 8, 32, 56, 80, 104, 16, 40, 64, 88, 112,
4, 28, 52, 76, 100, 12, 36, 60, 84, 108, 20, 44, 68, 92, 116,
1, 25, 49, 73, 97, 9, 33, 57, 81, 105, 17, 41, 65, 89, 113,
5, 29, 53, 77, 101, 13, 37, 61, 85, 109, 21, 45, 69, 93, 117,
2, 26, 50, 74, 98, 10, 34, 58, 82, 106, 18, 42, 66, 90, 114,
6, 30, 54, 78, 102, 14, 38, 62, 86, 110, 22, 46, 70, 94, 118,
3, 27, 51, 75, 99, 11, 35, 59, 83, 107, 19, 43, 67, 91, 115,
7, 31, 55, 79, 103, 15, 39, 63, 87, 111, 23, 47, 71, 95, 119,
};
#endif
#ifndef FFT_BITREV60
#define FFT_BITREV60
static const opus_int16 fft_bitrev60[60] = {
0, 15, 30, 45, 5, 20, 35, 50, 10, 25, 40, 55, 1, 16, 31,
46, 6, 21, 36, 51, 11, 26, 41, 56, 2, 17, 32, 47, 7, 22,
37, 52, 12, 27, 42, 57, 3, 18, 33, 48, 8, 23, 38, 53, 13,
28, 43, 58, 4, 19, 34, 49, 9, 24, 39, 54, 14, 29, 44, 59,
0, 12, 24, 36, 48, 4, 16, 28, 40, 52, 8, 20, 32, 44, 56,
1, 13, 25, 37, 49, 5, 17, 29, 41, 53, 9, 21, 33, 45, 57,
2, 14, 26, 38, 50, 6, 18, 30, 42, 54, 10, 22, 34, 46, 58,
3, 15, 27, 39, 51, 7, 19, 31, 43, 55, 11, 23, 35, 47, 59,
};
#endif
@ -426,10 +431,17 @@ static const opus_int16 fft_bitrev60[60] = {
#define FFT_STATE48000_960_0
static const kiss_fft_state fft_state48000_960_0 = {
480, /* nfft */
17476, /* scale */
8, /* scale_shift */
-1, /* shift */
{4, 120, 4, 30, 2, 15, 3, 5, 5, 1, 0, 0, 0, 0, 0, 0, }, /* factors */
{5, 96, 3, 32, 4, 8, 2, 4, 4, 1, 0, 0, 0, 0, 0, 0, }, /* factors */
fft_bitrev480, /* bitrev */
fft_twiddles48000_960, /* bitrev */
#ifdef OVERRIDE_FFT
(arch_fft_state *)&cfg_arch_480,
#else
NULL,
#endif
};
#endif
@ -437,10 +449,17 @@ fft_twiddles48000_960, /* bitrev */
#define FFT_STATE48000_960_1
static const kiss_fft_state fft_state48000_960_1 = {
240, /* nfft */
17476, /* scale */
7, /* scale_shift */
1, /* shift */
{4, 60, 4, 15, 3, 5, 5, 1, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
{5, 48, 3, 16, 4, 4, 4, 1, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
fft_bitrev240, /* bitrev */
fft_twiddles48000_960, /* bitrev */
#ifdef OVERRIDE_FFT
(arch_fft_state *)&cfg_arch_240,
#else
NULL,
#endif
};
#endif
@ -448,10 +467,17 @@ fft_twiddles48000_960, /* bitrev */
#define FFT_STATE48000_960_2
static const kiss_fft_state fft_state48000_960_2 = {
120, /* nfft */
17476, /* scale */
6, /* scale_shift */
2, /* shift */
{4, 30, 2, 15, 3, 5, 5, 1, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
{5, 24, 3, 8, 2, 4, 4, 1, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
fft_bitrev120, /* bitrev */
fft_twiddles48000_960, /* bitrev */
#ifdef OVERRIDE_FFT
(arch_fft_state *)&cfg_arch_120,
#else
NULL,
#endif
};
#endif
@ -459,10 +485,17 @@ fft_twiddles48000_960, /* bitrev */
#define FFT_STATE48000_960_3
static const kiss_fft_state fft_state48000_960_3 = {
60, /* nfft */
17476, /* scale */
5, /* scale_shift */
3, /* shift */
{4, 15, 3, 5, 5, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
{5, 12, 3, 4, 4, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
fft_bitrev60, /* bitrev */
fft_twiddles48000_960, /* bitrev */
#ifdef OVERRIDE_FFT
(arch_fft_state *)&cfg_arch_60,
#else
NULL,
#endif
};
#endif
@ -470,104 +503,368 @@ fft_twiddles48000_960, /* bitrev */
#ifndef MDCT_TWIDDLES960
#define MDCT_TWIDDLES960
static const opus_val16 mdct_twiddles960[481] = {
32767, 32767, 32767, 32767, 32766,
32763, 32762, 32759, 32757, 32753,
32751, 32747, 32743, 32738, 32733,
32729, 32724, 32717, 32711, 32705,
32698, 32690, 32683, 32676, 32667,
32658, 32650, 32640, 32631, 32620,
32610, 32599, 32588, 32577, 32566,
32554, 32541, 32528, 32515, 32502,
32487, 32474, 32459, 32444, 32429,
32413, 32397, 32381, 32364, 32348,
32331, 32313, 32294, 32277, 32257,
32239, 32219, 32200, 32180, 32159,
32138, 32118, 32096, 32074, 32051,
32029, 32006, 31984, 31960, 31936,
31912, 31888, 31863, 31837, 31812,
31786, 31760, 31734, 31707, 31679,
31652, 31624, 31596, 31567, 31539,
31508, 31479, 31450, 31419, 31388,
31357, 31326, 31294, 31262, 31230,
31198, 31164, 31131, 31097, 31063,
31030, 30994, 30959, 30924, 30889,
30853, 30816, 30779, 30743, 30705,
30668, 30629, 30592, 30553, 30515,
30475, 30435, 30396, 30356, 30315,
30274, 30233, 30191, 30149, 30107,
30065, 30022, 29979, 29936, 29891,
29847, 29803, 29758, 29713, 29668,
29622, 29577, 29529, 29483, 29436,
29390, 29341, 29293, 29246, 29197,
29148, 29098, 29050, 29000, 28949,
28899, 28848, 28797, 28746, 28694,
28642, 28590, 28537, 28485, 28432,
28378, 28324, 28271, 28217, 28162,
28106, 28051, 27995, 27940, 27884,
27827, 27770, 27713, 27657, 27598,
27540, 27481, 27423, 27365, 27305,
27246, 27187, 27126, 27066, 27006,
26945, 26883, 26822, 26760, 26698,
26636, 26574, 26510, 26448, 26383,
26320, 26257, 26191, 26127, 26062,
25997, 25931, 25866, 25800, 25734,
25667, 25601, 25533, 25466, 25398,
25330, 25262, 25194, 25125, 25056,
24987, 24917, 24848, 24778, 24707,
24636, 24566, 24495, 24424, 24352,
24280, 24208, 24135, 24063, 23990,
23917, 23842, 23769, 23695, 23622,
23546, 23472, 23398, 23322, 23246,
23171, 23095, 23018, 22942, 22866,
22788, 22711, 22634, 22557, 22478,
22400, 22322, 22244, 22165, 22085,
22006, 21927, 21846, 21766, 21687,
21606, 21524, 21443, 21363, 21282,
21199, 21118, 21035, 20954, 20870,
20788, 20705, 20621, 20538, 20455,
20371, 20286, 20202, 20118, 20034,
19947, 19863, 19777, 19692, 19606,
19520, 19434, 19347, 19260, 19174,
19088, 18999, 18911, 18825, 18737,
18648, 18560, 18472, 18384, 18294,
18205, 18116, 18025, 17936, 17846,
17757, 17666, 17576, 17485, 17395,
17303, 17212, 17122, 17030, 16937,
16846, 16755, 16662, 16569, 16477,
16385, 16291, 16198, 16105, 16012,
15917, 15824, 15730, 15636, 15541,
15447, 15352, 15257, 15162, 15067,
14973, 14875, 14781, 14685, 14589,
14493, 14396, 14300, 14204, 14107,
14010, 13914, 13815, 13718, 13621,
13524, 13425, 13328, 13230, 13133,
13033, 12935, 12836, 12738, 12638,
12540, 12441, 12341, 12241, 12142,
12044, 11943, 11843, 11744, 11643,
11542, 11442, 11342, 11241, 11139,
11039, 10939, 10836, 10736, 10635,
10534, 10431, 10330, 10228, 10127,
10024, 9921, 9820, 9718, 9614,
9512, 9410, 9306, 9204, 9101,
8998, 8895, 8791, 8689, 8585,
8481, 8377, 8274, 8171, 8067,
7962, 7858, 7753, 7650, 7545,
7441, 7336, 7231, 7129, 7023,
6917, 6813, 6709, 6604, 6498,
6393, 6288, 6182, 6077, 5973,
5867, 5760, 5656, 5549, 5445,
5339, 5232, 5127, 5022, 4914,
4809, 4703, 4596, 4490, 4384,
4278, 4171, 4065, 3958, 3852,
3745, 3640, 3532, 3426, 3318,
3212, 3106, 2998, 2891, 2786,
2679, 2570, 2465, 2358, 2251,
2143, 2037, 1929, 1823, 1715,
1609, 1501, 1393, 1287, 1180,
1073, 964, 858, 751, 644,
535, 429, 322, 214, 107,
0, };
static const opus_val16 mdct_twiddles960[1800] = {
32767, 32767, 32767, 32766, 32765,
32763, 32761, 32759, 32756, 32753,
32750, 32746, 32742, 32738, 32733,
32728, 32722, 32717, 32710, 32704,
32697, 32690, 32682, 32674, 32666,
32657, 32648, 32639, 32629, 32619,
32609, 32598, 32587, 32576, 32564,
32552, 32539, 32526, 32513, 32500,
32486, 32472, 32457, 32442, 32427,
32411, 32395, 32379, 32362, 32345,
32328, 32310, 32292, 32274, 32255,
32236, 32217, 32197, 32177, 32157,
32136, 32115, 32093, 32071, 32049,
32027, 32004, 31981, 31957, 31933,
31909, 31884, 31859, 31834, 31809,
31783, 31756, 31730, 31703, 31676,
31648, 31620, 31592, 31563, 31534,
31505, 31475, 31445, 31415, 31384,
31353, 31322, 31290, 31258, 31226,
31193, 31160, 31127, 31093, 31059,
31025, 30990, 30955, 30920, 30884,
30848, 30812, 30775, 30738, 30701,
30663, 30625, 30587, 30548, 30509,
30470, 30430, 30390, 30350, 30309,
30269, 30227, 30186, 30144, 30102,
30059, 30016, 29973, 29930, 29886,
29842, 29797, 29752, 29707, 29662,
29616, 29570, 29524, 29477, 29430,
29383, 29335, 29287, 29239, 29190,
29142, 29092, 29043, 28993, 28943,
28892, 28842, 28791, 28739, 28688,
28636, 28583, 28531, 28478, 28425,
28371, 28317, 28263, 28209, 28154,
28099, 28044, 27988, 27932, 27876,
27820, 27763, 27706, 27648, 27591,
27533, 27474, 27416, 27357, 27298,
27238, 27178, 27118, 27058, 26997,
26936, 26875, 26814, 26752, 26690,
26628, 26565, 26502, 26439, 26375,
26312, 26247, 26183, 26119, 26054,
25988, 25923, 25857, 25791, 25725,
25658, 25592, 25524, 25457, 25389,
25322, 25253, 25185, 25116, 25047,
24978, 24908, 24838, 24768, 24698,
24627, 24557, 24485, 24414, 24342,
24270, 24198, 24126, 24053, 23980,
23907, 23834, 23760, 23686, 23612,
23537, 23462, 23387, 23312, 23237,
23161, 23085, 23009, 22932, 22856,
22779, 22701, 22624, 22546, 22468,
22390, 22312, 22233, 22154, 22075,
21996, 21916, 21836, 21756, 21676,
21595, 21515, 21434, 21352, 21271,
21189, 21107, 21025, 20943, 20860,
20777, 20694, 20611, 20528, 20444,
20360, 20276, 20192, 20107, 20022,
19937, 19852, 19767, 19681, 19595,
19509, 19423, 19336, 19250, 19163,
19076, 18988, 18901, 18813, 18725,
18637, 18549, 18460, 18372, 18283,
18194, 18104, 18015, 17925, 17835,
17745, 17655, 17565, 17474, 17383,
17292, 17201, 17110, 17018, 16927,
16835, 16743, 16650, 16558, 16465,
16372, 16279, 16186, 16093, 15999,
15906, 15812, 15718, 15624, 15529,
15435, 15340, 15245, 15150, 15055,
14960, 14864, 14769, 14673, 14577,
14481, 14385, 14288, 14192, 14095,
13998, 13901, 13804, 13706, 13609,
13511, 13414, 13316, 13218, 13119,
13021, 12923, 12824, 12725, 12626,
12527, 12428, 12329, 12230, 12130,
12030, 11930, 11831, 11730, 11630,
11530, 11430, 11329, 11228, 11128,
11027, 10926, 10824, 10723, 10622,
10520, 10419, 10317, 10215, 10113,
10011, 9909, 9807, 9704, 9602,
9499, 9397, 9294, 9191, 9088,
8985, 8882, 8778, 8675, 8572,
8468, 8364, 8261, 8157, 8053,
7949, 7845, 7741, 7637, 7532,
7428, 7323, 7219, 7114, 7009,
6905, 6800, 6695, 6590, 6485,
6380, 6274, 6169, 6064, 5958,
5853, 5747, 5642, 5536, 5430,
5325, 5219, 5113, 5007, 4901,
4795, 4689, 4583, 4476, 4370,
4264, 4157, 4051, 3945, 3838,
3732, 3625, 3518, 3412, 3305,
3198, 3092, 2985, 2878, 2771,
2664, 2558, 2451, 2344, 2237,
2130, 2023, 1916, 1809, 1702,
1594, 1487, 1380, 1273, 1166,
1059, 952, 844, 737, 630,
523, 416, 308, 201, 94,
-13, -121, -228, -335, -442,
-550, -657, -764, -871, -978,
-1086, -1193, -1300, -1407, -1514,
-1621, -1728, -1835, -1942, -2049,
-2157, -2263, -2370, -2477, -2584,
-2691, -2798, -2905, -3012, -3118,
-3225, -3332, -3439, -3545, -3652,
-3758, -3865, -3971, -4078, -4184,
-4290, -4397, -4503, -4609, -4715,
-4821, -4927, -5033, -5139, -5245,
-5351, -5457, -5562, -5668, -5774,
-5879, -5985, -6090, -6195, -6301,
-6406, -6511, -6616, -6721, -6826,
-6931, -7036, -7140, -7245, -7349,
-7454, -7558, -7663, -7767, -7871,
-7975, -8079, -8183, -8287, -8390,
-8494, -8597, -8701, -8804, -8907,
-9011, -9114, -9217, -9319, -9422,
-9525, -9627, -9730, -9832, -9934,
-10037, -10139, -10241, -10342, -10444,
-10546, -10647, -10748, -10850, -10951,
-11052, -11153, -11253, -11354, -11455,
-11555, -11655, -11756, -11856, -11955,
-12055, -12155, -12254, -12354, -12453,
-12552, -12651, -12750, -12849, -12947,
-13046, -13144, -13242, -13340, -13438,
-13536, -13633, -13731, -13828, -13925,
-14022, -14119, -14216, -14312, -14409,
-14505, -14601, -14697, -14793, -14888,
-14984, -15079, -15174, -15269, -15364,
-15459, -15553, -15647, -15741, -15835,
-15929, -16023, -16116, -16210, -16303,
-16396, -16488, -16581, -16673, -16766,
-16858, -16949, -17041, -17133, -17224,
-17315, -17406, -17497, -17587, -17678,
-17768, -17858, -17948, -18037, -18127,
-18216, -18305, -18394, -18483, -18571,
-18659, -18747, -18835, -18923, -19010,
-19098, -19185, -19271, -19358, -19444,
-19531, -19617, -19702, -19788, -19873,
-19959, -20043, -20128, -20213, -20297,
-20381, -20465, -20549, -20632, -20715,
-20798, -20881, -20963, -21046, -21128,
-21210, -21291, -21373, -21454, -21535,
-21616, -21696, -21776, -21856, -21936,
-22016, -22095, -22174, -22253, -22331,
-22410, -22488, -22566, -22643, -22721,
-22798, -22875, -22951, -23028, -23104,
-23180, -23256, -23331, -23406, -23481,
-23556, -23630, -23704, -23778, -23852,
-23925, -23998, -24071, -24144, -24216,
-24288, -24360, -24432, -24503, -24574,
-24645, -24716, -24786, -24856, -24926,
-24995, -25064, -25133, -25202, -25270,
-25339, -25406, -25474, -25541, -25608,
-25675, -25742, -25808, -25874, -25939,
-26005, -26070, -26135, -26199, -26264,
-26327, -26391, -26455, -26518, -26581,
-26643, -26705, -26767, -26829, -26891,
-26952, -27013, -27073, -27133, -27193,
-27253, -27312, -27372, -27430, -27489,
-27547, -27605, -27663, -27720, -27777,
-27834, -27890, -27946, -28002, -28058,
-28113, -28168, -28223, -28277, -28331,
-28385, -28438, -28491, -28544, -28596,
-28649, -28701, -28752, -28803, -28854,
-28905, -28955, -29006, -29055, -29105,
-29154, -29203, -29251, -29299, -29347,
-29395, -29442, -29489, -29535, -29582,
-29628, -29673, -29719, -29764, -29808,
-29853, -29897, -29941, -29984, -30027,
-30070, -30112, -30154, -30196, -30238,
-30279, -30320, -30360, -30400, -30440,
-30480, -30519, -30558, -30596, -30635,
-30672, -30710, -30747, -30784, -30821,
-30857, -30893, -30929, -30964, -30999,
-31033, -31068, -31102, -31135, -31168,
-31201, -31234, -31266, -31298, -31330,
-31361, -31392, -31422, -31453, -31483,
-31512, -31541, -31570, -31599, -31627,
-31655, -31682, -31710, -31737, -31763,
-31789, -31815, -31841, -31866, -31891,
-31915, -31939, -31963, -31986, -32010,
-32032, -32055, -32077, -32099, -32120,
-32141, -32162, -32182, -32202, -32222,
-32241, -32260, -32279, -32297, -32315,
-32333, -32350, -32367, -32383, -32399,
-32415, -32431, -32446, -32461, -32475,
-32489, -32503, -32517, -32530, -32542,
-32555, -32567, -32579, -32590, -32601,
-32612, -32622, -32632, -32641, -32651,
-32659, -32668, -32676, -32684, -32692,
-32699, -32706, -32712, -32718, -32724,
-32729, -32734, -32739, -32743, -32747,
-32751, -32754, -32757, -32760, -32762,
-32764, -32765, -32767, -32767, -32767,
32767, 32767, 32765, 32761, 32756,
32750, 32742, 32732, 32722, 32710,
32696, 32681, 32665, 32647, 32628,
32608, 32586, 32562, 32538, 32512,
32484, 32455, 32425, 32393, 32360,
32326, 32290, 32253, 32214, 32174,
32133, 32090, 32046, 32001, 31954,
31906, 31856, 31805, 31753, 31700,
31645, 31588, 31530, 31471, 31411,
31349, 31286, 31222, 31156, 31089,
31020, 30951, 30880, 30807, 30733,
30658, 30582, 30504, 30425, 30345,
30263, 30181, 30096, 30011, 29924,
29836, 29747, 29656, 29564, 29471,
29377, 29281, 29184, 29086, 28987,
28886, 28784, 28681, 28577, 28471,
28365, 28257, 28147, 28037, 27925,
27812, 27698, 27583, 27467, 27349,
27231, 27111, 26990, 26868, 26744,
26620, 26494, 26367, 26239, 26110,
25980, 25849, 25717, 25583, 25449,
25313, 25176, 25038, 24900, 24760,
24619, 24477, 24333, 24189, 24044,
23898, 23751, 23602, 23453, 23303,
23152, 22999, 22846, 22692, 22537,
22380, 22223, 22065, 21906, 21746,
21585, 21423, 21261, 21097, 20933,
20767, 20601, 20434, 20265, 20096,
19927, 19756, 19584, 19412, 19239,
19065, 18890, 18714, 18538, 18361,
18183, 18004, 17824, 17644, 17463,
17281, 17098, 16915, 16731, 16546,
16361, 16175, 15988, 15800, 15612,
15423, 15234, 15043, 14852, 14661,
14469, 14276, 14083, 13889, 13694,
13499, 13303, 13107, 12910, 12713,
12515, 12317, 12118, 11918, 11718,
11517, 11316, 11115, 10913, 10710,
10508, 10304, 10100, 9896, 9691,
9486, 9281, 9075, 8869, 8662,
8455, 8248, 8040, 7832, 7623,
7415, 7206, 6996, 6787, 6577,
6366, 6156, 5945, 5734, 5523,
5311, 5100, 4888, 4675, 4463,
4251, 4038, 3825, 3612, 3399,
3185, 2972, 2758, 2544, 2330,
2116, 1902, 1688, 1474, 1260,
1045, 831, 617, 402, 188,
-27, -241, -456, -670, -885,
-1099, -1313, -1528, -1742, -1956,
-2170, -2384, -2598, -2811, -3025,
-3239, -3452, -3665, -3878, -4091,
-4304, -4516, -4728, -4941, -5153,
-5364, -5576, -5787, -5998, -6209,
-6419, -6629, -6839, -7049, -7258,
-7467, -7676, -7884, -8092, -8300,
-8507, -8714, -8920, -9127, -9332,
-9538, -9743, -9947, -10151, -10355,
-10558, -10761, -10963, -11165, -11367,
-11568, -11768, -11968, -12167, -12366,
-12565, -12762, -12960, -13156, -13352,
-13548, -13743, -13937, -14131, -14324,
-14517, -14709, -14900, -15091, -15281,
-15470, -15659, -15847, -16035, -16221,
-16407, -16593, -16777, -16961, -17144,
-17326, -17508, -17689, -17869, -18049,
-18227, -18405, -18582, -18758, -18934,
-19108, -19282, -19455, -19627, -19799,
-19969, -20139, -20308, -20475, -20642,
-20809, -20974, -21138, -21301, -21464,
-21626, -21786, -21946, -22105, -22263,
-22420, -22575, -22730, -22884, -23037,
-23189, -23340, -23490, -23640, -23788,
-23935, -24080, -24225, -24369, -24512,
-24654, -24795, -24934, -25073, -25211,
-25347, -25482, -25617, -25750, -25882,
-26013, -26143, -26272, -26399, -26526,
-26651, -26775, -26898, -27020, -27141,
-27260, -27379, -27496, -27612, -27727,
-27841, -27953, -28065, -28175, -28284,
-28391, -28498, -28603, -28707, -28810,
-28911, -29012, -29111, -29209, -29305,
-29401, -29495, -29587, -29679, -29769,
-29858, -29946, -30032, -30118, -30201,
-30284, -30365, -30445, -30524, -30601,
-30677, -30752, -30825, -30897, -30968,
-31038, -31106, -31172, -31238, -31302,
-31365, -31426, -31486, -31545, -31602,
-31658, -31713, -31766, -31818, -31869,
-31918, -31966, -32012, -32058, -32101,
-32144, -32185, -32224, -32262, -32299,
-32335, -32369, -32401, -32433, -32463,
-32491, -32518, -32544, -32568, -32591,
-32613, -32633, -32652, -32669, -32685,
-32700, -32713, -32724, -32735, -32744,
-32751, -32757, -32762, -32766, -32767,
32767, 32764, 32755, 32741, 32720,
32694, 32663, 32626, 32583, 32535,
32481, 32421, 32356, 32286, 32209,
32128, 32041, 31948, 31850, 31747,
31638, 31523, 31403, 31278, 31148,
31012, 30871, 30724, 30572, 30415,
30253, 30086, 29913, 29736, 29553,
29365, 29172, 28974, 28771, 28564,
28351, 28134, 27911, 27684, 27452,
27216, 26975, 26729, 26478, 26223,
25964, 25700, 25432, 25159, 24882,
24601, 24315, 24026, 23732, 23434,
23133, 22827, 22517, 22204, 21886,
21565, 21240, 20912, 20580, 20244,
19905, 19563, 19217, 18868, 18516,
18160, 17802, 17440, 17075, 16708,
16338, 15964, 15588, 15210, 14829,
14445, 14059, 13670, 13279, 12886,
12490, 12093, 11693, 11291, 10888,
10482, 10075, 9666, 9255, 8843,
8429, 8014, 7597, 7180, 6760,
6340, 5919, 5496, 5073, 4649,
4224, 3798, 3372, 2945, 2517,
2090, 1661, 1233, 804, 375,
-54, -483, -911, -1340, -1768,
-2197, -2624, -3052, -3479, -3905,
-4330, -4755, -5179, -5602, -6024,
-6445, -6865, -7284, -7702, -8118,
-8533, -8946, -9358, -9768, -10177,
-10584, -10989, -11392, -11793, -12192,
-12589, -12984, -13377, -13767, -14155,
-14541, -14924, -15305, -15683, -16058,
-16430, -16800, -17167, -17531, -17892,
-18249, -18604, -18956, -19304, -19649,
-19990, -20329, -20663, -20994, -21322,
-21646, -21966, -22282, -22595, -22904,
-23208, -23509, -23806, -24099, -24387,
-24672, -24952, -25228, -25499, -25766,
-26029, -26288, -26541, -26791, -27035,
-27275, -27511, -27741, -27967, -28188,
-28405, -28616, -28823, -29024, -29221,
-29412, -29599, -29780, -29957, -30128,
-30294, -30455, -30611, -30761, -30906,
-31046, -31181, -31310, -31434, -31552,
-31665, -31773, -31875, -31972, -32063,
-32149, -32229, -32304, -32373, -32437,
-32495, -32547, -32594, -32635, -32671,
-32701, -32726, -32745, -32758, -32766,
32767, 32754, 32717, 32658, 32577,
32473, 32348, 32200, 32029, 31837,
31624, 31388, 31131, 30853, 30553,
30232, 29891, 29530, 29148, 28746,
28324, 27883, 27423, 26944, 26447,
25931, 25398, 24847, 24279, 23695,
23095, 22478, 21846, 21199, 20538,
19863, 19174, 18472, 17757, 17030,
16291, 15541, 14781, 14010, 13230,
12441, 11643, 10837, 10024, 9204,
8377, 7545, 6708, 5866, 5020,
4171, 3319, 2464, 1608, 751,
-107, -965, -1822, -2678, -3532,
-4383, -5232, -6077, -6918, -7754,
-8585, -9409, -10228, -11039, -11843,
-12639, -13426, -14204, -14972, -15730,
-16477, -17213, -17937, -18648, -19347,
-20033, -20705, -21363, -22006, -22634,
-23246, -23843, -24423, -24986, -25533,
-26062, -26573, -27066, -27540, -27995,
-28431, -28848, -29245, -29622, -29979,
-30315, -30630, -30924, -31197, -31449,
-31679, -31887, -32074, -32239, -32381,
-32501, -32600, -32675, -32729, -32759,
};
#endif
static const CELTMode mode48000_960_120 = {

388
Frameworks/Opus/Opus/opus/celt/static_modes_fixed_arm_ne10.h Normal file
View file

@ -0,0 +1,388 @@
/* The contents of this file was automatically generated by
* dump_mode_arm_ne10.c with arguments: 48000 960
* It contains static definitions for some pre-defined modes. */
#include <NE10_init.h>
#ifndef NE10_FFT_PARAMS48000_960
#define NE10_FFT_PARAMS48000_960
static const ne10_int32_t ne10_factors_480[64] = {
4, 40, 4, 30, 2, 15, 5, 3, 3, 1, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, };
static const ne10_int32_t ne10_factors_240[64] = {
3, 20, 4, 15, 5, 3, 3, 1, 1, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, };
static const ne10_int32_t ne10_factors_120[64] = {
3, 10, 2, 15, 5, 3, 3, 1, 1, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, };
static const ne10_int32_t ne10_factors_60[64] = {
2, 5, 5, 3, 3, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, };
static const ne10_fft_cpx_int32_t ne10_twiddles_480[480] = {
{0,0}, {2147483647,0}, {2147483647,0},
{2147483647,0}, {1961823921,-873460313}, {1436946998,-1595891394},
{2147483647,0}, {1436946998,-1595891394}, {-224473265,-2135719496},
{2147483647,0}, {663608871,-2042378339}, {-1737350854,-1262259096},
{2147483647,0}, {-224473265,-2135719496}, {-2100555935,446487152},
{2147483647,0}, {2100555974,-446486968}, {1961823921,-873460313},
{1737350743,-1262259248}, {1436946998,-1595891394}, {1073741769,-1859775424},
{663608871,-2042378339}, {224473078,-2135719516}, {-224473265,-2135719496},
{-663609049,-2042378281}, {-1073741932,-1859775330}, {-1436947137,-1595891268},
{-1737350854,-1262259096}, {-1961823997,-873460141}, {-2100556013,-446486785},
{2147483647,0}, {2144540595,-112390613}, {2135719506,-224473172},
{2121044558,-335940465}, {2100555974,-446486968}, {2074309912,-555809682},
{2042378310,-663608960}, {2004848691,-769589332}, {1961823921,-873460313},
{1913421927,-974937199}, {1859775377,-1073741851}, {1801031311,-1169603450},
{1737350743,-1262259248}, {1668908218,-1351455280}, {1595891331,-1436947067},
{1518500216,-1518500282}, {1436946998,-1595891394}, {1351455207,-1668908277},
{1262259172,-1737350799}, {1169603371,-1801031362}, {1073741769,-1859775424},
{974937230,-1913421912}, {873460227,-1961823959}, {769589125,-2004848771},
{663608871,-2042378339}, {555809715,-2074309903}, {446486876,-2100555994},
{335940246,-2121044593}, {224473078,-2135719516}, {112390647,-2144540593},
{2147483647,0}, {2135719506,-224473172}, {2100555974,-446486968},
{2042378310,-663608960}, {1961823921,-873460313}, {1859775377,-1073741851},
{1737350743,-1262259248}, {1595891331,-1436947067}, {1436946998,-1595891394},
{1262259172,-1737350799}, {1073741769,-1859775424}, {873460227,-1961823959},
{663608871,-2042378339}, {446486876,-2100555994}, {224473078,-2135719516},
{-94,-2147483647}, {-224473265,-2135719496}, {-446487060,-2100555955},
{-663609049,-2042378281}, {-873460398,-1961823883}, {-1073741932,-1859775330},
{-1262259116,-1737350839}, {-1436947137,-1595891268}, {-1595891628,-1436946738},
{-1737350854,-1262259096}, {-1859775343,-1073741910}, {-1961823997,-873460141},
{-2042378447,-663608538}, {-2100556013,-446486785}, {-2135719499,-224473240},
{2147483647,0}, {2121044558,-335940465}, {2042378310,-663608960},
{1913421927,-974937199}, {1737350743,-1262259248}, {1518500216,-1518500282},
{1262259172,-1737350799}, {974937230,-1913421912}, {663608871,-2042378339},
{335940246,-2121044593}, {-94,-2147483647}, {-335940431,-2121044564},
{-663609049,-2042378281}, {-974937397,-1913421827}, {-1262259116,-1737350839},
{-1518500258,-1518500240}, {-1737350854,-1262259096}, {-1913422071,-974936918},
{-2042378447,-663608538}, {-2121044568,-335940406}, {-2147483647,188},
{-2121044509,335940777}, {-2042378331,663608895}, {-1913421900,974937252},
{-1737350633,1262259400}, {-1518499993,1518500506}, {-1262258813,1737351059},
{-974936606,1913422229}, {-663609179,2042378239}, {-335940566,2121044542},
{2147483647,0}, {2147299667,-28109693}, {2146747758,-56214570},
{2145828015,-84309815}, {2144540595,-112390613}, {2142885719,-140452154},
{2140863671,-168489630}, {2138474797,-196498235}, {2135719506,-224473172},
{2132598271,-252409646}, {2129111626,-280302871}, {2125260168,-308148068},
{2121044558,-335940465}, {2116465518,-363675300}, {2111523833,-391347822},
{2106220349,-418953288}, {2100555974,-446486968}, {2094531681,-473944146},
{2088148500,-501320115}, {2081407525,-528610186}, {2074309912,-555809682},
{2066856885,-582913912}, {2059049696,-609918325}, {2050889698,-636818231},
{2042378310,-663608960}, {2033516972,-690285983}, {2024307180,-716844791},
{2014750533,-743280770}, {2004848691,-769589332}, {1994603329,-795766029},
{1984016179,-821806435}, {1973089077,-847706028}, {1961823921,-873460313},
{1950222618,-899064934}, {1938287127,-924515564}, {1926019520,-949807783},
{1913421927,-974937199}, {1900496481,-999899565}, {1887245364,-1024690661},
{1873670877,-1049306180}, {1859775377,-1073741851}, {1845561215,-1097993541},
{1831030826,-1122057097}, {1816186632,-1145928502}, {1801031311,-1169603450},
{1785567394,-1193077993}, {1769797456,-1216348214}, {1753724345,-1239409914},
{1737350743,-1262259248}, {1720679456,-1284892300}, {1703713340,-1307305194},
{1686455222,-1329494189}, {1668908218,-1351455280}, {1651075255,-1373184807},
{1632959307,-1394679144}, {1614563642,-1415934412}, {1595891331,-1436947067},
{1576945572,-1457713510}, {1557729613,-1478230181}, {1538246655,-1498493658},
{1518500216,-1518500282}, {1498493590,-1538246721}, {1478230113,-1557729677},
{1457713441,-1576945636}, {1436946998,-1595891394}, {1415934341,-1614563704},
{1394679073,-1632959368}, {1373184735,-1651075315}, {1351455207,-1668908277},
{1329494115,-1686455280}, {1307305120,-1703713397}, {1284892225,-1720679512},
{1262259172,-1737350799}, {1239409837,-1753724400}, {1216348136,-1769797510},
{1193077915,-1785567446}, {1169603371,-1801031362}, {1145928423,-1816186682},
{1122057017,-1831030875}, {1097993571,-1845561197}, {1073741769,-1859775424},
{1049305987,-1873670985}, {1024690635,-1887245378}, {999899482,-1900496524},
{974937230,-1913421912}, {949807699,-1926019561}, {924515422,-1938287195},
{899064965,-1950222603}, {873460227,-1961823959}, {847705824,-1973089164},
{821806407,-1984016190}, {795765941,-1994603364}, {769589125,-2004848771},
{743280682,-2014750566}, {716844642,-2024307233}, {690286016,-2033516961},
{663608871,-2042378339}, {636818019,-2050889764}, {609918296,-2059049705},
{582913822,-2066856911}, {555809715,-2074309903}, {528610126,-2081407540},
{501319962,-2088148536}, {473944148,-2094531680}, {446486876,-2100555994},
{418953102,-2106220386}, {391347792,-2111523838}, {363675176,-2116465540},
{335940246,-2121044593}, {308148006,-2125260177}, {280302715,-2129111646},
{252409648,-2132598271}, {224473078,-2135719516}, {196498046,-2138474814},
{168489600,-2140863674}, {140452029,-2142885728}, {112390647,-2144540593},
{84309753,-2145828017}, {56214412,-2146747762}, {28109695,-2147299667},
{2147483647,0}, {2146747758,-56214570}, {2144540595,-112390613},
{2140863671,-168489630}, {2135719506,-224473172}, {2129111626,-280302871},
{2121044558,-335940465}, {2111523833,-391347822}, {2100555974,-446486968},
{2088148500,-501320115}, {2074309912,-555809682}, {2059049696,-609918325},
{2042378310,-663608960}, {2024307180,-716844791}, {2004848691,-769589332},
{1984016179,-821806435}, {1961823921,-873460313}, {1938287127,-924515564},
{1913421927,-974937199}, {1887245364,-1024690661}, {1859775377,-1073741851},
{1831030826,-1122057097}, {1801031311,-1169603450}, {1769797456,-1216348214},
{1737350743,-1262259248}, {1703713340,-1307305194}, {1668908218,-1351455280},
{1632959307,-1394679144}, {1595891331,-1436947067}, {1557729613,-1478230181},
{1518500216,-1518500282}, {1478230113,-1557729677}, {1436946998,-1595891394},
{1394679073,-1632959368}, {1351455207,-1668908277}, {1307305120,-1703713397},
{1262259172,-1737350799}, {1216348136,-1769797510}, {1169603371,-1801031362},
{1122057017,-1831030875}, {1073741769,-1859775424}, {1024690635,-1887245378},
{974937230,-1913421912}, {924515422,-1938287195}, {873460227,-1961823959},
{821806407,-1984016190}, {769589125,-2004848771}, {716844642,-2024307233},
{663608871,-2042378339}, {609918296,-2059049705}, {555809715,-2074309903},
{501319962,-2088148536}, {446486876,-2100555994}, {391347792,-2111523838},
{335940246,-2121044593}, {280302715,-2129111646}, {224473078,-2135719516},
{168489600,-2140863674}, {112390647,-2144540593}, {56214412,-2146747762},
{-94,-2147483647}, {-56214600,-2146747757}, {-112390835,-2144540584},
{-168489787,-2140863659}, {-224473265,-2135719496}, {-280302901,-2129111622},
{-335940431,-2121044564}, {-391347977,-2111523804}, {-446487060,-2100555955},
{-501320144,-2088148493}, {-555809896,-2074309855}, {-609918476,-2059049651},
{-663609049,-2042378281}, {-716844819,-2024307170}, {-769589300,-2004848703},
{-821806581,-1984016118}, {-873460398,-1961823883}, {-924515591,-1938287114},
{-974937397,-1913421827}, {-1024690575,-1887245411}, {-1073741932,-1859775330},
{-1122057395,-1831030643}, {-1169603421,-1801031330}, {-1216348291,-1769797403},
{-1262259116,-1737350839}, {-1307305268,-1703713283}, {-1351455453,-1668908078},
{-1394679021,-1632959413}, {-1436947137,-1595891268}, {-1478230435,-1557729372},
{-1518500258,-1518500240}, {-1557729742,-1478230045}, {-1595891628,-1436946738},
{-1632959429,-1394679001}, {-1668908417,-1351455035}, {-1703713298,-1307305248},
{-1737350854,-1262259096}, {-1769797708,-1216347848}, {-1801031344,-1169603400},
{-1831030924,-1122056937}, {-1859775343,-1073741910}, {-1887245423,-1024690552},
{-1913422071,-974936918}, {-1938287125,-924515568}, {-1961823997,-873460141},
{-1984016324,-821806084}, {-2004848713,-769589276}, {-2024307264,-716844553},
{-2042378447,-663608538}, {-2059049731,-609918206}, {-2074309994,-555809377},
{-2088148499,-501320119}, {-2100556013,-446486785}, {-2111523902,-391347448},
{-2121044568,-335940406}, {-2129111659,-280302621}, {-2135719499,-224473240},
{-2140863681,-168489506}, {-2144540612,-112390298}, {-2146747758,-56214574},
{2147483647,0}, {2145828015,-84309815}, {2140863671,-168489630},
{2132598271,-252409646}, {2121044558,-335940465}, {2106220349,-418953288},
{2088148500,-501320115}, {2066856885,-582913912}, {2042378310,-663608960},
{2014750533,-743280770}, {1984016179,-821806435}, {1950222618,-899064934},
{1913421927,-974937199}, {1873670877,-1049306180}, {1831030826,-1122057097},
{1785567394,-1193077993}, {1737350743,-1262259248}, {1686455222,-1329494189},
{1632959307,-1394679144}, {1576945572,-1457713510}, {1518500216,-1518500282},
{1457713441,-1576945636}, {1394679073,-1632959368}, {1329494115,-1686455280},
{1262259172,-1737350799}, {1193077915,-1785567446}, {1122057017,-1831030875},
{1049305987,-1873670985}, {974937230,-1913421912}, {899064965,-1950222603},
{821806407,-1984016190}, {743280682,-2014750566}, {663608871,-2042378339},
{582913822,-2066856911}, {501319962,-2088148536}, {418953102,-2106220386},
{335940246,-2121044593}, {252409648,-2132598271}, {168489600,-2140863674},
{84309753,-2145828017}, {-94,-2147483647}, {-84309940,-2145828010},
{-168489787,-2140863659}, {-252409834,-2132598249}, {-335940431,-2121044564},
{-418953286,-2106220349}, {-501320144,-2088148493}, {-582914003,-2066856860},
{-663609049,-2042378281}, {-743280858,-2014750501}, {-821806581,-1984016118},
{-899065136,-1950222525}, {-974937397,-1913421827}, {-1049306374,-1873670768},
{-1122057395,-1831030643}, {-1193078284,-1785567199}, {-1262259116,-1737350839},
{-1329494061,-1686455323}, {-1394679021,-1632959413}, {-1457713485,-1576945595},
{-1518500258,-1518500240}, {-1576945613,-1457713466}, {-1632959429,-1394679001},
{-1686455338,-1329494041}, {-1737350854,-1262259096}, {-1785567498,-1193077837},
{-1831030924,-1122056937}, {-1873671031,-1049305905}, {-1913422071,-974936918},
{-1950222750,-899064648}, {-1984016324,-821806084}, {-2014750687,-743280354},
{-2042378447,-663608538}, {-2066856867,-582913978}, {-2088148499,-501320119},
{-2106220354,-418953261}, {-2121044568,-335940406}, {-2132598282,-252409555},
{-2140863681,-168489506}, {-2145828021,-84309659}, {-2147483647,188},
{-2145828006,84310034}, {-2140863651,168489881}, {-2132598237,252409928},
{-2121044509,335940777}, {-2106220281,418953629}, {-2088148411,501320484},
{-2066856765,582914339}, {-2042378331,663608895}, {-2014750557,743280706},
{-1984016181,821806431}, {-1950222593,899064989}, {-1913421900,974937252},
{-1873670848,1049306232}, {-1831030728,1122057257}, {-1785567289,1193078149},
{-1737350633,1262259400}, {-1686455106,1329494336}, {-1632959185,1394679287},
{-1576945358,1457713742}, {-1518499993,1518500506}, {-1457713209,1576945850},
{-1394678735,1632959656}, {-1329493766,1686455555}, {-1262258813,1737351059},
{-1193077546,1785567692}, {-1122056638,1831031107}, {-1049305599,1873671202},
{-974936606,1913422229}, {-899064330,1950222896}, {-821805761,1984016458},
{-743280025,2014750808}, {-663609179,2042378239}, {-582914134,2066856823},
{-501320277,2088148461}, {-418953420,2106220322}, {-335940566,2121044542},
{-252409716,2132598263}, {-168489668,2140863668}, {-84309821,2145828015},
};
static const ne10_fft_cpx_int32_t ne10_twiddles_240[240] = {
{0,0}, {2147483647,0}, {2147483647,0},
{2147483647,0}, {1961823921,-873460313}, {1436946998,-1595891394},
{2147483647,0}, {1436946998,-1595891394}, {-224473265,-2135719496},
{2147483647,0}, {663608871,-2042378339}, {-1737350854,-1262259096},
{2147483647,0}, {-224473265,-2135719496}, {-2100555935,446487152},
{2147483647,0}, {2135719506,-224473172}, {2100555974,-446486968},
{2042378310,-663608960}, {1961823921,-873460313}, {1859775377,-1073741851},
{1737350743,-1262259248}, {1595891331,-1436947067}, {1436946998,-1595891394},
{1262259172,-1737350799}, {1073741769,-1859775424}, {873460227,-1961823959},
{663608871,-2042378339}, {446486876,-2100555994}, {224473078,-2135719516},
{2147483647,0}, {2100555974,-446486968}, {1961823921,-873460313},
{1737350743,-1262259248}, {1436946998,-1595891394}, {1073741769,-1859775424},
{663608871,-2042378339}, {224473078,-2135719516}, {-224473265,-2135719496},
{-663609049,-2042378281}, {-1073741932,-1859775330}, {-1436947137,-1595891268},
{-1737350854,-1262259096}, {-1961823997,-873460141}, {-2100556013,-446486785},
{2147483647,0}, {2042378310,-663608960}, {1737350743,-1262259248},
{1262259172,-1737350799}, {663608871,-2042378339}, {-94,-2147483647},
{-663609049,-2042378281}, {-1262259116,-1737350839}, {-1737350854,-1262259096},
{-2042378447,-663608538}, {-2147483647,188}, {-2042378331,663608895},
{-1737350633,1262259400}, {-1262258813,1737351059}, {-663609179,2042378239},
{2147483647,0}, {2146747758,-56214570}, {2144540595,-112390613},
{2140863671,-168489630}, {2135719506,-224473172}, {2129111626,-280302871},
{2121044558,-335940465}, {2111523833,-391347822}, {2100555974,-446486968},
{2088148500,-501320115}, {2074309912,-555809682}, {2059049696,-609918325},
{2042378310,-663608960}, {2024307180,-716844791}, {2004848691,-769589332},
{1984016179,-821806435}, {1961823921,-873460313}, {1938287127,-924515564},
{1913421927,-974937199}, {1887245364,-1024690661}, {1859775377,-1073741851},
{1831030826,-1122057097}, {1801031311,-1169603450}, {1769797456,-1216348214},
{1737350743,-1262259248}, {1703713340,-1307305194}, {1668908218,-1351455280},
{1632959307,-1394679144}, {1595891331,-1436947067}, {1557729613,-1478230181},
{1518500216,-1518500282}, {1478230113,-1557729677}, {1436946998,-1595891394},
{1394679073,-1632959368}, {1351455207,-1668908277}, {1307305120,-1703713397},
{1262259172,-1737350799}, {1216348136,-1769797510}, {1169603371,-1801031362},
{1122057017,-1831030875}, {1073741769,-1859775424}, {1024690635,-1887245378},
{974937230,-1913421912}, {924515422,-1938287195}, {873460227,-1961823959},
{821806407,-1984016190}, {769589125,-2004848771}, {716844642,-2024307233},
{663608871,-2042378339}, {609918296,-2059049705}, {555809715,-2074309903},
{501319962,-2088148536}, {446486876,-2100555994}, {391347792,-2111523838},
{335940246,-2121044593}, {280302715,-2129111646}, {224473078,-2135719516},
{168489600,-2140863674}, {112390647,-2144540593}, {56214412,-2146747762},
{2147483647,0}, {2144540595,-112390613}, {2135719506,-224473172},
{2121044558,-335940465}, {2100555974,-446486968}, {2074309912,-555809682},
{2042378310,-663608960}, {2004848691,-769589332}, {1961823921,-873460313},
{1913421927,-974937199}, {1859775377,-1073741851}, {1801031311,-1169603450},
{1737350743,-1262259248}, {1668908218,-1351455280}, {1595891331,-1436947067},
{1518500216,-1518500282}, {1436946998,-1595891394}, {1351455207,-1668908277},
{1262259172,-1737350799}, {1169603371,-1801031362}, {1073741769,-1859775424},
{974937230,-1913421912}, {873460227,-1961823959}, {769589125,-2004848771},
{663608871,-2042378339}, {555809715,-2074309903}, {446486876,-2100555994},
{335940246,-2121044593}, {224473078,-2135719516}, {112390647,-2144540593},
{-94,-2147483647}, {-112390835,-2144540584}, {-224473265,-2135719496},
{-335940431,-2121044564}, {-446487060,-2100555955}, {-555809896,-2074309855},
{-663609049,-2042378281}, {-769589300,-2004848703}, {-873460398,-1961823883},
{-974937397,-1913421827}, {-1073741932,-1859775330}, {-1169603421,-1801031330},
{-1262259116,-1737350839}, {-1351455453,-1668908078}, {-1436947137,-1595891268},
{-1518500258,-1518500240}, {-1595891628,-1436946738}, {-1668908417,-1351455035},
{-1737350854,-1262259096}, {-1801031344,-1169603400}, {-1859775343,-1073741910},
{-1913422071,-974936918}, {-1961823997,-873460141}, {-2004848713,-769589276},
{-2042378447,-663608538}, {-2074309994,-555809377}, {-2100556013,-446486785},
{-2121044568,-335940406}, {-2135719499,-224473240}, {-2144540612,-112390298},
{2147483647,0}, {2140863671,-168489630}, {2121044558,-335940465},
{2088148500,-501320115}, {2042378310,-663608960}, {1984016179,-821806435},
{1913421927,-974937199}, {1831030826,-1122057097}, {1737350743,-1262259248},
{1632959307,-1394679144}, {1518500216,-1518500282}, {1394679073,-1632959368},
{1262259172,-1737350799}, {1122057017,-1831030875}, {974937230,-1913421912},
{821806407,-1984016190}, {663608871,-2042378339}, {501319962,-2088148536},
{335940246,-2121044593}, {168489600,-2140863674}, {-94,-2147483647},
{-168489787,-2140863659}, {-335940431,-2121044564}, {-501320144,-2088148493},
{-663609049,-2042378281}, {-821806581,-1984016118}, {-974937397,-1913421827},
{-1122057395,-1831030643}, {-1262259116,-1737350839}, {-1394679021,-1632959413},
{-1518500258,-1518500240}, {-1632959429,-1394679001}, {-1737350854,-1262259096},
{-1831030924,-1122056937}, {-1913422071,-974936918}, {-1984016324,-821806084},
{-2042378447,-663608538}, {-2088148499,-501320119}, {-2121044568,-335940406},
{-2140863681,-168489506}, {-2147483647,188}, {-2140863651,168489881},
{-2121044509,335940777}, {-2088148411,501320484}, {-2042378331,663608895},
{-1984016181,821806431}, {-1913421900,974937252}, {-1831030728,1122057257},
{-1737350633,1262259400}, {-1632959185,1394679287}, {-1518499993,1518500506},
{-1394678735,1632959656}, {-1262258813,1737351059}, {-1122056638,1831031107},
{-974936606,1913422229}, {-821805761,1984016458}, {-663609179,2042378239},
{-501320277,2088148461}, {-335940566,2121044542}, {-168489668,2140863668},
};
static const ne10_fft_cpx_int32_t ne10_twiddles_120[120] = {
{0,0}, {2147483647,0}, {2147483647,0},
{2147483647,0}, {1961823921,-873460313}, {1436946998,-1595891394},
{2147483647,0}, {1436946998,-1595891394}, {-224473265,-2135719496},
{2147483647,0}, {663608871,-2042378339}, {-1737350854,-1262259096},
{2147483647,0}, {-224473265,-2135719496}, {-2100555935,446487152},
{2147483647,0}, {2100555974,-446486968}, {1961823921,-873460313},
{1737350743,-1262259248}, {1436946998,-1595891394}, {1073741769,-1859775424},
{663608871,-2042378339}, {224473078,-2135719516}, {-224473265,-2135719496},
{-663609049,-2042378281}, {-1073741932,-1859775330}, {-1436947137,-1595891268},
{-1737350854,-1262259096}, {-1961823997,-873460141}, {-2100556013,-446486785},
{2147483647,0}, {2144540595,-112390613}, {2135719506,-224473172},
{2121044558,-335940465}, {2100555974,-446486968}, {2074309912,-555809682},
{2042378310,-663608960}, {2004848691,-769589332}, {1961823921,-873460313},
{1913421927,-974937199}, {1859775377,-1073741851}, {1801031311,-1169603450},
{1737350743,-1262259248}, {1668908218,-1351455280}, {1595891331,-1436947067},
{1518500216,-1518500282}, {1436946998,-1595891394}, {1351455207,-1668908277},
{1262259172,-1737350799}, {1169603371,-1801031362}, {1073741769,-1859775424},
{974937230,-1913421912}, {873460227,-1961823959}, {769589125,-2004848771},
{663608871,-2042378339}, {555809715,-2074309903}, {446486876,-2100555994},
{335940246,-2121044593}, {224473078,-2135719516}, {112390647,-2144540593},
{2147483647,0}, {2135719506,-224473172}, {2100555974,-446486968},
{2042378310,-663608960}, {1961823921,-873460313}, {1859775377,-1073741851},
{1737350743,-1262259248}, {1595891331,-1436947067}, {1436946998,-1595891394},
{1262259172,-1737350799}, {1073741769,-1859775424}, {873460227,-1961823959},
{663608871,-2042378339}, {446486876,-2100555994}, {224473078,-2135719516},
{-94,-2147483647}, {-224473265,-2135719496}, {-446487060,-2100555955},
{-663609049,-2042378281}, {-873460398,-1961823883}, {-1073741932,-1859775330},
{-1262259116,-1737350839}, {-1436947137,-1595891268}, {-1595891628,-1436946738},
{-1737350854,-1262259096}, {-1859775343,-1073741910}, {-1961823997,-873460141},
{-2042378447,-663608538}, {-2100556013,-446486785}, {-2135719499,-224473240},
{2147483647,0}, {2121044558,-335940465}, {2042378310,-663608960},
{1913421927,-974937199}, {1737350743,-1262259248}, {1518500216,-1518500282},
{1262259172,-1737350799}, {974937230,-1913421912}, {663608871,-2042378339},
{335940246,-2121044593}, {-94,-2147483647}, {-335940431,-2121044564},
{-663609049,-2042378281}, {-974937397,-1913421827}, {-1262259116,-1737350839},
{-1518500258,-1518500240}, {-1737350854,-1262259096}, {-1913422071,-974936918},
{-2042378447,-663608538}, {-2121044568,-335940406}, {-2147483647,188},
{-2121044509,335940777}, {-2042378331,663608895}, {-1913421900,974937252},
{-1737350633,1262259400}, {-1518499993,1518500506}, {-1262258813,1737351059},
{-974936606,1913422229}, {-663609179,2042378239}, {-335940566,2121044542},
};
static const ne10_fft_cpx_int32_t ne10_twiddles_60[60] = {
{0,0}, {2147483647,0}, {2147483647,0},
{2147483647,0}, {1961823921,-873460313}, {1436946998,-1595891394},
{2147483647,0}, {1436946998,-1595891394}, {-224473265,-2135719496},
{2147483647,0}, {663608871,-2042378339}, {-1737350854,-1262259096},
{2147483647,0}, {-224473265,-2135719496}, {-2100555935,446487152},
{2147483647,0}, {2135719506,-224473172}, {2100555974,-446486968},
{2042378310,-663608960}, {1961823921,-873460313}, {1859775377,-1073741851},
{1737350743,-1262259248}, {1595891331,-1436947067}, {1436946998,-1595891394},
{1262259172,-1737350799}, {1073741769,-1859775424}, {873460227,-1961823959},
{663608871,-2042378339}, {446486876,-2100555994}, {224473078,-2135719516},
{2147483647,0}, {2100555974,-446486968}, {1961823921,-873460313},
{1737350743,-1262259248}, {1436946998,-1595891394}, {1073741769,-1859775424},
{663608871,-2042378339}, {224473078,-2135719516}, {-224473265,-2135719496},
{-663609049,-2042378281}, {-1073741932,-1859775330}, {-1436947137,-1595891268},
{-1737350854,-1262259096}, {-1961823997,-873460141}, {-2100556013,-446486785},
{2147483647,0}, {2042378310,-663608960}, {1737350743,-1262259248},
{1262259172,-1737350799}, {663608871,-2042378339}, {-94,-2147483647},
{-663609049,-2042378281}, {-1262259116,-1737350839}, {-1737350854,-1262259096},
{-2042378447,-663608538}, {-2147483647,188}, {-2042378331,663608895},
{-1737350633,1262259400}, {-1262258813,1737351059}, {-663609179,2042378239},
};
static const ne10_fft_state_int32_t ne10_fft_state_int32_t_480 = {
120,
(ne10_int32_t *)ne10_factors_480,
(ne10_fft_cpx_int32_t *)ne10_twiddles_480,
NULL,
(ne10_fft_cpx_int32_t *)&ne10_twiddles_480[120],
};
static const arch_fft_state cfg_arch_480 = {
1,
(void *)&ne10_fft_state_int32_t_480,
};
static const ne10_fft_state_int32_t ne10_fft_state_int32_t_240 = {
60,
(ne10_int32_t *)ne10_factors_240,
(ne10_fft_cpx_int32_t *)ne10_twiddles_240,
NULL,
(ne10_fft_cpx_int32_t *)&ne10_twiddles_240[60],
};
static const arch_fft_state cfg_arch_240 = {
1,
(void *)&ne10_fft_state_int32_t_240,
};
static const ne10_fft_state_int32_t ne10_fft_state_int32_t_120 = {
30,
(ne10_int32_t *)ne10_factors_120,
(ne10_fft_cpx_int32_t *)ne10_twiddles_120,
NULL,
(ne10_fft_cpx_int32_t *)&ne10_twiddles_120[30],
};
static const arch_fft_state cfg_arch_120 = {
1,
(void *)&ne10_fft_state_int32_t_120,
};
static const ne10_fft_state_int32_t ne10_fft_state_int32_t_60 = {
15,
(ne10_int32_t *)ne10_factors_60,
(ne10_fft_cpx_int32_t *)ne10_twiddles_60,
NULL,
(ne10_fft_cpx_int32_t *)&ne10_twiddles_60[15],
};
static const arch_fft_state cfg_arch_60 = {
1,
(void *)&ne10_fft_state_int32_t_60,
};
#endif /* end NE10_FFT_PARAMS48000_960 */

613
Frameworks/Opus/Opus/opus/celt/static_modes_float.h
View file

@ -4,6 +4,11 @@
#include "modes.h"
#include "rate.h"
#ifdef HAVE_ARM_NE10
#define OVERRIDE_FFT 1
#include "static_modes_float_arm_ne10.h"
#endif
#ifndef DEF_WINDOW120
#define DEF_WINDOW120
static const opus_val16 window120[120] = {
@ -341,84 +346,84 @@ static const kiss_twiddle_cpx fft_twiddles48000_960[480] = {
#ifndef FFT_BITREV480
#define FFT_BITREV480
static const opus_int16 fft_bitrev480[480] = {
0, 120, 240, 360, 30, 150, 270, 390, 60, 180, 300, 420, 90, 210, 330,
450, 15, 135, 255, 375, 45, 165, 285, 405, 75, 195, 315, 435, 105, 225,
345, 465, 5, 125, 245, 365, 35, 155, 275, 395, 65, 185, 305, 425, 95,
215, 335, 455, 20, 140, 260, 380, 50, 170, 290, 410, 80, 200, 320, 440,
110, 230, 350, 470, 10, 130, 250, 370, 40, 160, 280, 400, 70, 190, 310,
430, 100, 220, 340, 460, 25, 145, 265, 385, 55, 175, 295, 415, 85, 205,
325, 445, 115, 235, 355, 475, 1, 121, 241, 361, 31, 151, 271, 391, 61,
181, 301, 421, 91, 211, 331, 451, 16, 136, 256, 376, 46, 166, 286, 406,
76, 196, 316, 436, 106, 226, 346, 466, 6, 126, 246, 366, 36, 156, 276,
396, 66, 186, 306, 426, 96, 216, 336, 456, 21, 141, 261, 381, 51, 171,
291, 411, 81, 201, 321, 441, 111, 231, 351, 471, 11, 131, 251, 371, 41,
161, 281, 401, 71, 191, 311, 431, 101, 221, 341, 461, 26, 146, 266, 386,
56, 176, 296, 416, 86, 206, 326, 446, 116, 236, 356, 476, 2, 122, 242,
362, 32, 152, 272, 392, 62, 182, 302, 422, 92, 212, 332, 452, 17, 137,
257, 377, 47, 167, 287, 407, 77, 197, 317, 437, 107, 227, 347, 467, 7,
127, 247, 367, 37, 157, 277, 397, 67, 187, 307, 427, 97, 217, 337, 457,
22, 142, 262, 382, 52, 172, 292, 412, 82, 202, 322, 442, 112, 232, 352,
472, 12, 132, 252, 372, 42, 162, 282, 402, 72, 192, 312, 432, 102, 222,
342, 462, 27, 147, 267, 387, 57, 177, 297, 417, 87, 207, 327, 447, 117,
237, 357, 477, 3, 123, 243, 363, 33, 153, 273, 393, 63, 183, 303, 423,
93, 213, 333, 453, 18, 138, 258, 378, 48, 168, 288, 408, 78, 198, 318,
438, 108, 228, 348, 468, 8, 128, 248, 368, 38, 158, 278, 398, 68, 188,
308, 428, 98, 218, 338, 458, 23, 143, 263, 383, 53, 173, 293, 413, 83,
203, 323, 443, 113, 233, 353, 473, 13, 133, 253, 373, 43, 163, 283, 403,
73, 193, 313, 433, 103, 223, 343, 463, 28, 148, 268, 388, 58, 178, 298,
418, 88, 208, 328, 448, 118, 238, 358, 478, 4, 124, 244, 364, 34, 154,
274, 394, 64, 184, 304, 424, 94, 214, 334, 454, 19, 139, 259, 379, 49,
169, 289, 409, 79, 199, 319, 439, 109, 229, 349, 469, 9, 129, 249, 369,
39, 159, 279, 399, 69, 189, 309, 429, 99, 219, 339, 459, 24, 144, 264,
384, 54, 174, 294, 414, 84, 204, 324, 444, 114, 234, 354, 474, 14, 134,
254, 374, 44, 164, 284, 404, 74, 194, 314, 434, 104, 224, 344, 464, 29,
149, 269, 389, 59, 179, 299, 419, 89, 209, 329, 449, 119, 239, 359, 479,
0, 96, 192, 288, 384, 32, 128, 224, 320, 416, 64, 160, 256, 352, 448,
8, 104, 200, 296, 392, 40, 136, 232, 328, 424, 72, 168, 264, 360, 456,
16, 112, 208, 304, 400, 48, 144, 240, 336, 432, 80, 176, 272, 368, 464,
24, 120, 216, 312, 408, 56, 152, 248, 344, 440, 88, 184, 280, 376, 472,
4, 100, 196, 292, 388, 36, 132, 228, 324, 420, 68, 164, 260, 356, 452,
12, 108, 204, 300, 396, 44, 140, 236, 332, 428, 76, 172, 268, 364, 460,
20, 116, 212, 308, 404, 52, 148, 244, 340, 436, 84, 180, 276, 372, 468,
28, 124, 220, 316, 412, 60, 156, 252, 348, 444, 92, 188, 284, 380, 476,
1, 97, 193, 289, 385, 33, 129, 225, 321, 417, 65, 161, 257, 353, 449,
9, 105, 201, 297, 393, 41, 137, 233, 329, 425, 73, 169, 265, 361, 457,
17, 113, 209, 305, 401, 49, 145, 241, 337, 433, 81, 177, 273, 369, 465,
25, 121, 217, 313, 409, 57, 153, 249, 345, 441, 89, 185, 281, 377, 473,
5, 101, 197, 293, 389, 37, 133, 229, 325, 421, 69, 165, 261, 357, 453,
13, 109, 205, 301, 397, 45, 141, 237, 333, 429, 77, 173, 269, 365, 461,
21, 117, 213, 309, 405, 53, 149, 245, 341, 437, 85, 181, 277, 373, 469,
29, 125, 221, 317, 413, 61, 157, 253, 349, 445, 93, 189, 285, 381, 477,
2, 98, 194, 290, 386, 34, 130, 226, 322, 418, 66, 162, 258, 354, 450,
10, 106, 202, 298, 394, 42, 138, 234, 330, 426, 74, 170, 266, 362, 458,
18, 114, 210, 306, 402, 50, 146, 242, 338, 434, 82, 178, 274, 370, 466,
26, 122, 218, 314, 410, 58, 154, 250, 346, 442, 90, 186, 282, 378, 474,
6, 102, 198, 294, 390, 38, 134, 230, 326, 422, 70, 166, 262, 358, 454,
14, 110, 206, 302, 398, 46, 142, 238, 334, 430, 78, 174, 270, 366, 462,
22, 118, 214, 310, 406, 54, 150, 246, 342, 438, 86, 182, 278, 374, 470,
30, 126, 222, 318, 414, 62, 158, 254, 350, 446, 94, 190, 286, 382, 478,
3, 99, 195, 291, 387, 35, 131, 227, 323, 419, 67, 163, 259, 355, 451,
11, 107, 203, 299, 395, 43, 139, 235, 331, 427, 75, 171, 267, 363, 459,
19, 115, 211, 307, 403, 51, 147, 243, 339, 435, 83, 179, 275, 371, 467,
27, 123, 219, 315, 411, 59, 155, 251, 347, 443, 91, 187, 283, 379, 475,
7, 103, 199, 295, 391, 39, 135, 231, 327, 423, 71, 167, 263, 359, 455,
15, 111, 207, 303, 399, 47, 143, 239, 335, 431, 79, 175, 271, 367, 463,
23, 119, 215, 311, 407, 55, 151, 247, 343, 439, 87, 183, 279, 375, 471,
31, 127, 223, 319, 415, 63, 159, 255, 351, 447, 95, 191, 287, 383, 479,
};
#endif
#ifndef FFT_BITREV240
#define FFT_BITREV240
static const opus_int16 fft_bitrev240[240] = {
0, 60, 120, 180, 15, 75, 135, 195, 30, 90, 150, 210, 45, 105, 165,
225, 5, 65, 125, 185, 20, 80, 140, 200, 35, 95, 155, 215, 50, 110,
170, 230, 10, 70, 130, 190, 25, 85, 145, 205, 40, 100, 160, 220, 55,
115, 175, 235, 1, 61, 121, 181, 16, 76, 136, 196, 31, 91, 151, 211,
46, 106, 166, 226, 6, 66, 126, 186, 21, 81, 141, 201, 36, 96, 156,
216, 51, 111, 171, 231, 11, 71, 131, 191, 26, 86, 146, 206, 41, 101,
161, 221, 56, 116, 176, 236, 2, 62, 122, 182, 17, 77, 137, 197, 32,
92, 152, 212, 47, 107, 167, 227, 7, 67, 127, 187, 22, 82, 142, 202,
37, 97, 157, 217, 52, 112, 172, 232, 12, 72, 132, 192, 27, 87, 147,
207, 42, 102, 162, 222, 57, 117, 177, 237, 3, 63, 123, 183, 18, 78,
138, 198, 33, 93, 153, 213, 48, 108, 168, 228, 8, 68, 128, 188, 23,
83, 143, 203, 38, 98, 158, 218, 53, 113, 173, 233, 13, 73, 133, 193,
28, 88, 148, 208, 43, 103, 163, 223, 58, 118, 178, 238, 4, 64, 124,
184, 19, 79, 139, 199, 34, 94, 154, 214, 49, 109, 169, 229, 9, 69,
129, 189, 24, 84, 144, 204, 39, 99, 159, 219, 54, 114, 174, 234, 14,
74, 134, 194, 29, 89, 149, 209, 44, 104, 164, 224, 59, 119, 179, 239,
0, 48, 96, 144, 192, 16, 64, 112, 160, 208, 32, 80, 128, 176, 224,
4, 52, 100, 148, 196, 20, 68, 116, 164, 212, 36, 84, 132, 180, 228,
8, 56, 104, 152, 200, 24, 72, 120, 168, 216, 40, 88, 136, 184, 232,
12, 60, 108, 156, 204, 28, 76, 124, 172, 220, 44, 92, 140, 188, 236,
1, 49, 97, 145, 193, 17, 65, 113, 161, 209, 33, 81, 129, 177, 225,
5, 53, 101, 149, 197, 21, 69, 117, 165, 213, 37, 85, 133, 181, 229,
9, 57, 105, 153, 201, 25, 73, 121, 169, 217, 41, 89, 137, 185, 233,
13, 61, 109, 157, 205, 29, 77, 125, 173, 221, 45, 93, 141, 189, 237,
2, 50, 98, 146, 194, 18, 66, 114, 162, 210, 34, 82, 130, 178, 226,
6, 54, 102, 150, 198, 22, 70, 118, 166, 214, 38, 86, 134, 182, 230,
10, 58, 106, 154, 202, 26, 74, 122, 170, 218, 42, 90, 138, 186, 234,
14, 62, 110, 158, 206, 30, 78, 126, 174, 222, 46, 94, 142, 190, 238,
3, 51, 99, 147, 195, 19, 67, 115, 163, 211, 35, 83, 131, 179, 227,
7, 55, 103, 151, 199, 23, 71, 119, 167, 215, 39, 87, 135, 183, 231,
11, 59, 107, 155, 203, 27, 75, 123, 171, 219, 43, 91, 139, 187, 235,
15, 63, 111, 159, 207, 31, 79, 127, 175, 223, 47, 95, 143, 191, 239,
};
#endif
#ifndef FFT_BITREV120
#define FFT_BITREV120
static const opus_int16 fft_bitrev120[120] = {
0, 30, 60, 90, 15, 45, 75, 105, 5, 35, 65, 95, 20, 50, 80,
110, 10, 40, 70, 100, 25, 55, 85, 115, 1, 31, 61, 91, 16, 46,
76, 106, 6, 36, 66, 96, 21, 51, 81, 111, 11, 41, 71, 101, 26,
56, 86, 116, 2, 32, 62, 92, 17, 47, 77, 107, 7, 37, 67, 97,
22, 52, 82, 112, 12, 42, 72, 102, 27, 57, 87, 117, 3, 33, 63,
93, 18, 48, 78, 108, 8, 38, 68, 98, 23, 53, 83, 113, 13, 43,
73, 103, 28, 58, 88, 118, 4, 34, 64, 94, 19, 49, 79, 109, 9,
39, 69, 99, 24, 54, 84, 114, 14, 44, 74, 104, 29, 59, 89, 119,
0, 24, 48, 72, 96, 8, 32, 56, 80, 104, 16, 40, 64, 88, 112,
4, 28, 52, 76, 100, 12, 36, 60, 84, 108, 20, 44, 68, 92, 116,
1, 25, 49, 73, 97, 9, 33, 57, 81, 105, 17, 41, 65, 89, 113,
5, 29, 53, 77, 101, 13, 37, 61, 85, 109, 21, 45, 69, 93, 117,
2, 26, 50, 74, 98, 10, 34, 58, 82, 106, 18, 42, 66, 90, 114,
6, 30, 54, 78, 102, 14, 38, 62, 86, 110, 22, 46, 70, 94, 118,
3, 27, 51, 75, 99, 11, 35, 59, 83, 107, 19, 43, 67, 91, 115,
7, 31, 55, 79, 103, 15, 39, 63, 87, 111, 23, 47, 71, 95, 119,
};
#endif
#ifndef FFT_BITREV60
#define FFT_BITREV60
static const opus_int16 fft_bitrev60[60] = {
0, 15, 30, 45, 5, 20, 35, 50, 10, 25, 40, 55, 1, 16, 31,
46, 6, 21, 36, 51, 11, 26, 41, 56, 2, 17, 32, 47, 7, 22,
37, 52, 12, 27, 42, 57, 3, 18, 33, 48, 8, 23, 38, 53, 13,
28, 43, 58, 4, 19, 34, 49, 9, 24, 39, 54, 14, 29, 44, 59,
0, 12, 24, 36, 48, 4, 16, 28, 40, 52, 8, 20, 32, 44, 56,
1, 13, 25, 37, 49, 5, 17, 29, 41, 53, 9, 21, 33, 45, 57,
2, 14, 26, 38, 50, 6, 18, 30, 42, 54, 10, 22, 34, 46, 58,
3, 15, 27, 39, 51, 7, 19, 31, 43, 55, 11, 23, 35, 47, 59,
};
#endif
@ -428,9 +433,14 @@ static const kiss_fft_state fft_state48000_960_0 = {
480, /* nfft */
0.002083333f, /* scale */
-1, /* shift */
{4, 120, 4, 30, 2, 15, 3, 5, 5, 1, 0, 0, 0, 0, 0, 0, }, /* factors */
{5, 96, 3, 32, 4, 8, 2, 4, 4, 1, 0, 0, 0, 0, 0, 0, }, /* factors */
fft_bitrev480, /* bitrev */
fft_twiddles48000_960, /* bitrev */
#ifdef OVERRIDE_FFT
(arch_fft_state *)&cfg_arch_480,
#else
NULL,
#endif
};
#endif
@ -440,9 +450,14 @@ static const kiss_fft_state fft_state48000_960_1 = {
240, /* nfft */
0.004166667f, /* scale */
1, /* shift */
{4, 60, 4, 15, 3, 5, 5, 1, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
{5, 48, 3, 16, 4, 4, 4, 1, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
fft_bitrev240, /* bitrev */
fft_twiddles48000_960, /* bitrev */
#ifdef OVERRIDE_FFT
(arch_fft_state *)&cfg_arch_240,
#else
NULL,
#endif
};
#endif
@ -452,9 +467,14 @@ static const kiss_fft_state fft_state48000_960_2 = {
120, /* nfft */
0.008333333f, /* scale */
2, /* shift */
{4, 30, 2, 15, 3, 5, 5, 1, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
{5, 24, 3, 8, 2, 4, 4, 1, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
fft_bitrev120, /* bitrev */
fft_twiddles48000_960, /* bitrev */
#ifdef OVERRIDE_FFT
(arch_fft_state *)&cfg_arch_120,
#else
NULL,
#endif
};
#endif
@ -464,9 +484,14 @@ static const kiss_fft_state fft_state48000_960_3 = {
60, /* nfft */
0.016666667f, /* scale */
3, /* shift */
{4, 15, 3, 5, 5, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
{5, 12, 3, 4, 4, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, }, /* factors */
fft_bitrev60, /* bitrev */
fft_twiddles48000_960, /* bitrev */
#ifdef OVERRIDE_FFT
(arch_fft_state *)&cfg_arch_60,
#else
NULL,
#endif
};
#endif
@ -474,104 +499,368 @@ fft_twiddles48000_960, /* bitrev */
#ifndef MDCT_TWIDDLES960
#define MDCT_TWIDDLES960
static const opus_val16 mdct_twiddles960[481] = {
1.0000000f, 0.99999465f, 0.99997858f, 0.99995181f, 0.99991433f,
0.99986614f, 0.99980724f, 0.99973764f, 0.99965732f, 0.99956631f,
0.99946459f, 0.99935216f, 0.99922904f, 0.99909521f, 0.99895068f,
0.99879546f, 0.99862953f, 0.99845292f, 0.99826561f, 0.99806761f,
0.99785892f, 0.99763955f, 0.99740949f, 0.99716875f, 0.99691733f,
0.99665524f, 0.99638247f, 0.99609903f, 0.99580493f, 0.99550016f,
0.99518473f, 0.99485864f, 0.99452190f, 0.99417450f, 0.99381646f,
0.99344778f, 0.99306846f, 0.99267850f, 0.99227791f, 0.99186670f,
0.99144486f, 0.99101241f, 0.99056934f, 0.99011566f, 0.98965139f,
0.98917651f, 0.98869104f, 0.98819498f, 0.98768834f, 0.98717112f,
0.98664333f, 0.98610497f, 0.98555606f, 0.98499659f, 0.98442657f,
0.98384600f, 0.98325491f, 0.98265328f, 0.98204113f, 0.98141846f,
0.98078528f, 0.98014159f, 0.97948742f, 0.97882275f, 0.97814760f,
0.97746197f, 0.97676588f, 0.97605933f, 0.97534232f, 0.97461487f,
0.97387698f, 0.97312866f, 0.97236992f, 0.97160077f, 0.97082121f,
0.97003125f, 0.96923091f, 0.96842019f, 0.96759909f, 0.96676764f,
0.96592582f, 0.96507367f, 0.96421118f, 0.96333837f, 0.96245523f,
0.96156180f, 0.96065806f, 0.95974403f, 0.95881973f, 0.95788517f,
0.95694034f, 0.95598526f, 0.95501995f, 0.95404440f, 0.95305864f,
0.95206267f, 0.95105651f, 0.95004016f, 0.94901364f, 0.94797697f,
0.94693013f, 0.94587315f, 0.94480604f, 0.94372882f, 0.94264149f,
0.94154406f, 0.94043656f, 0.93931897f, 0.93819133f, 0.93705365f,
0.93590592f, 0.93474818f, 0.93358042f, 0.93240268f, 0.93121493f,
0.93001722f, 0.92880955f, 0.92759193f, 0.92636438f, 0.92512690f,
0.92387953f, 0.92262225f, 0.92135509f, 0.92007809f, 0.91879121f,
0.91749449f, 0.91618795f, 0.91487161f, 0.91354545f, 0.91220952f,
0.91086382f, 0.90950836f, 0.90814316f, 0.90676824f, 0.90538363f,
0.90398929f, 0.90258528f, 0.90117161f, 0.89974828f, 0.89831532f,
0.89687273f, 0.89542055f, 0.89395877f, 0.89248742f, 0.89100652f,
0.88951606f, 0.88801610f, 0.88650661f, 0.88498764f, 0.88345918f,
0.88192125f, 0.88037390f, 0.87881711f, 0.87725090f, 0.87567531f,
0.87409035f, 0.87249599f, 0.87089232f, 0.86927933f, 0.86765699f,
0.86602540f, 0.86438453f, 0.86273437f, 0.86107503f, 0.85940641f,
0.85772862f, 0.85604161f, 0.85434547f, 0.85264014f, 0.85092572f,
0.84920218f, 0.84746955f, 0.84572781f, 0.84397704f, 0.84221721f,
0.84044838f, 0.83867056f, 0.83688375f, 0.83508799f, 0.83328325f,
0.83146961f, 0.82964704f, 0.82781562f, 0.82597530f, 0.82412620f,
0.82226820f, 0.82040144f, 0.81852589f, 0.81664154f, 0.81474847f,
0.81284665f, 0.81093620f, 0.80901698f, 0.80708914f, 0.80515262f,
0.80320752f, 0.80125378f, 0.79929149f, 0.79732067f, 0.79534125f,
0.79335335f, 0.79135691f, 0.78935204f, 0.78733867f, 0.78531691f,
0.78328674f, 0.78124818f, 0.77920122f, 0.77714595f, 0.77508232f,
0.77301043f, 0.77093026f, 0.76884183f, 0.76674517f, 0.76464026f,
0.76252720f, 0.76040593f, 0.75827656f, 0.75613907f, 0.75399349f,
0.75183978f, 0.74967807f, 0.74750833f, 0.74533054f, 0.74314481f,
0.74095112f, 0.73874950f, 0.73653993f, 0.73432251f, 0.73209718f,
0.72986405f, 0.72762307f, 0.72537438f, 0.72311787f, 0.72085359f,
0.71858162f, 0.71630192f, 0.71401459f, 0.71171956f, 0.70941701f,
0.70710677f, 0.70478900f, 0.70246363f, 0.70013079f, 0.69779041f,
0.69544260f, 0.69308738f, 0.69072466f, 0.68835458f, 0.68597709f,
0.68359229f, 0.68120013f, 0.67880072f, 0.67639404f, 0.67398011f,
0.67155892f, 0.66913059f, 0.66669509f, 0.66425240f, 0.66180265f,
0.65934581f, 0.65688191f, 0.65441092f, 0.65193298f, 0.64944801f,
0.64695613f, 0.64445727f, 0.64195160f, 0.63943902f, 0.63691954f,
0.63439328f, 0.63186019f, 0.62932037f, 0.62677377f, 0.62422055f,
0.62166055f, 0.61909394f, 0.61652065f, 0.61394081f, 0.61135435f,
0.60876139f, 0.60616195f, 0.60355593f, 0.60094349f, 0.59832457f,
0.59569929f, 0.59306758f, 0.59042957f, 0.58778523f, 0.58513460f,
0.58247766f, 0.57981452f, 0.57714518f, 0.57446961f, 0.57178793f,
0.56910013f, 0.56640624f, 0.56370623f, 0.56100023f, 0.55828818f,
0.55557020f, 0.55284627f, 0.55011641f, 0.54738067f, 0.54463901f,
0.54189157f, 0.53913828f, 0.53637921f, 0.53361450f, 0.53084398f,
0.52806787f, 0.52528601f, 0.52249852f, 0.51970543f, 0.51690688f,
0.51410279f, 0.51129310f, 0.50847793f, 0.50565732f, 0.50283139f,
0.49999997f, 0.49716321f, 0.49432122f, 0.49147383f, 0.48862118f,
0.48576340f, 0.48290042f, 0.48003216f, 0.47715876f, 0.47428025f,
0.47139677f, 0.46850813f, 0.46561448f, 0.46271584f, 0.45981235f,
0.45690383f, 0.45399042f, 0.45107214f, 0.44814915f, 0.44522124f,
0.44228868f, 0.43935137f, 0.43640926f, 0.43346247f, 0.43051104f,
0.42755511f, 0.42459449f, 0.42162932f, 0.41865964f, 0.41568558f,
0.41270697f, 0.40972393f, 0.40673661f, 0.40374494f, 0.40074884f,
0.39774844f, 0.39474390f, 0.39173501f, 0.38872193f, 0.38570469f,
0.38268343f, 0.37965796f, 0.37662842f, 0.37359496f, 0.37055739f,
0.36751585f, 0.36447038f, 0.36142122f, 0.35836797f, 0.35531089f,
0.35225000f, 0.34918544f, 0.34611704f, 0.34304493f, 0.33996926f,
0.33688983f, 0.33380680f, 0.33072019f, 0.32763015f, 0.32453650f,
0.32143936f, 0.31833890f, 0.31523503f, 0.31212767f, 0.30901696f,
0.30590306f, 0.30278577f, 0.29966524f, 0.29654150f, 0.29341470f,
0.29028464f, 0.28715147f, 0.28401522f, 0.28087605f, 0.27773376f,
0.27458861f, 0.27144052f, 0.26828940f, 0.26513541f, 0.26197859f,
0.25881907f, 0.25565666f, 0.25249152f, 0.24932367f, 0.24615327f,
0.24298012f, 0.23980436f, 0.23662604f, 0.23344530f, 0.23026206f,
0.22707623f, 0.22388809f, 0.22069744f, 0.21750443f, 0.21430908f,
0.21111156f, 0.20791165f, 0.20470953f, 0.20150520f, 0.19829884f,
0.19509024f, 0.19187955f, 0.18866692f, 0.18545227f, 0.18223552f,
0.17901681f, 0.17579631f, 0.17257380f, 0.16934945f, 0.16612328f,
0.16289546f, 0.15966577f, 0.15643437f, 0.15320141f, 0.14996669f,
0.14673037f, 0.14349260f, 0.14025329f, 0.13701235f, 0.13376995f,
0.13052612f, 0.12728101f, 0.12403442f, 0.12078650f, 0.11753740f,
0.11428693f, 0.11103523f, 0.10778234f, 0.10452842f, 0.10127326f,
0.098017137f, 0.094759842f, 0.091501652f, 0.088242363f, 0.084982129f,
0.081721103f, 0.078459084f, 0.075196224f, 0.071932560f, 0.068668243f,
0.065403073f, 0.062137201f, 0.058870665f, 0.055603617f, 0.052335974f,
0.049067651f, 0.045798921f, 0.042529582f, 0.039259788f, 0.035989573f,
0.032719092f, 0.029448142f, 0.026176876f, 0.022905329f, 0.019633657f,
0.016361655f, 0.013089478f, 0.0098171604f, 0.0065449764f, 0.0032724839f,
-4.3711390e-08f, };
static const opus_val16 mdct_twiddles960[1800] = {
0.99999994f, 0.99999321f, 0.99997580f, 0.99994773f, 0.99990886f,
0.99985933f, 0.99979913f, 0.99972820f, 0.99964654f, 0.99955416f,
0.99945110f, 0.99933738f, 0.99921292f, 0.99907774f, 0.99893188f,
0.99877530f, 0.99860805f, 0.99843007f, 0.99824142f, 0.99804211f,
0.99783206f, 0.99761140f, 0.99737996f, 0.99713790f, 0.99688518f,
0.99662173f, 0.99634761f, 0.99606287f, 0.99576741f, 0.99546129f,
0.99514455f, 0.99481714f, 0.99447906f, 0.99413031f, 0.99377096f,
0.99340093f, 0.99302030f, 0.99262899f, 0.99222708f, 0.99181455f,
0.99139136f, 0.99095762f, 0.99051321f, 0.99005818f, 0.98959261f,
0.98911643f, 0.98862964f, 0.98813224f, 0.98762429f, 0.98710573f,
0.98657662f, 0.98603696f, 0.98548669f, 0.98492593f, 0.98435456f,
0.98377270f, 0.98318028f, 0.98257732f, 0.98196387f, 0.98133987f,
0.98070538f, 0.98006040f, 0.97940493f, 0.97873890f, 0.97806245f,
0.97737551f, 0.97667813f, 0.97597027f, 0.97525197f, 0.97452319f,
0.97378403f, 0.97303438f, 0.97227436f, 0.97150391f, 0.97072303f,
0.96993178f, 0.96913016f, 0.96831810f, 0.96749574f, 0.96666300f,
0.96581990f, 0.96496642f, 0.96410263f, 0.96322852f, 0.96234411f,
0.96144938f, 0.96054435f, 0.95962906f, 0.95870346f, 0.95776761f,
0.95682150f, 0.95586514f, 0.95489854f, 0.95392174f, 0.95293468f,
0.95193744f, 0.95093000f, 0.94991243f, 0.94888461f, 0.94784665f,
0.94679856f, 0.94574034f, 0.94467193f, 0.94359344f, 0.94250488f,
0.94140619f, 0.94029742f, 0.93917859f, 0.93804967f, 0.93691075f,
0.93576175f, 0.93460274f, 0.93343377f, 0.93225473f, 0.93106574f,
0.92986679f, 0.92865789f, 0.92743903f, 0.92621022f, 0.92497152f,
0.92372292f, 0.92246443f, 0.92119598f, 0.91991776f, 0.91862965f,
0.91733170f, 0.91602397f, 0.91470635f, 0.91337901f, 0.91204184f,
0.91069490f, 0.90933824f, 0.90797186f, 0.90659571f, 0.90520984f,
0.90381432f, 0.90240908f, 0.90099424f, 0.89956969f, 0.89813554f,
0.89669174f, 0.89523834f, 0.89377540f, 0.89230281f, 0.89082074f,
0.88932908f, 0.88782793f, 0.88631725f, 0.88479710f, 0.88326746f,
0.88172835f, 0.88017982f, 0.87862182f, 0.87705445f, 0.87547767f,
0.87389153f, 0.87229604f, 0.87069118f, 0.86907703f, 0.86745358f,
0.86582077f, 0.86417878f, 0.86252749f, 0.86086690f, 0.85919720f,
0.85751826f, 0.85583007f, 0.85413277f, 0.85242635f, 0.85071075f,
0.84898609f, 0.84725231f, 0.84550947f, 0.84375757f, 0.84199661f,
0.84022665f, 0.83844769f, 0.83665979f, 0.83486289f, 0.83305705f,
0.83124226f, 0.82941860f, 0.82758605f, 0.82574469f, 0.82389444f,
0.82203537f, 0.82016748f, 0.81829083f, 0.81640542f, 0.81451124f,
0.81260836f, 0.81069672f, 0.80877650f, 0.80684757f, 0.80490994f,
0.80296379f, 0.80100900f, 0.79904562f, 0.79707366f, 0.79509324f,
0.79310423f, 0.79110676f, 0.78910083f, 0.78708643f, 0.78506362f,
0.78303236f, 0.78099275f, 0.77894479f, 0.77688843f, 0.77482378f,
0.77275085f, 0.77066964f, 0.76858020f, 0.76648247f, 0.76437658f,
0.76226246f, 0.76014024f, 0.75800985f, 0.75587130f, 0.75372469f,
0.75157005f, 0.74940729f, 0.74723655f, 0.74505776f, 0.74287105f,
0.74067634f, 0.73847371f, 0.73626316f, 0.73404479f, 0.73181850f,
0.72958434f, 0.72734243f, 0.72509271f, 0.72283524f, 0.72057003f,
0.71829706f, 0.71601641f, 0.71372813f, 0.71143216f, 0.70912862f,
0.70681745f, 0.70449871f, 0.70217246f, 0.69983864f, 0.69749737f,
0.69514859f, 0.69279242f, 0.69042879f, 0.68805778f, 0.68567938f,
0.68329364f, 0.68090063f, 0.67850029f, 0.67609268f, 0.67367786f,
0.67125577f, 0.66882652f, 0.66639012f, 0.66394657f, 0.66149592f,
0.65903819f, 0.65657341f, 0.65410155f, 0.65162271f, 0.64913690f,
0.64664418f, 0.64414448f, 0.64163786f, 0.63912445f, 0.63660413f,
0.63407701f, 0.63154310f, 0.62900239f, 0.62645501f, 0.62390089f,
0.62134010f, 0.61877263f, 0.61619854f, 0.61361790f, 0.61103064f,
0.60843682f, 0.60583651f, 0.60322970f, 0.60061646f, 0.59799677f,
0.59537065f, 0.59273821f, 0.59009939f, 0.58745426f, 0.58480281f,
0.58214509f, 0.57948118f, 0.57681108f, 0.57413477f, 0.57145232f,
0.56876373f, 0.56606907f, 0.56336832f, 0.56066155f, 0.55794877f,
0.55523002f, 0.55250537f, 0.54977477f, 0.54703826f, 0.54429591f,
0.54154772f, 0.53879374f, 0.53603399f, 0.53326851f, 0.53049731f,
0.52772039f, 0.52493787f, 0.52214974f, 0.51935595f, 0.51655668f,
0.51375180f, 0.51094145f, 0.50812566f, 0.50530440f, 0.50247771f,
0.49964568f, 0.49680826f, 0.49396557f, 0.49111754f, 0.48826426f,
0.48540577f, 0.48254207f, 0.47967321f, 0.47679919f, 0.47392011f,
0.47103590f, 0.46814668f, 0.46525243f, 0.46235323f, 0.45944905f,
0.45653993f, 0.45362595f, 0.45070711f, 0.44778344f, 0.44485497f,
0.44192174f, 0.43898380f, 0.43604112f, 0.43309379f, 0.43014181f,
0.42718524f, 0.42422408f, 0.42125839f, 0.41828820f, 0.41531351f,
0.41233435f, 0.40935081f, 0.40636289f, 0.40337059f, 0.40037400f,
0.39737311f, 0.39436796f, 0.39135858f, 0.38834500f, 0.38532731f,
0.38230544f, 0.37927949f, 0.37624949f, 0.37321547f, 0.37017745f,
0.36713544f, 0.36408952f, 0.36103970f, 0.35798600f, 0.35492846f,
0.35186714f, 0.34880206f, 0.34573323f, 0.34266070f, 0.33958447f,
0.33650464f, 0.33342120f, 0.33033419f, 0.32724363f, 0.32414958f,
0.32105204f, 0.31795108f, 0.31484672f, 0.31173897f, 0.30862790f,
0.30551350f, 0.30239585f, 0.29927495f, 0.29615086f, 0.29302359f,
0.28989318f, 0.28675964f, 0.28362307f, 0.28048345f, 0.27734083f,
0.27419522f, 0.27104670f, 0.26789525f, 0.26474094f, 0.26158381f,
0.25842386f, 0.25526115f, 0.25209570f, 0.24892756f, 0.24575676f,
0.24258332f, 0.23940729f, 0.23622867f, 0.23304754f, 0.22986393f,
0.22667783f, 0.22348931f, 0.22029841f, 0.21710514f, 0.21390954f,
0.21071166f, 0.20751151f, 0.20430915f, 0.20110460f, 0.19789790f,
0.19468907f, 0.19147816f, 0.18826519f, 0.18505022f, 0.18183327f,
0.17861435f, 0.17539354f, 0.17217083f, 0.16894630f, 0.16571994f,
0.16249183f, 0.15926196f, 0.15603039f, 0.15279715f, 0.14956227f,
0.14632578f, 0.14308774f, 0.13984816f, 0.13660708f, 0.13336454f,
0.13012058f, 0.12687522f, 0.12362850f, 0.12038045f, 0.11713112f,
0.11388054f, 0.11062872f, 0.10737573f, 0.10412160f, 0.10086634f,
0.097609997f, 0.094352618f, 0.091094226f, 0.087834857f, 0.084574550f,
0.081313334f, 0.078051247f, 0.074788325f, 0.071524605f, 0.068260118f,
0.064994894f, 0.061728980f, 0.058462404f, 0.055195201f, 0.051927410f,
0.048659060f, 0.045390189f, 0.042120833f, 0.038851023f, 0.035580799f,
0.032310195f, 0.029039243f, 0.025767982f, 0.022496443f, 0.019224664f,
0.015952680f, 0.012680525f, 0.0094082337f, 0.0061358409f, 0.0028633832f,
-0.00040910527f, -0.0036815894f, -0.0069540343f, -0.010226404f, -0.013498665f,
-0.016770782f, -0.020042717f, -0.023314439f, -0.026585912f, -0.029857099f,
-0.033127967f, -0.036398482f, -0.039668605f, -0.042938303f, -0.046207540f,
-0.049476285f, -0.052744497f, -0.056012146f, -0.059279196f, -0.062545612f,
-0.065811358f, -0.069076397f, -0.072340697f, -0.075604223f, -0.078866936f,
-0.082128808f, -0.085389800f, -0.088649876f, -0.091909006f, -0.095167145f,
-0.098424271f, -0.10168034f, -0.10493532f, -0.10818918f, -0.11144188f,
-0.11469338f, -0.11794366f, -0.12119267f, -0.12444039f, -0.12768677f,
-0.13093179f, -0.13417540f, -0.13741758f, -0.14065829f, -0.14389749f,
-0.14713514f, -0.15037122f, -0.15360570f, -0.15683852f, -0.16006967f,
-0.16329910f, -0.16652679f, -0.16975269f, -0.17297678f, -0.17619900f,
-0.17941935f, -0.18263777f, -0.18585424f, -0.18906870f, -0.19228116f,
-0.19549155f, -0.19869985f, -0.20190603f, -0.20511003f, -0.20831184f,
-0.21151142f, -0.21470875f, -0.21790376f, -0.22109644f, -0.22428675f,
-0.22747467f, -0.23066014f, -0.23384315f, -0.23702365f, -0.24020162f,
-0.24337701f, -0.24654980f, -0.24971995f, -0.25288740f, -0.25605217f,
-0.25921419f, -0.26237345f, -0.26552987f, -0.26868346f, -0.27183419f,
-0.27498198f, -0.27812684f, -0.28126872f, -0.28440759f, -0.28754342f,
-0.29067615f, -0.29380578f, -0.29693225f, -0.30005556f, -0.30317566f,
-0.30629250f, -0.30940607f, -0.31251630f, -0.31562322f, -0.31872672f,
-0.32182685f, -0.32492352f, -0.32801670f, -0.33110636f, -0.33419248f,
-0.33727503f, -0.34035397f, -0.34342924f, -0.34650084f, -0.34956875f,
-0.35263291f, -0.35569328f, -0.35874987f, -0.36180258f, -0.36485144f,
-0.36789638f, -0.37093741f, -0.37397444f, -0.37700745f, -0.38003644f,
-0.38306138f, -0.38608220f, -0.38909888f, -0.39211139f, -0.39511973f,
-0.39812380f, -0.40112361f, -0.40411916f, -0.40711036f, -0.41009718f,
-0.41307965f, -0.41605768f, -0.41903123f, -0.42200032f, -0.42496487f,
-0.42792490f, -0.43088034f, -0.43383113f, -0.43677729f, -0.43971881f,
-0.44265559f, -0.44558764f, -0.44851488f, -0.45143735f, -0.45435500f,
-0.45726776f, -0.46017563f, -0.46307856f, -0.46597654f, -0.46886954f,
-0.47175750f, -0.47464043f, -0.47751826f, -0.48039100f, -0.48325855f,
-0.48612097f, -0.48897815f, -0.49183011f, -0.49467680f, -0.49751821f,
-0.50035429f, -0.50318497f, -0.50601029f, -0.50883019f, -0.51164466f,
-0.51445359f, -0.51725709f, -0.52005500f, -0.52284735f, -0.52563411f,
-0.52841520f, -0.53119069f, -0.53396046f, -0.53672451f, -0.53948283f,
-0.54223537f, -0.54498214f, -0.54772300f, -0.55045801f, -0.55318713f,
-0.55591035f, -0.55862761f, -0.56133890f, -0.56404412f, -0.56674337f,
-0.56943649f, -0.57212353f, -0.57480448f, -0.57747924f, -0.58014780f,
-0.58281022f, -0.58546633f, -0.58811617f, -0.59075975f, -0.59339696f,
-0.59602785f, -0.59865236f, -0.60127044f, -0.60388207f, -0.60648727f,
-0.60908598f, -0.61167812f, -0.61426371f, -0.61684275f, -0.61941516f,
-0.62198097f, -0.62454009f, -0.62709254f, -0.62963831f, -0.63217729f,
-0.63470948f, -0.63723493f, -0.63975352f, -0.64226526f, -0.64477009f,
-0.64726806f, -0.64975911f, -0.65224314f, -0.65472025f, -0.65719032f,
-0.65965337f, -0.66210932f, -0.66455823f, -0.66700000f, -0.66943461f,
-0.67186207f, -0.67428231f, -0.67669535f, -0.67910111f, -0.68149966f,
-0.68389088f, -0.68627477f, -0.68865126f, -0.69102043f, -0.69338220f,
-0.69573659f, -0.69808346f, -0.70042288f, -0.70275480f, -0.70507920f,
-0.70739603f, -0.70970529f, -0.71200693f, -0.71430099f, -0.71658736f,
-0.71886611f, -0.72113711f, -0.72340041f, -0.72565591f, -0.72790372f,
-0.73014367f, -0.73237586f, -0.73460019f, -0.73681659f, -0.73902518f,
-0.74122584f, -0.74341851f, -0.74560326f, -0.74778003f, -0.74994880f,
-0.75210953f, -0.75426215f, -0.75640678f, -0.75854325f, -0.76067162f,
-0.76279181f, -0.76490390f, -0.76700771f, -0.76910341f, -0.77119076f,
-0.77326995f, -0.77534080f, -0.77740335f, -0.77945763f, -0.78150350f,
-0.78354102f, -0.78557014f, -0.78759086f, -0.78960317f, -0.79160696f,
-0.79360235f, -0.79558921f, -0.79756755f, -0.79953730f, -0.80149853f,
-0.80345118f, -0.80539525f, -0.80733067f, -0.80925739f, -0.81117553f,
-0.81308490f, -0.81498563f, -0.81687760f, -0.81876087f, -0.82063532f,
-0.82250100f, -0.82435787f, -0.82620591f, -0.82804507f, -0.82987541f,
-0.83169687f, -0.83350939f, -0.83531296f, -0.83710766f, -0.83889335f,
-0.84067005f, -0.84243774f, -0.84419644f, -0.84594607f, -0.84768665f,
-0.84941816f, -0.85114056f, -0.85285389f, -0.85455805f, -0.85625303f,
-0.85793889f, -0.85961550f, -0.86128294f, -0.86294121f, -0.86459017f,
-0.86622989f, -0.86786032f, -0.86948150f, -0.87109333f, -0.87269586f,
-0.87428904f, -0.87587279f, -0.87744725f, -0.87901229f, -0.88056785f,
-0.88211405f, -0.88365078f, -0.88517809f, -0.88669586f, -0.88820416f,
-0.88970292f, -0.89119220f, -0.89267188f, -0.89414203f, -0.89560264f,
-0.89705360f, -0.89849502f, -0.89992678f, -0.90134889f, -0.90276134f,
-0.90416414f, -0.90555727f, -0.90694070f, -0.90831441f, -0.90967834f,
-0.91103262f, -0.91237706f, -0.91371179f, -0.91503674f, -0.91635185f,
-0.91765714f, -0.91895264f, -0.92023826f, -0.92151409f, -0.92277998f,
-0.92403603f, -0.92528218f, -0.92651838f, -0.92774469f, -0.92896110f,
-0.93016750f, -0.93136400f, -0.93255049f, -0.93372697f, -0.93489349f,
-0.93604994f, -0.93719643f, -0.93833286f, -0.93945926f, -0.94057560f,
-0.94168180f, -0.94277799f, -0.94386405f, -0.94494003f, -0.94600588f,
-0.94706154f, -0.94810712f, -0.94914252f, -0.95016778f, -0.95118284f,
-0.95218778f, -0.95318246f, -0.95416695f, -0.95514119f, -0.95610523f,
-0.95705903f, -0.95800257f, -0.95893586f, -0.95985889f, -0.96077162f,
-0.96167403f, -0.96256620f, -0.96344805f, -0.96431959f, -0.96518075f,
-0.96603161f, -0.96687216f, -0.96770233f, -0.96852213f, -0.96933156f,
-0.97013056f, -0.97091925f, -0.97169751f, -0.97246534f, -0.97322279f,
-0.97396982f, -0.97470641f, -0.97543252f, -0.97614825f, -0.97685349f,
-0.97754824f, -0.97823256f, -0.97890645f, -0.97956979f, -0.98022264f,
-0.98086500f, -0.98149687f, -0.98211825f, -0.98272908f, -0.98332942f,
-0.98391914f, -0.98449844f, -0.98506713f, -0.98562527f, -0.98617285f,
-0.98670989f, -0.98723638f, -0.98775226f, -0.98825759f, -0.98875231f,
-0.98923647f, -0.98971003f, -0.99017298f, -0.99062532f, -0.99106705f,
-0.99149817f, -0.99191868f, -0.99232858f, -0.99272782f, -0.99311644f,
-0.99349445f, -0.99386179f, -0.99421853f, -0.99456459f, -0.99489999f,
-0.99522477f, -0.99553883f, -0.99584228f, -0.99613506f, -0.99641716f,
-0.99668860f, -0.99694937f, -0.99719942f, -0.99743885f, -0.99766755f,
-0.99788558f, -0.99809295f, -0.99828959f, -0.99847561f, -0.99865085f,
-0.99881548f, -0.99896932f, -0.99911255f, -0.99924499f, -0.99936682f,
-0.99947786f, -0.99957830f, -0.99966794f, -0.99974692f, -0.99981517f,
-0.99987274f, -0.99991959f, -0.99995571f, -0.99998116f, -0.99999589f,
0.99999964f, 0.99997288f, 0.99990326f, 0.99979085f, 0.99963558f,
0.99943751f, 0.99919659f, 0.99891287f, 0.99858636f, 0.99821711f,
0.99780506f, 0.99735034f, 0.99685282f, 0.99631262f, 0.99572974f,
0.99510419f, 0.99443603f, 0.99372530f, 0.99297196f, 0.99217612f,
0.99133772f, 0.99045694f, 0.98953366f, 0.98856801f, 0.98756003f,
0.98650974f, 0.98541719f, 0.98428243f, 0.98310548f, 0.98188645f,
0.98062533f, 0.97932225f, 0.97797716f, 0.97659022f, 0.97516143f,
0.97369087f, 0.97217858f, 0.97062469f, 0.96902919f, 0.96739221f,
0.96571374f, 0.96399397f, 0.96223283f, 0.96043050f, 0.95858705f,
0.95670253f, 0.95477700f, 0.95281059f, 0.95080340f, 0.94875544f,
0.94666684f, 0.94453770f, 0.94236809f, 0.94015813f, 0.93790787f,
0.93561745f, 0.93328691f, 0.93091643f, 0.92850608f, 0.92605597f,
0.92356616f, 0.92103678f, 0.91846794f, 0.91585976f, 0.91321236f,
0.91052586f, 0.90780038f, 0.90503591f, 0.90223277f, 0.89939094f,
0.89651060f, 0.89359182f, 0.89063478f, 0.88763964f, 0.88460642f,
0.88153529f, 0.87842643f, 0.87527996f, 0.87209594f, 0.86887461f,
0.86561602f, 0.86232042f, 0.85898781f, 0.85561842f, 0.85221243f,
0.84876984f, 0.84529096f, 0.84177583f, 0.83822471f, 0.83463764f,
0.83101481f, 0.82735640f, 0.82366252f, 0.81993335f, 0.81616908f,
0.81236988f, 0.80853581f, 0.80466717f, 0.80076402f, 0.79682660f,
0.79285502f, 0.78884947f, 0.78481019f, 0.78073722f, 0.77663082f,
0.77249116f, 0.76831841f, 0.76411277f, 0.75987434f, 0.75560343f,
0.75130010f, 0.74696463f, 0.74259710f, 0.73819780f, 0.73376691f,
0.72930455f, 0.72481096f, 0.72028631f, 0.71573079f, 0.71114463f,
0.70652801f, 0.70188117f, 0.69720417f, 0.69249737f, 0.68776089f,
0.68299496f, 0.67819971f, 0.67337549f, 0.66852236f, 0.66364062f,
0.65873051f, 0.65379208f, 0.64882571f, 0.64383155f, 0.63880974f,
0.63376063f, 0.62868434f, 0.62358117f, 0.61845124f, 0.61329484f,
0.60811216f, 0.60290343f, 0.59766883f, 0.59240872f, 0.58712316f,
0.58181250f, 0.57647687f, 0.57111657f, 0.56573176f, 0.56032276f,
0.55488980f, 0.54943299f, 0.54395270f, 0.53844911f, 0.53292239f,
0.52737290f, 0.52180082f, 0.51620632f, 0.51058978f, 0.50495136f,
0.49929130f, 0.49360985f, 0.48790723f, 0.48218375f, 0.47643960f,
0.47067502f, 0.46489030f, 0.45908567f, 0.45326138f, 0.44741765f,
0.44155475f, 0.43567297f, 0.42977250f, 0.42385364f, 0.41791660f,
0.41196167f, 0.40598908f, 0.39999911f, 0.39399201f, 0.38796803f,
0.38192743f, 0.37587047f, 0.36979741f, 0.36370850f, 0.35760403f,
0.35148421f, 0.34534934f, 0.33919969f, 0.33303553f, 0.32685706f,
0.32066461f, 0.31445843f, 0.30823877f, 0.30200592f, 0.29576012f,
0.28950164f, 0.28323078f, 0.27694780f, 0.27065292f, 0.26434645f,
0.25802869f, 0.25169984f, 0.24536023f, 0.23901010f, 0.23264973f,
0.22627939f, 0.21989937f, 0.21350993f, 0.20711134f, 0.20070387f,
0.19428782f, 0.18786344f, 0.18143101f, 0.17499080f, 0.16854310f,
0.16208819f, 0.15562633f, 0.14915779f, 0.14268288f, 0.13620184f,
0.12971498f, 0.12322257f, 0.11672486f, 0.11022217f, 0.10371475f,
0.097202882f, 0.090686858f, 0.084166944f, 0.077643424f, 0.071116582f,
0.064586692f, 0.058054037f, 0.051518895f, 0.044981543f, 0.038442269f,
0.031901345f, 0.025359053f, 0.018815678f, 0.012271495f, 0.0057267868f,
-0.00081816671f, -0.0073630852f, -0.013907688f, -0.020451695f, -0.026994826f,
-0.033536803f, -0.040077340f, -0.046616159f, -0.053152986f, -0.059687532f,
-0.066219524f, -0.072748676f, -0.079274714f, -0.085797355f, -0.092316322f,
-0.098831341f, -0.10534211f, -0.11184838f, -0.11834986f, -0.12484626f,
-0.13133731f, -0.13782275f, -0.14430228f, -0.15077563f, -0.15724251f,
-0.16370267f, -0.17015581f, -0.17660165f, -0.18303993f, -0.18947038f,
-0.19589271f, -0.20230664f, -0.20871192f, -0.21510825f, -0.22149536f,
-0.22787298f, -0.23424086f, -0.24059868f, -0.24694622f, -0.25328314f,
-0.25960925f, -0.26592422f, -0.27222782f, -0.27851975f, -0.28479972f,
-0.29106751f, -0.29732284f, -0.30356544f, -0.30979502f, -0.31601134f,
-0.32221413f, -0.32840309f, -0.33457801f, -0.34073856f, -0.34688455f,
-0.35301566f, -0.35913166f, -0.36523229f, -0.37131724f, -0.37738630f,
-0.38343921f, -0.38947567f, -0.39549544f, -0.40149832f, -0.40748394f,
-0.41345215f, -0.41940263f, -0.42533514f, -0.43124944f, -0.43714526f,
-0.44302234f, -0.44888046f, -0.45471936f, -0.46053877f, -0.46633846f,
-0.47211814f, -0.47787762f, -0.48361665f, -0.48933494f, -0.49503228f,
-0.50070840f, -0.50636309f, -0.51199609f, -0.51760709f, -0.52319598f,
-0.52876246f, -0.53430629f, -0.53982723f, -0.54532504f, -0.55079949f,
-0.55625033f, -0.56167740f, -0.56708032f, -0.57245898f, -0.57781315f,
-0.58314258f, -0.58844697f, -0.59372622f, -0.59897995f, -0.60420811f,
-0.60941035f, -0.61458647f, -0.61973625f, -0.62485951f, -0.62995601f,
-0.63502556f, -0.64006782f, -0.64508271f, -0.65007001f, -0.65502942f,
-0.65996075f, -0.66486382f, -0.66973841f, -0.67458433f, -0.67940134f,
-0.68418926f, -0.68894786f, -0.69367695f, -0.69837630f, -0.70304573f,
-0.70768511f, -0.71229410f, -0.71687263f, -0.72142041f, -0.72593731f,
-0.73042315f, -0.73487765f, -0.73930067f, -0.74369204f, -0.74805158f,
-0.75237900f, -0.75667429f, -0.76093709f, -0.76516730f, -0.76936477f,
-0.77352923f, -0.77766061f, -0.78175867f, -0.78582323f, -0.78985411f,
-0.79385114f, -0.79781419f, -0.80174309f, -0.80563760f, -0.80949765f,
-0.81332302f, -0.81711352f, -0.82086903f, -0.82458937f, -0.82827437f,
-0.83192390f, -0.83553779f, -0.83911592f, -0.84265804f, -0.84616417f,
-0.84963393f, -0.85306740f, -0.85646427f, -0.85982448f, -0.86314780f,
-0.86643422f, -0.86968350f, -0.87289548f, -0.87607014f, -0.87920725f,
-0.88230664f, -0.88536829f, -0.88839203f, -0.89137769f, -0.89432514f,
-0.89723432f, -0.90010506f, -0.90293723f, -0.90573072f, -0.90848541f,
-0.91120118f, -0.91387796f, -0.91651553f, -0.91911387f, -0.92167282f,
-0.92419231f, -0.92667222f, -0.92911243f, -0.93151283f, -0.93387336f,
-0.93619382f, -0.93847424f, -0.94071442f, -0.94291431f, -0.94507378f,
-0.94719279f, -0.94927126f, -0.95130903f, -0.95330608f, -0.95526224f,
-0.95717752f, -0.95905179f, -0.96088499f, -0.96267700f, -0.96442777f,
-0.96613729f, -0.96780539f, -0.96943200f, -0.97101706f, -0.97256058f,
-0.97406244f, -0.97552258f, -0.97694093f, -0.97831738f, -0.97965199f,
-0.98094457f, -0.98219514f, -0.98340368f, -0.98457009f, -0.98569429f,
-0.98677629f, -0.98781598f, -0.98881340f, -0.98976845f, -0.99068111f,
-0.99155134f, -0.99237907f, -0.99316430f, -0.99390697f, -0.99460709f,
-0.99526459f, -0.99587947f, -0.99645168f, -0.99698120f, -0.99746799f,
-0.99791211f, -0.99831343f, -0.99867201f, -0.99898779f, -0.99926084f,
-0.99949104f, -0.99967843f, -0.99982297f, -0.99992472f, -0.99998361f,
0.99999869f, 0.99989158f, 0.99961317f, 0.99916345f, 0.99854255f,
0.99775058f, 0.99678761f, 0.99565387f, 0.99434954f, 0.99287480f,
0.99122995f, 0.98941529f, 0.98743105f, 0.98527765f, 0.98295540f,
0.98046476f, 0.97780609f, 0.97497988f, 0.97198665f, 0.96882683f,
0.96550101f, 0.96200979f, 0.95835376f, 0.95453346f, 0.95054960f,
0.94640291f, 0.94209403f, 0.93762374f, 0.93299282f, 0.92820197f,
0.92325211f, 0.91814411f, 0.91287869f, 0.90745693f, 0.90187967f,
0.89614785f, 0.89026248f, 0.88422459f, 0.87803519f, 0.87169534f,
0.86520612f, 0.85856867f, 0.85178405f, 0.84485358f, 0.83777827f,
0.83055943f, 0.82319832f, 0.81569612f, 0.80805415f, 0.80027372f,
0.79235619f, 0.78430289f, 0.77611518f, 0.76779449f, 0.75934225f,
0.75075996f, 0.74204898f, 0.73321080f, 0.72424710f, 0.71515924f,
0.70594883f, 0.69661748f, 0.68716675f, 0.67759830f, 0.66791373f,
0.65811473f, 0.64820296f, 0.63818014f, 0.62804794f, 0.61780810f,
0.60746247f, 0.59701276f, 0.58646071f, 0.57580817f, 0.56505698f,
0.55420899f, 0.54326600f, 0.53222996f, 0.52110273f, 0.50988621f,
0.49858227f, 0.48719296f, 0.47572014f, 0.46416581f, 0.45253196f,
0.44082057f, 0.42903364f, 0.41717321f, 0.40524128f, 0.39323992f,
0.38117120f, 0.36903715f, 0.35683987f, 0.34458145f, 0.33226398f,
0.31988961f, 0.30746040f, 0.29497850f, 0.28244606f, 0.26986524f,
0.25723818f, 0.24456702f, 0.23185398f, 0.21910121f, 0.20631088f,
0.19348522f, 0.18062639f, 0.16773662f, 0.15481812f, 0.14187308f,
0.12890373f, 0.11591230f, 0.10290100f, 0.089872077f, 0.076827750f,
0.063770257f, 0.050701842f, 0.037624735f, 0.024541186f, 0.011453429f,
-0.0016362892f, -0.014725727f, -0.027812643f, -0.040894791f, -0.053969935f,
-0.067035832f, -0.080090240f, -0.093130924f, -0.10615565f, -0.11916219f,
-0.13214831f, -0.14511178f, -0.15805040f, -0.17096193f, -0.18384418f,
-0.19669491f, -0.20951195f, -0.22229309f, -0.23503613f, -0.24773891f,
-0.26039925f, -0.27301496f, -0.28558388f, -0.29810387f, -0.31057280f,
-0.32298848f, -0.33534884f, -0.34765175f, -0.35989508f, -0.37207675f,
-0.38419467f, -0.39624676f, -0.40823093f, -0.42014518f, -0.43198743f,
-0.44375566f, -0.45544785f, -0.46706200f, -0.47859612f, -0.49004826f,
-0.50141639f, -0.51269865f, -0.52389306f, -0.53499764f, -0.54601061f,
-0.55693001f, -0.56775403f, -0.57848072f, -0.58910829f, -0.59963489f,
-0.61005878f, -0.62037814f, -0.63059121f, -0.64069623f, -0.65069145f,
-0.66057515f, -0.67034572f, -0.68000144f, -0.68954057f, -0.69896162f,
-0.70826286f, -0.71744281f, -0.72649974f, -0.73543227f, -0.74423873f,
-0.75291771f, -0.76146764f, -0.76988715f, -0.77817470f, -0.78632891f,
-0.79434842f, -0.80223179f, -0.80997771f, -0.81758487f, -0.82505190f,
-0.83237761f, -0.83956063f, -0.84659988f, -0.85349399f, -0.86024189f,
-0.86684239f, -0.87329435f, -0.87959671f, -0.88574833f, -0.89174819f,
-0.89759529f, -0.90328854f, -0.90882701f, -0.91420978f, -0.91943592f,
-0.92450452f, -0.92941469f, -0.93416560f, -0.93875647f, -0.94318646f,
-0.94745487f, -0.95156091f, -0.95550388f, -0.95928317f, -0.96289814f,
-0.96634805f, -0.96963239f, -0.97275060f, -0.97570217f, -0.97848648f,
-0.98110318f, -0.98355180f, -0.98583186f, -0.98794299f, -0.98988485f,
-0.99165714f, -0.99325943f, -0.99469161f, -0.99595332f, -0.99704438f,
-0.99796462f, -0.99871385f, -0.99929196f, -0.99969882f, -0.99993443f,
0.99999464f, 0.99956632f, 0.99845290f, 0.99665523f, 0.99417448f,
0.99101239f, 0.98717111f, 0.98265326f, 0.97746199f, 0.97160077f,
0.96507365f, 0.95788515f, 0.95004016f, 0.94154406f, 0.93240267f,
0.92262226f, 0.91220951f, 0.90117162f, 0.88951606f, 0.87725091f,
0.86438453f, 0.85092574f, 0.83688372f, 0.82226819f, 0.80708915f,
0.79135692f, 0.77508235f, 0.75827658f, 0.74095112f, 0.72311783f,
0.70478898f, 0.68597710f, 0.66669506f, 0.64695615f, 0.62677377f,
0.60616189f, 0.58513457f, 0.56370622f, 0.54189157f, 0.51970547f,
0.49716324f, 0.47428027f, 0.45107225f, 0.42755505f, 0.40374488f,
0.37965798f, 0.35531086f, 0.33072025f, 0.30590299f, 0.28087607f,
0.25565663f, 0.23026201f, 0.20470956f, 0.17901683f, 0.15320139f,
0.12728097f, 0.10127331f, 0.075196236f, 0.049067631f, 0.022905400f,
-0.0032725304f, -0.029448219f, -0.055603724f, -0.081721120f, -0.10778251f,
-0.13377003f, -0.15966587f, -0.18545228f, -0.21111161f, -0.23662624f,
-0.26197869f, -0.28715160f, -0.31212771f, -0.33688989f, -0.36142120f,
-0.38570482f, -0.40972409f, -0.43346253f, -0.45690393f, -0.48003218f,
-0.50283146f, -0.52528608f, -0.54738069f, -0.56910020f, -0.59042966f,
-0.61135447f, -0.63186026f, -0.65193301f, -0.67155898f, -0.69072473f,
-0.70941705f, -0.72762316f, -0.74533063f, -0.76252723f, -0.77920127f,
-0.79534131f, -0.81093621f, -0.82597536f, -0.84044844f, -0.85434550f,
-0.86765707f, -0.88037395f, -0.89248747f, -0.90398932f, -0.91487163f,
-0.92512697f, -0.93474823f, -0.94372886f, -0.95206273f, -0.95974404f,
-0.96676767f, -0.97312868f, -0.97882277f, -0.98384601f, -0.98819500f,
-0.99186671f, -0.99485862f, -0.99716878f, -0.99879545f, -0.99973762f,
};
#endif
static const CELTMode mode48000_960_120 = {

404
Frameworks/Opus/Opus/opus/celt/static_modes_float_arm_ne10.h Normal file
View file

@ -0,0 +1,404 @@
/* The contents of this file was automatically generated by
* dump_mode_arm_ne10.c with arguments: 48000 960
* It contains static definitions for some pre-defined modes. */
#include <NE10_init.h>
#ifndef NE10_FFT_PARAMS48000_960
#define NE10_FFT_PARAMS48000_960
static const ne10_int32_t ne10_factors_480[64] = {
4, 40, 4, 30, 2, 15, 5, 3, 3, 1, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, };
static const ne10_int32_t ne10_factors_240[64] = {
3, 20, 4, 15, 5, 3, 3, 1, 1, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, };
static const ne10_int32_t ne10_factors_120[64] = {
3, 10, 2, 15, 5, 3, 3, 1, 1, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, };
static const ne10_int32_t ne10_factors_60[64] = {
2, 5, 5, 3, 3, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, };
static const ne10_fft_cpx_float32_t ne10_twiddles_480[480] = {
{1.0000000f,0.0000000f}, {1.0000000f,-0.0000000f}, {1.0000000f,-0.0000000f},
{1.0000000f,-0.0000000f}, {0.91354543f,-0.40673664f}, {0.66913056f,-0.74314487f},
{1.0000000f,-0.0000000f}, {0.66913056f,-0.74314487f}, {-0.10452851f,-0.99452192f},
{1.0000000f,-0.0000000f}, {0.30901697f,-0.95105654f}, {-0.80901700f,-0.58778518f},
{1.0000000f,-0.0000000f}, {-0.10452851f,-0.99452192f}, {-0.97814757f,0.20791179f},
{1.0000000f,-0.0000000f}, {0.97814763f,-0.20791170f}, {0.91354543f,-0.40673664f},
{0.80901700f,-0.58778524f}, {0.66913056f,-0.74314487f}, {0.49999997f,-0.86602545f},
{0.30901697f,-0.95105654f}, {0.10452842f,-0.99452192f}, {-0.10452851f,-0.99452192f},
{-0.30901703f,-0.95105648f}, {-0.50000006f,-0.86602533f}, {-0.66913068f,-0.74314475f},
{-0.80901700f,-0.58778518f}, {-0.91354549f,-0.40673658f}, {-0.97814763f,-0.20791161f},
{1.0000000f,-0.0000000f}, {0.99862951f,-0.052335959f}, {0.99452192f,-0.10452846f},
{0.98768836f,-0.15643448f}, {0.97814763f,-0.20791170f}, {0.96592581f,-0.25881904f},
{0.95105648f,-0.30901700f}, {0.93358040f,-0.35836795f}, {0.91354543f,-0.40673664f},
{0.89100653f,-0.45399052f}, {0.86602545f,-0.50000000f}, {0.83867055f,-0.54463905f},
{0.80901700f,-0.58778524f}, {0.77714598f,-0.62932038f}, {0.74314475f,-0.66913062f},
{0.70710677f,-0.70710683f}, {0.66913056f,-0.74314487f}, {0.62932038f,-0.77714598f},
{0.58778524f,-0.80901700f}, {0.54463899f,-0.83867055f}, {0.49999997f,-0.86602545f},
{0.45399052f,-0.89100653f}, {0.40673661f,-0.91354549f}, {0.35836786f,-0.93358046f},
{0.30901697f,-0.95105654f}, {0.25881907f,-0.96592581f}, {0.20791166f,-0.97814763f},
{0.15643437f,-0.98768836f}, {0.10452842f,-0.99452192f}, {0.052335974f,-0.99862951f},
{1.0000000f,-0.0000000f}, {0.99452192f,-0.10452846f}, {0.97814763f,-0.20791170f},
{0.95105648f,-0.30901700f}, {0.91354543f,-0.40673664f}, {0.86602545f,-0.50000000f},
{0.80901700f,-0.58778524f}, {0.74314475f,-0.66913062f}, {0.66913056f,-0.74314487f},
{0.58778524f,-0.80901700f}, {0.49999997f,-0.86602545f}, {0.40673661f,-0.91354549f},
{0.30901697f,-0.95105654f}, {0.20791166f,-0.97814763f}, {0.10452842f,-0.99452192f},
{-4.3711388e-08f,-1.0000000f}, {-0.10452851f,-0.99452192f}, {-0.20791174f,-0.97814757f},
{-0.30901703f,-0.95105648f}, {-0.40673670f,-0.91354543f}, {-0.50000006f,-0.86602533f},
{-0.58778518f,-0.80901700f}, {-0.66913068f,-0.74314475f}, {-0.74314493f,-0.66913044f},
{-0.80901700f,-0.58778518f}, {-0.86602539f,-0.50000006f}, {-0.91354549f,-0.40673658f},
{-0.95105654f,-0.30901679f}, {-0.97814763f,-0.20791161f}, {-0.99452192f,-0.10452849f},
{1.0000000f,-0.0000000f}, {0.98768836f,-0.15643448f}, {0.95105648f,-0.30901700f},
{0.89100653f,-0.45399052f}, {0.80901700f,-0.58778524f}, {0.70710677f,-0.70710683f},
{0.58778524f,-0.80901700f}, {0.45399052f,-0.89100653f}, {0.30901697f,-0.95105654f},
{0.15643437f,-0.98768836f}, {-4.3711388e-08f,-1.0000000f}, {-0.15643445f,-0.98768836f},
{-0.30901703f,-0.95105648f}, {-0.45399061f,-0.89100647f}, {-0.58778518f,-0.80901700f},
{-0.70710677f,-0.70710677f}, {-0.80901700f,-0.58778518f}, {-0.89100659f,-0.45399037f},
{-0.95105654f,-0.30901679f}, {-0.98768836f,-0.15643445f}, {-1.0000000f,8.7422777e-08f},
{-0.98768830f,0.15643461f}, {-0.95105654f,0.30901697f}, {-0.89100653f,0.45399055f},
{-0.80901694f,0.58778536f}, {-0.70710665f,0.70710689f}, {-0.58778507f,0.80901712f},
{-0.45399022f,0.89100665f}, {-0.30901709f,0.95105648f}, {-0.15643452f,0.98768830f},
{1.0000000f,-0.0000000f}, {0.99991435f,-0.013089596f}, {0.99965733f,-0.026176950f},
{0.99922901f,-0.039259817f}, {0.99862951f,-0.052335959f}, {0.99785894f,-0.065403134f},
{0.99691731f,-0.078459099f}, {0.99580491f,-0.091501623f}, {0.99452192f,-0.10452846f},
{0.99306846f,-0.11753740f}, {0.99144489f,-0.13052620f}, {0.98965138f,-0.14349262f},
{0.98768836f,-0.15643448f}, {0.98555607f,-0.16934951f}, {0.98325491f,-0.18223552f},
{0.98078525f,-0.19509032f}, {0.97814763f,-0.20791170f}, {0.97534233f,-0.22069745f},
{0.97236991f,-0.23344538f}, {0.96923089f,-0.24615330f}, {0.96592581f,-0.25881904f},
{0.96245521f,-0.27144045f}, {0.95881975f,-0.28401536f}, {0.95501995f,-0.29654160f},
{0.95105648f,-0.30901700f}, {0.94693011f,-0.32143945f}, {0.94264150f,-0.33380687f},
{0.93819129f,-0.34611708f}, {0.93358040f,-0.35836795f}, {0.92880952f,-0.37055743f},
{0.92387956f,-0.38268346f}, {0.91879117f,-0.39474389f}, {0.91354543f,-0.40673664f},
{0.90814316f,-0.41865975f}, {0.90258527f,-0.43051112f}, {0.89687270f,-0.44228873f},
{0.89100653f,-0.45399052f}, {0.88498765f,-0.46561453f}, {0.87881708f,-0.47715878f},
{0.87249601f,-0.48862126f}, {0.86602545f,-0.50000000f}, {0.85940641f,-0.51129311f},
{0.85264015f,-0.52249855f}, {0.84572786f,-0.53361452f}, {0.83867055f,-0.54463905f},
{0.83146960f,-0.55557024f}, {0.82412618f,-0.56640625f}, {0.81664151f,-0.57714522f},
{0.80901700f,-0.58778524f}, {0.80125380f,-0.59832460f}, {0.79335332f,-0.60876143f},
{0.78531694f,-0.61909395f}, {0.77714598f,-0.62932038f}, {0.76884180f,-0.63943899f},
{0.76040596f,-0.64944810f}, {0.75183982f,-0.65934587f}, {0.74314475f,-0.66913062f},
{0.73432249f,-0.67880076f}, {0.72537434f,-0.68835455f}, {0.71630192f,-0.69779050f},
{0.70710677f,-0.70710683f}, {0.69779044f,-0.71630198f}, {0.68835455f,-0.72537440f},
{0.67880070f,-0.73432255f}, {0.66913056f,-0.74314487f}, {0.65934581f,-0.75183982f},
{0.64944804f,-0.76040596f}, {0.63943899f,-0.76884186f}, {0.62932038f,-0.77714598f},
{0.61909395f,-0.78531694f}, {0.60876137f,-0.79335338f}, {0.59832460f,-0.80125386f},
{0.58778524f,-0.80901700f}, {0.57714516f,-0.81664151f}, {0.56640625f,-0.82412618f},
{0.55557019f,-0.83146960f}, {0.54463899f,-0.83867055f}, {0.53361452f,-0.84572786f},
{0.52249849f,-0.85264015f}, {0.51129311f,-0.85940641f}, {0.49999997f,-0.86602545f},
{0.48862118f,-0.87249601f}, {0.47715876f,-0.87881708f}, {0.46561447f,-0.88498765f},
{0.45399052f,-0.89100653f}, {0.44228867f,-0.89687276f}, {0.43051103f,-0.90258533f},
{0.41865975f,-0.90814316f}, {0.40673661f,-0.91354549f}, {0.39474380f,-0.91879129f},
{0.38268343f,-0.92387956f}, {0.37055740f,-0.92880958f}, {0.35836786f,-0.93358046f},
{0.34611705f,-0.93819135f}, {0.33380681f,-0.94264150f}, {0.32143947f,-0.94693011f},
{0.30901697f,-0.95105654f}, {0.29654151f,-0.95501995f}, {0.28401533f,-0.95881975f},
{0.27144039f,-0.96245527f}, {0.25881907f,-0.96592581f}, {0.24615327f,-0.96923089f},
{0.23344530f,-0.97236991f}, {0.22069745f,-0.97534233f}, {0.20791166f,-0.97814763f},
{0.19509023f,-0.98078531f}, {0.18223552f,-0.98325491f}, {0.16934945f,-0.98555607f},
{0.15643437f,-0.98768836f}, {0.14349259f,-0.98965138f}, {0.13052613f,-0.99144489f},
{0.11753740f,-0.99306846f}, {0.10452842f,-0.99452192f}, {0.091501534f,-0.99580491f},
{0.078459084f,-0.99691731f}, {0.065403074f,-0.99785894f}, {0.052335974f,-0.99862951f},
{0.039259788f,-0.99922901f}, {0.026176875f,-0.99965733f}, {0.013089597f,-0.99991435f},
{1.0000000f,-0.0000000f}, {0.99965733f,-0.026176950f}, {0.99862951f,-0.052335959f},
{0.99691731f,-0.078459099f}, {0.99452192f,-0.10452846f}, {0.99144489f,-0.13052620f},
{0.98768836f,-0.15643448f}, {0.98325491f,-0.18223552f}, {0.97814763f,-0.20791170f},
{0.97236991f,-0.23344538f}, {0.96592581f,-0.25881904f}, {0.95881975f,-0.28401536f},
{0.95105648f,-0.30901700f}, {0.94264150f,-0.33380687f}, {0.93358040f,-0.35836795f},
{0.92387956f,-0.38268346f}, {0.91354543f,-0.40673664f}, {0.90258527f,-0.43051112f},
{0.89100653f,-0.45399052f}, {0.87881708f,-0.47715878f}, {0.86602545f,-0.50000000f},
{0.85264015f,-0.52249855f}, {0.83867055f,-0.54463905f}, {0.82412618f,-0.56640625f},
{0.80901700f,-0.58778524f}, {0.79335332f,-0.60876143f}, {0.77714598f,-0.62932038f},
{0.76040596f,-0.64944810f}, {0.74314475f,-0.66913062f}, {0.72537434f,-0.68835455f},
{0.70710677f,-0.70710683f}, {0.68835455f,-0.72537440f}, {0.66913056f,-0.74314487f},
{0.64944804f,-0.76040596f}, {0.62932038f,-0.77714598f}, {0.60876137f,-0.79335338f},
{0.58778524f,-0.80901700f}, {0.56640625f,-0.82412618f}, {0.54463899f,-0.83867055f},
{0.52249849f,-0.85264015f}, {0.49999997f,-0.86602545f}, {0.47715876f,-0.87881708f},
{0.45399052f,-0.89100653f}, {0.43051103f,-0.90258533f}, {0.40673661f,-0.91354549f},
{0.38268343f,-0.92387956f}, {0.35836786f,-0.93358046f}, {0.33380681f,-0.94264150f},
{0.30901697f,-0.95105654f}, {0.28401533f,-0.95881975f}, {0.25881907f,-0.96592581f},
{0.23344530f,-0.97236991f}, {0.20791166f,-0.97814763f}, {0.18223552f,-0.98325491f},
{0.15643437f,-0.98768836f}, {0.13052613f,-0.99144489f}, {0.10452842f,-0.99452192f},
{0.078459084f,-0.99691731f}, {0.052335974f,-0.99862951f}, {0.026176875f,-0.99965733f},
{-4.3711388e-08f,-1.0000000f}, {-0.026176963f,-0.99965733f}, {-0.052336060f,-0.99862951f},
{-0.078459173f,-0.99691731f}, {-0.10452851f,-0.99452192f}, {-0.13052621f,-0.99144489f},
{-0.15643445f,-0.98768836f}, {-0.18223560f,-0.98325491f}, {-0.20791174f,-0.97814757f},
{-0.23344538f,-0.97236991f}, {-0.25881916f,-0.96592581f}, {-0.28401542f,-0.95881969f},
{-0.30901703f,-0.95105648f}, {-0.33380687f,-0.94264150f}, {-0.35836795f,-0.93358040f},
{-0.38268352f,-0.92387950f}, {-0.40673670f,-0.91354543f}, {-0.43051112f,-0.90258527f},
{-0.45399061f,-0.89100647f}, {-0.47715873f,-0.87881708f}, {-0.50000006f,-0.86602533f},
{-0.52249867f,-0.85264009f}, {-0.54463905f,-0.83867055f}, {-0.56640631f,-0.82412612f},
{-0.58778518f,-0.80901700f}, {-0.60876143f,-0.79335332f}, {-0.62932050f,-0.77714586f},
{-0.64944804f,-0.76040596f}, {-0.66913068f,-0.74314475f}, {-0.68835467f,-0.72537428f},
{-0.70710677f,-0.70710677f}, {-0.72537446f,-0.68835449f}, {-0.74314493f,-0.66913044f},
{-0.76040596f,-0.64944804f}, {-0.77714604f,-0.62932026f}, {-0.79335332f,-0.60876143f},
{-0.80901700f,-0.58778518f}, {-0.82412624f,-0.56640613f}, {-0.83867055f,-0.54463899f},
{-0.85264021f,-0.52249849f}, {-0.86602539f,-0.50000006f}, {-0.87881714f,-0.47715873f},
{-0.89100659f,-0.45399037f}, {-0.90258527f,-0.43051112f}, {-0.91354549f,-0.40673658f},
{-0.92387956f,-0.38268328f}, {-0.93358040f,-0.35836792f}, {-0.94264150f,-0.33380675f},
{-0.95105654f,-0.30901679f}, {-0.95881975f,-0.28401530f}, {-0.96592587f,-0.25881892f},
{-0.97236991f,-0.23344538f}, {-0.97814763f,-0.20791161f}, {-0.98325491f,-0.18223536f},
{-0.98768836f,-0.15643445f}, {-0.99144489f,-0.13052608f}, {-0.99452192f,-0.10452849f},
{-0.99691737f,-0.078459039f}, {-0.99862957f,-0.052335810f}, {-0.99965733f,-0.026176952f},
{1.0000000f,-0.0000000f}, {0.99922901f,-0.039259817f}, {0.99691731f,-0.078459099f},
{0.99306846f,-0.11753740f}, {0.98768836f,-0.15643448f}, {0.98078525f,-0.19509032f},
{0.97236991f,-0.23344538f}, {0.96245521f,-0.27144045f}, {0.95105648f,-0.30901700f},
{0.93819129f,-0.34611708f}, {0.92387956f,-0.38268346f}, {0.90814316f,-0.41865975f},
{0.89100653f,-0.45399052f}, {0.87249601f,-0.48862126f}, {0.85264015f,-0.52249855f},
{0.83146960f,-0.55557024f}, {0.80901700f,-0.58778524f}, {0.78531694f,-0.61909395f},
{0.76040596f,-0.64944810f}, {0.73432249f,-0.67880076f}, {0.70710677f,-0.70710683f},
{0.67880070f,-0.73432255f}, {0.64944804f,-0.76040596f}, {0.61909395f,-0.78531694f},
{0.58778524f,-0.80901700f}, {0.55557019f,-0.83146960f}, {0.52249849f,-0.85264015f},
{0.48862118f,-0.87249601f}, {0.45399052f,-0.89100653f}, {0.41865975f,-0.90814316f},
{0.38268343f,-0.92387956f}, {0.34611705f,-0.93819135f}, {0.30901697f,-0.95105654f},
{0.27144039f,-0.96245527f}, {0.23344530f,-0.97236991f}, {0.19509023f,-0.98078531f},
{0.15643437f,-0.98768836f}, {0.11753740f,-0.99306846f}, {0.078459084f,-0.99691731f},
{0.039259788f,-0.99922901f}, {-4.3711388e-08f,-1.0000000f}, {-0.039259877f,-0.99922901f},
{-0.078459173f,-0.99691731f}, {-0.11753749f,-0.99306846f}, {-0.15643445f,-0.98768836f},
{-0.19509032f,-0.98078525f}, {-0.23344538f,-0.97236991f}, {-0.27144048f,-0.96245521f},
{-0.30901703f,-0.95105648f}, {-0.34611711f,-0.93819129f}, {-0.38268352f,-0.92387950f},
{-0.41865984f,-0.90814310f}, {-0.45399061f,-0.89100647f}, {-0.48862135f,-0.87249595f},
{-0.52249867f,-0.85264009f}, {-0.55557036f,-0.83146954f}, {-0.58778518f,-0.80901700f},
{-0.61909389f,-0.78531694f}, {-0.64944804f,-0.76040596f}, {-0.67880076f,-0.73432249f},
{-0.70710677f,-0.70710677f}, {-0.73432249f,-0.67880070f}, {-0.76040596f,-0.64944804f},
{-0.78531694f,-0.61909389f}, {-0.80901700f,-0.58778518f}, {-0.83146966f,-0.55557019f},
{-0.85264021f,-0.52249849f}, {-0.87249607f,-0.48862115f}, {-0.89100659f,-0.45399037f},
{-0.90814322f,-0.41865960f}, {-0.92387956f,-0.38268328f}, {-0.93819135f,-0.34611690f},
{-0.95105654f,-0.30901679f}, {-0.96245521f,-0.27144048f}, {-0.97236991f,-0.23344538f},
{-0.98078531f,-0.19509031f}, {-0.98768836f,-0.15643445f}, {-0.99306846f,-0.11753736f},
{-0.99691737f,-0.078459039f}, {-0.99922901f,-0.039259743f}, {-1.0000000f,8.7422777e-08f},
{-0.99922901f,0.039259918f}, {-0.99691731f,0.078459218f}, {-0.99306846f,0.11753753f},
{-0.98768830f,0.15643461f}, {-0.98078525f,0.19509049f}, {-0.97236985f,0.23344554f},
{-0.96245515f,0.27144065f}, {-0.95105654f,0.30901697f}, {-0.93819135f,0.34611705f},
{-0.92387956f,0.38268346f}, {-0.90814316f,0.41865975f}, {-0.89100653f,0.45399055f},
{-0.87249601f,0.48862129f}, {-0.85264015f,0.52249861f}, {-0.83146960f,0.55557030f},
{-0.80901694f,0.58778536f}, {-0.78531688f,0.61909401f}, {-0.76040590f,0.64944816f},
{-0.73432243f,0.67880082f}, {-0.70710665f,0.70710689f}, {-0.67880058f,0.73432261f},
{-0.64944792f,0.76040608f}, {-0.61909378f,0.78531706f}, {-0.58778507f,0.80901712f},
{-0.55557001f,0.83146977f}, {-0.52249837f,0.85264033f}, {-0.48862100f,0.87249613f},
{-0.45399022f,0.89100665f}, {-0.41865945f,0.90814328f}, {-0.38268313f,0.92387968f},
{-0.34611672f,0.93819147f}, {-0.30901709f,0.95105648f}, {-0.27144054f,0.96245521f},
{-0.23344545f,0.97236991f}, {-0.19509038f,0.98078525f}, {-0.15643452f,0.98768830f},
{-0.11753743f,0.99306846f}, {-0.078459114f,0.99691731f}, {-0.039259821f,0.99922901f},
};
static const ne10_fft_cpx_float32_t ne10_twiddles_240[240] = {
{1.0000000f,0.0000000f}, {1.0000000f,-0.0000000f}, {1.0000000f,-0.0000000f},
{1.0000000f,-0.0000000f}, {0.91354543f,-0.40673664f}, {0.66913056f,-0.74314487f},
{1.0000000f,-0.0000000f}, {0.66913056f,-0.74314487f}, {-0.10452851f,-0.99452192f},
{1.0000000f,-0.0000000f}, {0.30901697f,-0.95105654f}, {-0.80901700f,-0.58778518f},
{1.0000000f,-0.0000000f}, {-0.10452851f,-0.99452192f}, {-0.97814757f,0.20791179f},
{1.0000000f,-0.0000000f}, {0.99452192f,-0.10452846f}, {0.97814763f,-0.20791170f},
{0.95105648f,-0.30901700f}, {0.91354543f,-0.40673664f}, {0.86602545f,-0.50000000f},
{0.80901700f,-0.58778524f}, {0.74314475f,-0.66913062f}, {0.66913056f,-0.74314487f},
{0.58778524f,-0.80901700f}, {0.49999997f,-0.86602545f}, {0.40673661f,-0.91354549f},
{0.30901697f,-0.95105654f}, {0.20791166f,-0.97814763f}, {0.10452842f,-0.99452192f},
{1.0000000f,-0.0000000f}, {0.97814763f,-0.20791170f}, {0.91354543f,-0.40673664f},
{0.80901700f,-0.58778524f}, {0.66913056f,-0.74314487f}, {0.49999997f,-0.86602545f},
{0.30901697f,-0.95105654f}, {0.10452842f,-0.99452192f}, {-0.10452851f,-0.99452192f},
{-0.30901703f,-0.95105648f}, {-0.50000006f,-0.86602533f}, {-0.66913068f,-0.74314475f},
{-0.80901700f,-0.58778518f}, {-0.91354549f,-0.40673658f}, {-0.97814763f,-0.20791161f},
{1.0000000f,-0.0000000f}, {0.95105648f,-0.30901700f}, {0.80901700f,-0.58778524f},
{0.58778524f,-0.80901700f}, {0.30901697f,-0.95105654f}, {-4.3711388e-08f,-1.0000000f},
{-0.30901703f,-0.95105648f}, {-0.58778518f,-0.80901700f}, {-0.80901700f,-0.58778518f},
{-0.95105654f,-0.30901679f}, {-1.0000000f,8.7422777e-08f}, {-0.95105654f,0.30901697f},
{-0.80901694f,0.58778536f}, {-0.58778507f,0.80901712f}, {-0.30901709f,0.95105648f},
{1.0000000f,-0.0000000f}, {0.99965733f,-0.026176950f}, {0.99862951f,-0.052335959f},
{0.99691731f,-0.078459099f}, {0.99452192f,-0.10452846f}, {0.99144489f,-0.13052620f},
{0.98768836f,-0.15643448f}, {0.98325491f,-0.18223552f}, {0.97814763f,-0.20791170f},
{0.97236991f,-0.23344538f}, {0.96592581f,-0.25881904f}, {0.95881975f,-0.28401536f},
{0.95105648f,-0.30901700f}, {0.94264150f,-0.33380687f}, {0.93358040f,-0.35836795f},
{0.92387956f,-0.38268346f}, {0.91354543f,-0.40673664f}, {0.90258527f,-0.43051112f},
{0.89100653f,-0.45399052f}, {0.87881708f,-0.47715878f}, {0.86602545f,-0.50000000f},
{0.85264015f,-0.52249855f}, {0.83867055f,-0.54463905f}, {0.82412618f,-0.56640625f},
{0.80901700f,-0.58778524f}, {0.79335332f,-0.60876143f}, {0.77714598f,-0.62932038f},
{0.76040596f,-0.64944810f}, {0.74314475f,-0.66913062f}, {0.72537434f,-0.68835455f},
{0.70710677f,-0.70710683f}, {0.68835455f,-0.72537440f}, {0.66913056f,-0.74314487f},
{0.64944804f,-0.76040596f}, {0.62932038f,-0.77714598f}, {0.60876137f,-0.79335338f},
{0.58778524f,-0.80901700f}, {0.56640625f,-0.82412618f}, {0.54463899f,-0.83867055f},
{0.52249849f,-0.85264015f}, {0.49999997f,-0.86602545f}, {0.47715876f,-0.87881708f},
{0.45399052f,-0.89100653f}, {0.43051103f,-0.90258533f}, {0.40673661f,-0.91354549f},
{0.38268343f,-0.92387956f}, {0.35836786f,-0.93358046f}, {0.33380681f,-0.94264150f},
{0.30901697f,-0.95105654f}, {0.28401533f,-0.95881975f}, {0.25881907f,-0.96592581f},
{0.23344530f,-0.97236991f}, {0.20791166f,-0.97814763f}, {0.18223552f,-0.98325491f},
{0.15643437f,-0.98768836f}, {0.13052613f,-0.99144489f}, {0.10452842f,-0.99452192f},
{0.078459084f,-0.99691731f}, {0.052335974f,-0.99862951f}, {0.026176875f,-0.99965733f},
{1.0000000f,-0.0000000f}, {0.99862951f,-0.052335959f}, {0.99452192f,-0.10452846f},
{0.98768836f,-0.15643448f}, {0.97814763f,-0.20791170f}, {0.96592581f,-0.25881904f},
{0.95105648f,-0.30901700f}, {0.93358040f,-0.35836795f}, {0.91354543f,-0.40673664f},
{0.89100653f,-0.45399052f}, {0.86602545f,-0.50000000f}, {0.83867055f,-0.54463905f},
{0.80901700f,-0.58778524f}, {0.77714598f,-0.62932038f}, {0.74314475f,-0.66913062f},
{0.70710677f,-0.70710683f}, {0.66913056f,-0.74314487f}, {0.62932038f,-0.77714598f},
{0.58778524f,-0.80901700f}, {0.54463899f,-0.83867055f}, {0.49999997f,-0.86602545f},
{0.45399052f,-0.89100653f}, {0.40673661f,-0.91354549f}, {0.35836786f,-0.93358046f},
{0.30901697f,-0.95105654f}, {0.25881907f,-0.96592581f}, {0.20791166f,-0.97814763f},
{0.15643437f,-0.98768836f}, {0.10452842f,-0.99452192f}, {0.052335974f,-0.99862951f},
{-4.3711388e-08f,-1.0000000f}, {-0.052336060f,-0.99862951f}, {-0.10452851f,-0.99452192f},
{-0.15643445f,-0.98768836f}, {-0.20791174f,-0.97814757f}, {-0.25881916f,-0.96592581f},
{-0.30901703f,-0.95105648f}, {-0.35836795f,-0.93358040f}, {-0.40673670f,-0.91354543f},
{-0.45399061f,-0.89100647f}, {-0.50000006f,-0.86602533f}, {-0.54463905f,-0.83867055f},
{-0.58778518f,-0.80901700f}, {-0.62932050f,-0.77714586f}, {-0.66913068f,-0.74314475f},
{-0.70710677f,-0.70710677f}, {-0.74314493f,-0.66913044f}, {-0.77714604f,-0.62932026f},
{-0.80901700f,-0.58778518f}, {-0.83867055f,-0.54463899f}, {-0.86602539f,-0.50000006f},
{-0.89100659f,-0.45399037f}, {-0.91354549f,-0.40673658f}, {-0.93358040f,-0.35836792f},
{-0.95105654f,-0.30901679f}, {-0.96592587f,-0.25881892f}, {-0.97814763f,-0.20791161f},
{-0.98768836f,-0.15643445f}, {-0.99452192f,-0.10452849f}, {-0.99862957f,-0.052335810f},
{1.0000000f,-0.0000000f}, {0.99691731f,-0.078459099f}, {0.98768836f,-0.15643448f},
{0.97236991f,-0.23344538f}, {0.95105648f,-0.30901700f}, {0.92387956f,-0.38268346f},
{0.89100653f,-0.45399052f}, {0.85264015f,-0.52249855f}, {0.80901700f,-0.58778524f},
{0.76040596f,-0.64944810f}, {0.70710677f,-0.70710683f}, {0.64944804f,-0.76040596f},
{0.58778524f,-0.80901700f}, {0.52249849f,-0.85264015f}, {0.45399052f,-0.89100653f},
{0.38268343f,-0.92387956f}, {0.30901697f,-0.95105654f}, {0.23344530f,-0.97236991f},
{0.15643437f,-0.98768836f}, {0.078459084f,-0.99691731f}, {-4.3711388e-08f,-1.0000000f},
{-0.078459173f,-0.99691731f}, {-0.15643445f,-0.98768836f}, {-0.23344538f,-0.97236991f},
{-0.30901703f,-0.95105648f}, {-0.38268352f,-0.92387950f}, {-0.45399061f,-0.89100647f},
{-0.52249867f,-0.85264009f}, {-0.58778518f,-0.80901700f}, {-0.64944804f,-0.76040596f},
{-0.70710677f,-0.70710677f}, {-0.76040596f,-0.64944804f}, {-0.80901700f,-0.58778518f},
{-0.85264021f,-0.52249849f}, {-0.89100659f,-0.45399037f}, {-0.92387956f,-0.38268328f},
{-0.95105654f,-0.30901679f}, {-0.97236991f,-0.23344538f}, {-0.98768836f,-0.15643445f},
{-0.99691737f,-0.078459039f}, {-1.0000000f,8.7422777e-08f}, {-0.99691731f,0.078459218f},
{-0.98768830f,0.15643461f}, {-0.97236985f,0.23344554f}, {-0.95105654f,0.30901697f},
{-0.92387956f,0.38268346f}, {-0.89100653f,0.45399055f}, {-0.85264015f,0.52249861f},
{-0.80901694f,0.58778536f}, {-0.76040590f,0.64944816f}, {-0.70710665f,0.70710689f},
{-0.64944792f,0.76040608f}, {-0.58778507f,0.80901712f}, {-0.52249837f,0.85264033f},
{-0.45399022f,0.89100665f}, {-0.38268313f,0.92387968f}, {-0.30901709f,0.95105648f},
{-0.23344545f,0.97236991f}, {-0.15643452f,0.98768830f}, {-0.078459114f,0.99691731f},
};
static const ne10_fft_cpx_float32_t ne10_twiddles_120[120] = {
{1.0000000f,0.0000000f}, {1.0000000f,-0.0000000f}, {1.0000000f,-0.0000000f},
{1.0000000f,-0.0000000f}, {0.91354543f,-0.40673664f}, {0.66913056f,-0.74314487f},
{1.0000000f,-0.0000000f}, {0.66913056f,-0.74314487f}, {-0.10452851f,-0.99452192f},
{1.0000000f,-0.0000000f}, {0.30901697f,-0.95105654f}, {-0.80901700f,-0.58778518f},
{1.0000000f,-0.0000000f}, {-0.10452851f,-0.99452192f}, {-0.97814757f,0.20791179f},
{1.0000000f,-0.0000000f}, {0.97814763f,-0.20791170f}, {0.91354543f,-0.40673664f},
{0.80901700f,-0.58778524f}, {0.66913056f,-0.74314487f}, {0.49999997f,-0.86602545f},
{0.30901697f,-0.95105654f}, {0.10452842f,-0.99452192f}, {-0.10452851f,-0.99452192f},
{-0.30901703f,-0.95105648f}, {-0.50000006f,-0.86602533f}, {-0.66913068f,-0.74314475f},
{-0.80901700f,-0.58778518f}, {-0.91354549f,-0.40673658f}, {-0.97814763f,-0.20791161f},
{1.0000000f,-0.0000000f}, {0.99862951f,-0.052335959f}, {0.99452192f,-0.10452846f},
{0.98768836f,-0.15643448f}, {0.97814763f,-0.20791170f}, {0.96592581f,-0.25881904f},
{0.95105648f,-0.30901700f}, {0.93358040f,-0.35836795f}, {0.91354543f,-0.40673664f},
{0.89100653f,-0.45399052f}, {0.86602545f,-0.50000000f}, {0.83867055f,-0.54463905f},
{0.80901700f,-0.58778524f}, {0.77714598f,-0.62932038f}, {0.74314475f,-0.66913062f},
{0.70710677f,-0.70710683f}, {0.66913056f,-0.74314487f}, {0.62932038f,-0.77714598f},
{0.58778524f,-0.80901700f}, {0.54463899f,-0.83867055f}, {0.49999997f,-0.86602545f},
{0.45399052f,-0.89100653f}, {0.40673661f,-0.91354549f}, {0.35836786f,-0.93358046f},
{0.30901697f,-0.95105654f}, {0.25881907f,-0.96592581f}, {0.20791166f,-0.97814763f},
{0.15643437f,-0.98768836f}, {0.10452842f,-0.99452192f}, {0.052335974f,-0.99862951f},
{1.0000000f,-0.0000000f}, {0.99452192f,-0.10452846f}, {0.97814763f,-0.20791170f},
{0.95105648f,-0.30901700f}, {0.91354543f,-0.40673664f}, {0.86602545f,-0.50000000f},
{0.80901700f,-0.58778524f}, {0.74314475f,-0.66913062f}, {0.66913056f,-0.74314487f},
{0.58778524f,-0.80901700f}, {0.49999997f,-0.86602545f}, {0.40673661f,-0.91354549f},
{0.30901697f,-0.95105654f}, {0.20791166f,-0.97814763f}, {0.10452842f,-0.99452192f},
{-4.3711388e-08f,-1.0000000f}, {-0.10452851f,-0.99452192f}, {-0.20791174f,-0.97814757f},
{-0.30901703f,-0.95105648f}, {-0.40673670f,-0.91354543f}, {-0.50000006f,-0.86602533f},
{-0.58778518f,-0.80901700f}, {-0.66913068f,-0.74314475f}, {-0.74314493f,-0.66913044f},
{-0.80901700f,-0.58778518f}, {-0.86602539f,-0.50000006f}, {-0.91354549f,-0.40673658f},
{-0.95105654f,-0.30901679f}, {-0.97814763f,-0.20791161f}, {-0.99452192f,-0.10452849f},
{1.0000000f,-0.0000000f}, {0.98768836f,-0.15643448f}, {0.95105648f,-0.30901700f},
{0.89100653f,-0.45399052f}, {0.80901700f,-0.58778524f}, {0.70710677f,-0.70710683f},
{0.58778524f,-0.80901700f}, {0.45399052f,-0.89100653f}, {0.30901697f,-0.95105654f},
{0.15643437f,-0.98768836f}, {-4.3711388e-08f,-1.0000000f}, {-0.15643445f,-0.98768836f},
{-0.30901703f,-0.95105648f}, {-0.45399061f,-0.89100647f}, {-0.58778518f,-0.80901700f},
{-0.70710677f,-0.70710677f}, {-0.80901700f,-0.58778518f}, {-0.89100659f,-0.45399037f},
{-0.95105654f,-0.30901679f}, {-0.98768836f,-0.15643445f}, {-1.0000000f,8.7422777e-08f},
{-0.98768830f,0.15643461f}, {-0.95105654f,0.30901697f}, {-0.89100653f,0.45399055f},
{-0.80901694f,0.58778536f}, {-0.70710665f,0.70710689f}, {-0.58778507f,0.80901712f},
{-0.45399022f,0.89100665f}, {-0.30901709f,0.95105648f}, {-0.15643452f,0.98768830f},
};
static const ne10_fft_cpx_float32_t ne10_twiddles_60[60] = {
{1.0000000f,0.0000000f}, {1.0000000f,-0.0000000f}, {1.0000000f,-0.0000000f},
{1.0000000f,-0.0000000f}, {0.91354543f,-0.40673664f}, {0.66913056f,-0.74314487f},
{1.0000000f,-0.0000000f}, {0.66913056f,-0.74314487f}, {-0.10452851f,-0.99452192f},
{1.0000000f,-0.0000000f}, {0.30901697f,-0.95105654f}, {-0.80901700f,-0.58778518f},
{1.0000000f,-0.0000000f}, {-0.10452851f,-0.99452192f}, {-0.97814757f,0.20791179f},
{1.0000000f,-0.0000000f}, {0.99452192f,-0.10452846f}, {0.97814763f,-0.20791170f},
{0.95105648f,-0.30901700f}, {0.91354543f,-0.40673664f}, {0.86602545f,-0.50000000f},
{0.80901700f,-0.58778524f}, {0.74314475f,-0.66913062f}, {0.66913056f,-0.74314487f},
{0.58778524f,-0.80901700f}, {0.49999997f,-0.86602545f}, {0.40673661f,-0.91354549f},
{0.30901697f,-0.95105654f}, {0.20791166f,-0.97814763f}, {0.10452842f,-0.99452192f},
{1.0000000f,-0.0000000f}, {0.97814763f,-0.20791170f}, {0.91354543f,-0.40673664f},
{0.80901700f,-0.58778524f}, {0.66913056f,-0.74314487f}, {0.49999997f,-0.86602545f},
{0.30901697f,-0.95105654f}, {0.10452842f,-0.99452192f}, {-0.10452851f,-0.99452192f},
{-0.30901703f,-0.95105648f}, {-0.50000006f,-0.86602533f}, {-0.66913068f,-0.74314475f},
{-0.80901700f,-0.58778518f}, {-0.91354549f,-0.40673658f}, {-0.97814763f,-0.20791161f},
{1.0000000f,-0.0000000f}, {0.95105648f,-0.30901700f}, {0.80901700f,-0.58778524f},
{0.58778524f,-0.80901700f}, {0.30901697f,-0.95105654f}, {-4.3711388e-08f,-1.0000000f},
{-0.30901703f,-0.95105648f}, {-0.58778518f,-0.80901700f}, {-0.80901700f,-0.58778518f},
{-0.95105654f,-0.30901679f}, {-1.0000000f,8.7422777e-08f}, {-0.95105654f,0.30901697f},
{-0.80901694f,0.58778536f}, {-0.58778507f,0.80901712f}, {-0.30901709f,0.95105648f},
};
static const ne10_fft_state_float32_t ne10_fft_state_float32_t_480 = {
120,
(ne10_int32_t *)ne10_factors_480,
(ne10_fft_cpx_float32_t *)ne10_twiddles_480,
NULL,
(ne10_fft_cpx_float32_t *)&ne10_twiddles_480[120],
/* is_forward_scaled = true */
(ne10_int32_t) 1,
/* is_backward_scaled = false */
(ne10_int32_t) 0,
};
static const arch_fft_state cfg_arch_480 = {
1,
(void *)&ne10_fft_state_float32_t_480,
};
static const ne10_fft_state_float32_t ne10_fft_state_float32_t_240 = {
60,
(ne10_int32_t *)ne10_factors_240,
(ne10_fft_cpx_float32_t *)ne10_twiddles_240,
NULL,
(ne10_fft_cpx_float32_t *)&ne10_twiddles_240[60],
/* is_forward_scaled = true */
(ne10_int32_t) 1,
/* is_backward_scaled = false */
(ne10_int32_t) 0,
};
static const arch_fft_state cfg_arch_240 = {
1,
(void *)&ne10_fft_state_float32_t_240,
};
static const ne10_fft_state_float32_t ne10_fft_state_float32_t_120 = {
30,
(ne10_int32_t *)ne10_factors_120,
(ne10_fft_cpx_float32_t *)ne10_twiddles_120,
NULL,
(ne10_fft_cpx_float32_t *)&ne10_twiddles_120[30],
/* is_forward_scaled = true */
(ne10_int32_t) 1,
/* is_backward_scaled = false */
(ne10_int32_t) 0,
};
static const arch_fft_state cfg_arch_120 = {
1,
(void *)&ne10_fft_state_float32_t_120,
};
static const ne10_fft_state_float32_t ne10_fft_state_float32_t_60 = {
15,
(ne10_int32_t *)ne10_factors_60,
(ne10_fft_cpx_float32_t *)ne10_twiddles_60,
NULL,
(ne10_fft_cpx_float32_t *)&ne10_twiddles_60[15],
/* is_forward_scaled = true */
(ne10_int32_t) 1,
/* is_backward_scaled = false */
(ne10_int32_t) 0,
};
static const arch_fft_state cfg_arch_60 = {
1,
(void *)&ne10_fft_state_float32_t_60,
};
#endif /* end NE10_FFT_PARAMS48000_960 */

104
Frameworks/Opus/Opus/opus/celt/tests/test_unit_cwrs32.c
View file

@ -53,22 +53,20 @@
#ifdef TEST_CUSTOM_MODES
#define NDIMS (46)
#define NDIMS (44)
static const int pn[NDIMS]={
2, 3, 4, 5, 6, 7, 8, 9, 10,
11, 12, 13, 14, 15, 16, 18, 20, 22,
24, 26, 28, 30, 32, 36, 40, 44, 48,
52, 56, 60, 64, 72, 80, 88, 96, 104,
112, 120, 128, 144, 160, 176, 192, 208, 224,
240
112, 120, 128, 144, 160, 176, 192, 208
};
static const int pkmax[NDIMS]={
128, 128, 128, 128, 88, 52, 36, 26, 22,
18, 16, 15, 13, 12, 12, 11, 10, 9,
9, 8, 8, 7, 7, 7, 7, 6, 6,
6, 6, 6, 5, 5, 5, 5, 5, 5,
4, 4, 4, 4, 4, 4, 4, 4, 4,
4
4, 4, 4, 4, 4, 4, 4, 4
};
#else /* TEST_CUSTOM_MODES */
@ -97,29 +95,39 @@ int main(void){
for(pseudo=1;pseudo<41;pseudo++)
{
int k;
#if defined(SMALL_FOOTPRINT)
opus_uint32 uu[KMAX+2U];
#endif
opus_uint32 inc;
opus_uint32 nc;
opus_uint32 i;
k=get_pulses(pseudo);
if (k>pkmax[t])break;
printf("Testing CWRS with N=%i, K=%i...\n",n,k);
#if defined(SMALL_FOOTPRINT)
nc=ncwrs_urow(n,k,uu);
#else
nc=CELT_PVQ_V(n,k);
#endif
inc=nc/20000;
if(inc<1)inc=1;
for(i=0;i<nc;i+=inc){
#if defined(SMALL_FOOTPRINT)
opus_uint32 u[KMAX+2U];
#endif
int y[NMAX];
int sy;
int yy[5];
opus_uint32 v;
opus_uint32 ii;
int kk;
int j;
#if defined(SMALL_FOOTPRINT)
memcpy(u,uu,(k+2U)*sizeof(*u));
cwrsi(n,k,i,y,u);
#else
cwrsi(n,k,i,y);
#endif
sy=0;
for(j=0;j<n;j++)sy+=ABS(y[j]);
for(j=0;j<n;j++)sy+=abs(y[j]);
if(sy!=k){
fprintf(stderr,"N=%d Pulse count mismatch in cwrsi (%d!=%d).\n",
n,sy,k);
@ -128,7 +136,12 @@ int main(void){
/*printf("%6u of %u:",i,nc);
for(j=0;j<n;j++)printf(" %+3i",y[j]);
printf(" ->");*/
#if defined(SMALL_FOOTPRINT)
ii=icwrs(n,k,&v,y,u);
#else
ii=icwrs(n,y);
v=CELT_PVQ_V(n,k);
#endif
if(ii!=i){
fprintf(stderr,"Combination-index mismatch (%lu!=%lu).\n",
(long)ii,(long)i);
@ -139,81 +152,6 @@ int main(void){
(long)v,(long)nc);
return 2;
}
#ifndef SMALL_FOOTPRINT
if(n==2){
cwrsi2(k,i,yy);
for(j=0;j<2;j++)if(yy[j]!=y[j]){
fprintf(stderr,"N=2 pulse vector mismatch ({%i,%i}!={%i,%i}).\n",
yy[0],yy[1],y[0],y[1]);
return 3;
}
ii=icwrs2(yy,&kk);
if(ii!=i){
fprintf(stderr,"N=2 combination-index mismatch (%lu!=%lu).\n",
(long)ii,(long)i);
return 4;
}
if(kk!=k){
fprintf(stderr,"N=2 pulse count mismatch (%i,%i).\n",kk,k);
return 5;
}
v=ncwrs2(k);
if(v!=nc){
fprintf(stderr,"N=2 combination count mismatch (%lu,%lu).\n",
(long)v,(long)nc);
return 6;
}
}
else if(n==3){
cwrsi3(k,i,yy);
for(j=0;j<3;j++)if(yy[j]!=y[j]){
fprintf(stderr,"N=3 pulse vector mismatch "
"({%i,%i,%i}!={%i,%i,%i}).\n",yy[0],yy[1],yy[2],y[0],y[1],y[2]);
return 7;
}
ii=icwrs3(yy,&kk);
if(ii!=i){
fprintf(stderr,"N=3 combination-index mismatch (%lu!=%lu).\n",
(long)ii,(long)i);
return 8;
}
if(kk!=k){
fprintf(stderr,"N=3 pulse count mismatch (%i!=%i).\n",kk,k);
return 9;
}
v=ncwrs3(k);
if(v!=nc){
fprintf(stderr,"N=3 combination count mismatch (%lu!=%lu).\n",
(long)v,(long)nc);
return 10;
}
}
else if(n==4){
cwrsi4(k,i,yy);
for(j=0;j<4;j++)if(yy[j]!=y[j]){
fprintf(stderr,"N=4 pulse vector mismatch "
"({%i,%i,%i,%i}!={%i,%i,%i,%i}.\n",
yy[0],yy[1],yy[2],yy[3],y[0],y[1],y[2],y[3]);
return 11;
}
ii=icwrs4(yy,&kk);
if(ii!=i){
fprintf(stderr,"N=4 combination-index mismatch (%lu!=%lu).\n",
(long)ii,(long)i);
return 12;
}
if(kk!=k){
fprintf(stderr,"N=4 pulse count mismatch (%i!=%i).\n",kk,k);
return 13;
}
v=ncwrs4(k);
if(v!=nc){
fprintf(stderr,"N=4 combination count mismatch (%lu!=%lu).\n",
(long)v,(long)nc);
return 14;
}
}
#endif /* SMALL_FOOTPRINT */
/*printf(" %6u\n",i);*/
}
/*printf("\n");*/

61
Frameworks/Opus/Opus/opus/celt/tests/test_unit_dft.c
View file

@ -38,12 +38,29 @@
#include <stdio.h>
#define CELT_C
#define TEST_UNIT_DFT_C
#include "stack_alloc.h"
#include "kiss_fft.h"
#include "kiss_fft.c"
#include "mathops.c"
#include "entcode.c"
#if defined(OPUS_X86_MAY_HAVE_SSE2) || defined(OPUS_X86_MAY_HAVE_SSE4_1)
# include "x86/x86cpu.c"
#elif defined(OPUS_ARM_ASM) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
# include "arm/armcpu.c"
# include "celt_lpc.c"
# include "pitch.c"
# if defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
# include "arm/celt_neon_intr.c"
# if defined(HAVE_ARM_NE10)
# include "mdct.c"
# include "arm/celt_ne10_fft.c"
# include "arm/celt_ne10_mdct.c"
# endif
# endif
# include "arm/arm_celt_map.c"
#endif
#ifndef M_PI
#define M_PI 3.141592653
@ -92,13 +109,13 @@ void check(kiss_fft_cpx * in,kiss_fft_cpx * out,int nfft,int isinverse)
}
}
void test1d(int nfft,int isinverse)
void test1d(int nfft,int isinverse,int arch)
{
size_t buflen = sizeof(kiss_fft_cpx)*nfft;
kiss_fft_cpx * in = (kiss_fft_cpx*)malloc(buflen);
kiss_fft_cpx * out= (kiss_fft_cpx*)malloc(buflen);
kiss_fft_state *cfg = opus_fft_alloc(nfft,0,0);
kiss_fft_state *cfg = opus_fft_alloc(nfft,0,0,arch);
int k;
for (k=0;k<nfft;++k) {
@ -122,9 +139,9 @@ void test1d(int nfft,int isinverse)
/*for (k=0;k<nfft;++k) printf("%d %d ", in[k].r, in[k].i);printf("\n");*/
if (isinverse)
opus_ifft(cfg,in,out);
opus_ifft(cfg,in,out, arch);
else
opus_fft(cfg,in,out);
opus_fft(cfg,in,out, arch);
/*for (k=0;k<nfft;++k) printf("%d %d ", out[k].r, out[k].i);printf("\n");*/
@ -138,26 +155,34 @@ void test1d(int nfft,int isinverse)
int main(int argc,char ** argv)
{
ALLOC_STACK;
int arch = opus_select_arch();
if (argc>1) {
int k;
for (k=1;k<argc;++k) {
test1d(atoi(argv[k]),0);
test1d(atoi(argv[k]),1);
test1d(atoi(argv[k]),0,arch);
test1d(atoi(argv[k]),1,arch);
}
}else{
test1d(32,0);
test1d(32,1);
test1d(128,0);
test1d(128,1);
test1d(256,0);
test1d(256,1);
test1d(32,0,arch);
test1d(32,1,arch);
test1d(128,0,arch);
test1d(128,1,arch);
test1d(256,0,arch);
test1d(256,1,arch);
#ifndef RADIX_TWO_ONLY
test1d(36,0);
test1d(36,1);
test1d(50,0);
test1d(50,1);
test1d(120,0);
test1d(120,1);
test1d(36,0,arch);
test1d(36,1,arch);
test1d(50,0,arch);
test1d(50,1,arch);
test1d(60,0,arch);
test1d(60,1,arch);
test1d(120,0,arch);
test1d(120,1,arch);
test1d(240,0,arch);
test1d(240,1,arch);
test1d(480,0,arch);
test1d(480,1,arch);
#endif
}
return ret;

35
Frameworks/Opus/Opus/opus/celt/tests/test_unit_mathops.c
View file

@ -36,15 +36,46 @@
#define CELT_C
#include <stdio.h>
#include <math.h>
#include "mathops.c"
#include "entenc.c"
#include "entdec.c"
#include "entcode.c"
#include "bands.c"
#include "quant_bands.c"
#include "laplace.c"
#include "vq.c"
#include "cwrs.c"
#include <stdio.h>
#include <math.h>
#include "pitch.c"
#include "celt_lpc.c"
#include "celt.c"
#if defined(OPUS_X86_MAY_HAVE_SSE) || defined(OPUS_X86_MAY_HAVE_SSE2) || defined(OPUS_X86_MAY_HAVE_SSE4_1)
# if defined(OPUS_X86_MAY_HAVE_SSE)
# include "x86/pitch_sse.c"
# endif
# if defined(OPUS_X86_MAY_HAVE_SSE2)
# include "x86/pitch_sse2.c"
# endif
# if defined(OPUS_X86_MAY_HAVE_SSE4_1)
# include "x86/pitch_sse4_1.c"
# include "x86/celt_lpc_sse.c"
# endif
# include "x86/x86_celt_map.c"
#elif defined(OPUS_ARM_ASM) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
# include "arm/armcpu.c"
# if defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
# include "arm/celt_neon_intr.c"
# if defined(HAVE_ARM_NE10)
# include "kiss_fft.c"
# include "mdct.c"
# include "arm/celt_ne10_fft.c"
# include "arm/celt_ne10_mdct.c"
# endif
# endif
# include "arm/arm_celt_map.c"
#endif
#ifdef FIXED_POINT
#define WORD "%d"

75
Frameworks/Opus/Opus/opus/celt/tests/test_unit_mdct.c
View file

@ -46,6 +46,22 @@
#include "mathops.c"
#include "entcode.c"
#if defined(OPUS_X86_MAY_HAVE_SSE2) || defined(OPUS_X86_MAY_HAVE_SSE4_1)
# include "x86/x86cpu.c"
#elif defined(OPUS_ARM_ASM) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
# include "arm/armcpu.c"
# include "pitch.c"
# include "celt_lpc.c"
# if defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
# include "arm/celt_neon_intr.c"
# if defined(HAVE_ARM_NE10)
# include "arm/celt_ne10_fft.c"
# include "arm/celt_ne10_mdct.c"
# endif
# endif
# include "arm/arm_celt_map.c"
#endif
#ifndef M_PI
#define M_PI 3.141592653
#endif
@ -112,7 +128,7 @@ void check_inv(kiss_fft_scalar * in,kiss_fft_scalar * out,int nfft,int isinver
}
void test1d(int nfft,int isinverse)
void test1d(int nfft,int isinverse,int arch)
{
mdct_lookup cfg;
size_t buflen = sizeof(kiss_fft_scalar)*nfft;
@ -123,7 +139,7 @@ void test1d(int nfft,int isinverse)
opus_val16 * window= (opus_val16*)malloc(sizeof(opus_val16)*nfft/2);
int k;
clt_mdct_init(&cfg, nfft, 0);
clt_mdct_init(&cfg, nfft, 0, arch);
for (k=0;k<nfft;++k) {
in[k] = (rand() % 32768) - 16384;
}
@ -150,10 +166,13 @@ void test1d(int nfft,int isinverse)
{
for (k=0;k<nfft;++k)
out[k] = 0;
clt_mdct_backward(&cfg,in,out, window, nfft/2, 0, 1);
clt_mdct_backward(&cfg,in,out, window, nfft/2, 0, 1, arch);
/* apply TDAC because clt_mdct_backward() no longer does that */
for (k=0;k<nfft/4;++k)
out[nfft-k-1] = out[nfft/2+k];
check_inv(in,out,nfft,isinverse);
} else {
clt_mdct_forward(&cfg,in,out,window, nfft/2, 0, 1);
clt_mdct_forward(&cfg,in,out,window, nfft/2, 0, 1, arch);
check(in_copy,out,nfft,isinverse);
}
/*for (k=0;k<nfft;++k) printf("%d %d ", out[k].r, out[k].i);printf("\n");*/
@ -161,34 +180,48 @@ void test1d(int nfft,int isinverse)
free(in);
free(out);
clt_mdct_clear(&cfg);
clt_mdct_clear(&cfg, arch);
}
int main(int argc,char ** argv)
{
ALLOC_STACK;
int arch = opus_select_arch();
if (argc>1) {
int k;
for (k=1;k<argc;++k) {
test1d(atoi(argv[k]),0);
test1d(atoi(argv[k]),1);
test1d(atoi(argv[k]),0,arch);
test1d(atoi(argv[k]),1,arch);
}
}else{
test1d(32,0);
test1d(32,1);
test1d(256,0);
test1d(256,1);
test1d(512,0);
test1d(512,1);
test1d(32,0,arch);
test1d(32,1,arch);
test1d(256,0,arch);
test1d(256,1,arch);
test1d(512,0,arch);
test1d(512,1,arch);
test1d(1024,0,arch);
test1d(1024,1,arch);
test1d(2048,0,arch);
test1d(2048,1,arch);
#ifndef RADIX_TWO_ONLY
test1d(40,0);
test1d(40,1);
test1d(120,0);
test1d(120,1);
test1d(240,0);
test1d(240,1);
test1d(480,0);
test1d(480,1);
test1d(36,0,arch);
test1d(36,1,arch);
test1d(40,0,arch);
test1d(40,1,arch);
test1d(60,0,arch);
test1d(60,1,arch);
test1d(120,0,arch);
test1d(120,1,arch);
test1d(240,0,arch);
test1d(240,1,arch);
test1d(480,0,arch);
test1d(480,1,arch);
test1d(960,0,arch);
test1d(960,1,arch);
test1d(1920,0,arch);
test1d(1920,1,arch);
#endif
}
return ret;

30
Frameworks/Opus/Opus/opus/celt/tests/test_unit_rotation.c
View file

@ -44,7 +44,37 @@
#include "entdec.c"
#include "mathops.c"
#include "bands.h"
#include "pitch.c"
#include "celt_lpc.c"
#include "celt.c"
#include <math.h>
#if defined(OPUS_X86_MAY_HAVE_SSE) || defined(OPUS_X86_MAY_HAVE_SSE2) || defined(OPUS_X86_MAY_HAVE_SSE4_1)
# if defined(OPUS_X86_MAY_HAVE_SSE)
# include "x86/pitch_sse.c"
# endif
# if defined(OPUS_X86_MAY_HAVE_SSE2)
# include "x86/pitch_sse2.c"
# endif
# if defined(OPUS_X86_MAY_HAVE_SSE4_1)
# include "x86/pitch_sse4_1.c"
# include "x86/celt_lpc_sse.c"
# endif
# include "x86/x86_celt_map.c"
#elif defined(OPUS_ARM_ASM) || defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
# include "arm/armcpu.c"
# if defined(OPUS_ARM_MAY_HAVE_NEON_INTR)
# include "arm/celt_neon_intr.c"
# if defined(HAVE_ARM_NE10)
# include "kiss_fft.c"
# include "mdct.c"
# include "arm/celt_ne10_fft.c"
# include "arm/celt_ne10_mdct.c"
# endif
# endif
# include "arm/arm_celt_map.c"
#endif
#define MAX_SIZE 100
int ret=0;

55
Frameworks/Opus/Opus/opus/celt/vq.c
View file

@ -37,19 +37,23 @@
#include "os_support.h"
#include "bands.h"
#include "rate.h"
#include "pitch.h"
#ifndef OVERRIDE_vq_exp_rotation1
static void exp_rotation1(celt_norm *X, int len, int stride, opus_val16 c, opus_val16 s)
{
int i;
opus_val16 ms;
celt_norm *Xptr;
Xptr = X;
ms = NEG16(s);
for (i=0;i<len-stride;i++)
{
celt_norm x1, x2;
x1 = Xptr[0];
x2 = Xptr[stride];
Xptr[stride] = EXTRACT16(SHR32(MULT16_16(c,x2) + MULT16_16(s,x1), 15));
*Xptr++ = EXTRACT16(SHR32(MULT16_16(c,x1) - MULT16_16(s,x2), 15));
Xptr[stride] = EXTRACT16(PSHR32(MAC16_16(MULT16_16(c, x2), s, x1), 15));
*Xptr++ = EXTRACT16(PSHR32(MAC16_16(MULT16_16(c, x1), ms, x2), 15));
}
Xptr = &X[len-2*stride-1];
for (i=len-2*stride-1;i>=0;i--)
@ -57,10 +61,11 @@ static void exp_rotation1(celt_norm *X, int len, int stride, opus_val16 c, opus_
celt_norm x1, x2;
x1 = Xptr[0];
x2 = Xptr[stride];
Xptr[stride] = EXTRACT16(SHR32(MULT16_16(c,x2) + MULT16_16(s,x1), 15));
*Xptr-- = EXTRACT16(SHR32(MULT16_16(c,x1) - MULT16_16(s,x2), 15));
Xptr[stride] = EXTRACT16(PSHR32(MAC16_16(MULT16_16(c, x2), s, x1), 15));
*Xptr-- = EXTRACT16(PSHR32(MAC16_16(MULT16_16(c, x1), ms, x2), 15));
}
}
#endif /* OVERRIDE_vq_exp_rotation1 */
static void exp_rotation(celt_norm *X, int len, int dir, int stride, int K, int spread)
{
@ -91,7 +96,7 @@ static void exp_rotation(celt_norm *X, int len, int dir, int stride, int K, int
}
/*NOTE: As a minor optimization, we could be passing around log2(B), not B, for both this and for
extract_collapse_mask().*/
len /= stride;
len = celt_udiv(len, stride);
for (i=0;i<stride;i++)
{
if (dir < 0)
@ -140,13 +145,15 @@ static unsigned extract_collapse_mask(int *iy, int N, int B)
return 1;
/*NOTE: As a minor optimization, we could be passing around log2(B), not B, for both this and for
exp_rotation().*/
N0 = N/B;
N0 = celt_udiv(N, B);
collapse_mask = 0;
i=0; do {
int j;
unsigned tmp=0;
j=0; do {
collapse_mask |= (iy[i*N0+j]!=0)<<i;
tmp |= iy[i*N0+j];
} while (++j<N0);
collapse_mask |= (tmp!=0)<<i;
} while (++i<B);
return collapse_mask;
}
@ -322,7 +329,6 @@ unsigned alg_quant(celt_norm *X, int N, int K, int spread, int B, ec_enc *enc
unsigned alg_unquant(celt_norm *X, int N, int K, int spread, int B,
ec_dec *dec, opus_val16 gain)
{
int i;
opus_val32 Ryy;
unsigned collapse_mask;
VARDECL(int, iy);
@ -331,12 +337,7 @@ unsigned alg_unquant(celt_norm *X, int N, int K, int spread, int B,
celt_assert2(K>0, "alg_unquant() needs at least one pulse");
celt_assert2(N>1, "alg_unquant() needs at least two dimensions");
ALLOC(iy, N, int);
decode_pulses(iy, N, K, dec);
Ryy = 0;
i=0;
do {
Ryy = MAC16_16(Ryy, iy[i], iy[i]);
} while (++i < N);
Ryy = decode_pulses(iy, N, K, dec);
normalise_residual(iy, X, N, Ryy, gain);
exp_rotation(X, N, -1, B, K, spread);
collapse_mask = extract_collapse_mask(iy, N, B);
@ -344,21 +345,18 @@ unsigned alg_unquant(celt_norm *X, int N, int K, int spread, int B,
return collapse_mask;
}
void renormalise_vector(celt_norm *X, int N, opus_val16 gain)
#ifndef OVERRIDE_renormalise_vector
void renormalise_vector(celt_norm *X, int N, opus_val16 gain, int arch)
{
int i;
#ifdef FIXED_POINT
int k;
#endif
opus_val32 E = EPSILON;
opus_val32 E;
opus_val16 g;
opus_val32 t;
celt_norm *xptr = X;
for (i=0;i<N;i++)
{
E = MAC16_16(E, *xptr, *xptr);
xptr++;
}
celt_norm *xptr;
E = EPSILON + celt_inner_prod(X, X, N, arch);
#ifdef FIXED_POINT
k = celt_ilog2(E)>>1;
#endif
@ -373,8 +371,9 @@ void renormalise_vector(celt_norm *X, int N, opus_val16 gain)
}
/*return celt_sqrt(E);*/
}
#endif /* OVERRIDE_renormalise_vector */
int stereo_itheta(celt_norm *X, celt_norm *Y, int stereo, int N)
int stereo_itheta(const celt_norm *X, const celt_norm *Y, int stereo, int N, int arch)
{
int i;
int itheta;
@ -393,14 +392,8 @@ int stereo_itheta(celt_norm *X, celt_norm *Y, int stereo, int N)
Eside = MAC16_16(Eside, s, s);
}
} else {
for (i=0;i<N;i++)
{
celt_norm m, s;
m = X[i];
s = Y[i];
Emid = MAC16_16(Emid, m, m);
Eside = MAC16_16(Eside, s, s);
}
Emid += celt_inner_prod(X, X, N, arch);
Eside += celt_inner_prod(Y, Y, N, arch);
}
mid = celt_sqrt(Emid);
side = celt_sqrt(Eside);

16
Frameworks/Opus/Opus/opus/celt/vq.h
View file

@ -37,14 +37,17 @@
#include "entdec.h"
#include "modes.h"
#if defined(MIPSr1_ASM)
#include "mips/vq_mipsr1.h"
#endif
/** Algebraic pulse-vector quantiser. The signal x is replaced by the sum of
* the pitch and a combination of pulses such that its norm is still equal
* to 1. This is the function that will typically require the most CPU.
* @param x Residual signal to quantise/encode (returns quantised version)
* @param W Perceptual weight to use when optimising (currently unused)
* @param X Residual signal to quantise/encode (returns quantised version)
* @param N Number of samples to encode
* @param K Number of pulses to use
* @param p Pitch vector (it is assumed that p+x is a unit vector)
* @param enc Entropy encoder state
* @ret A mask indicating which blocks in the band received pulses
*/
@ -56,18 +59,17 @@ unsigned alg_quant(celt_norm *X, int N, int K, int spread, int B,
);
/** Algebraic pulse decoder
* @param x Decoded normalised spectrum (returned)
* @param X Decoded normalised spectrum (returned)
* @param N Number of samples to decode
* @param K Number of pulses to use
* @param p Pitch vector (automatically added to x)
* @param dec Entropy decoder state
* @ret A mask indicating which blocks in the band received pulses
*/
unsigned alg_unquant(celt_norm *X, int N, int K, int spread, int B,
ec_dec *dec, opus_val16 gain);
void renormalise_vector(celt_norm *X, int N, opus_val16 gain);
void renormalise_vector(celt_norm *X, int N, opus_val16 gain, int arch);
int stereo_itheta(celt_norm *X, celt_norm *Y, int stereo, int N);
int stereo_itheta(const celt_norm *X, const celt_norm *Y, int stereo, int N, int arch);
#endif /* VQ_H */

132
Frameworks/Opus/Opus/opus/celt/x86/celt_lpc_sse.c Normal file
View file

@ -0,0 +1,132 @@
/* Copyright (c) 2014, Cisco Systems, INC
Written by XiangMingZhu WeiZhou MinPeng YanWang
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#if defined(FIXED_POINT)
#include <xmmintrin.h>
#include <emmintrin.h>
#include <smmintrin.h>
#include "celt_lpc.h"
#include "stack_alloc.h"
#include "mathops.h"
#include "pitch.h"
#include "x86cpu.h"
void celt_fir_sse4_1(const opus_val16 *_x,
const opus_val16 *num,
opus_val16 *_y,
int N,
int ord,
opus_val16 *mem,
int arch)
{
int i,j;
VARDECL(opus_val16, rnum);
VARDECL(opus_val16, x);
__m128i vecNoA;
opus_int32 noA ;
SAVE_STACK;
ALLOC(rnum, ord, opus_val16);
ALLOC(x, N+ord, opus_val16);
for(i=0;i<ord;i++)
rnum[i] = num[ord-i-1];
for(i=0;i<ord;i++)
x[i] = mem[ord-i-1];
for (i=0;i<N-7;i+=8)
{
x[i+ord ]=_x[i ];
x[i+ord+1]=_x[i+1];
x[i+ord+2]=_x[i+2];
x[i+ord+3]=_x[i+3];
x[i+ord+4]=_x[i+4];
x[i+ord+5]=_x[i+5];
x[i+ord+6]=_x[i+6];
x[i+ord+7]=_x[i+7];
}
for (;i<N-3;i+=4)
{
x[i+ord ]=_x[i ];
x[i+ord+1]=_x[i+1];
x[i+ord+2]=_x[i+2];
x[i+ord+3]=_x[i+3];
}
for (;i<N;i++)
x[i+ord]=_x[i];
for(i=0;i<ord;i++)
mem[i] = _x[N-i-1];
#ifdef SMALL_FOOTPRINT
for (i=0;i<N;i++)
{
opus_val32 sum = SHL32(EXTEND32(_x[i]), SIG_SHIFT);
for (j=0;j<ord;j++)
{
sum = MAC16_16(sum,rnum[j],x[i+j]);
}
_y[i] = SATURATE16(PSHR32(sum, SIG_SHIFT));
}
#else
noA = EXTEND32(1) << SIG_SHIFT >> 1;
vecNoA = _mm_set_epi32(noA, noA, noA, noA);
for (i=0;i<N-3;i+=4)
{
opus_val32 sums[4] = {0};
__m128i vecSum, vecX;
xcorr_kernel(rnum, x+i, sums, ord, arch);
vecSum = _mm_loadu_si128((__m128i *)sums);
vecSum = _mm_add_epi32(vecSum, vecNoA);
vecSum = _mm_srai_epi32(vecSum, SIG_SHIFT);
vecX = OP_CVTEPI16_EPI32_M64(_x + i);
vecSum = _mm_add_epi32(vecSum, vecX);
vecSum = _mm_packs_epi32(vecSum, vecSum);
_mm_storel_epi64((__m128i *)(_y + i), vecSum);
}
for (;i<N;i++)
{
opus_val32 sum = 0;
for (j=0;j<ord;j++)
sum = MAC16_16(sum, rnum[j], x[i + j]);
_y[i] = SATURATE16(ADD32(EXTEND32(_x[i]), PSHR32(sum, SIG_SHIFT)));
}
#endif
RESTORE_STACK;
}
#endif

68
Frameworks/Opus/Opus/opus/celt/x86/celt_lpc_sse.h Normal file
View file

@ -0,0 +1,68 @@
/* Copyright (c) 2014, Cisco Systems, INC
Written by XiangMingZhu WeiZhou MinPeng YanWang
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef CELT_LPC_SSE_H
#define CELT_LPC_SSE_H
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#if defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)
#define OVERRIDE_CELT_FIR
void celt_fir_sse4_1(
const opus_val16 *x,
const opus_val16 *num,
opus_val16 *y,
int N,
int ord,
opus_val16 *mem,
int arch);
#if defined(OPUS_X86_PRESUME_SSE4_1)
#define celt_fir(x, num, y, N, ord, mem, arch) \
((void)arch, celt_fir_sse4_1(x, num, y, N, ord, mem, arch))
#else
extern void (*const CELT_FIR_IMPL[OPUS_ARCHMASK + 1])(
const opus_val16 *x,
const opus_val16 *num,
opus_val16 *y,
int N,
int ord,
opus_val16 *mem,
int arch);
# define celt_fir(x, num, y, N, ord, mem, arch) \
((*CELT_FIR_IMPL[(arch) & OPUS_ARCHMASK])(x, num, y, N, ord, mem, arch))
#endif
#endif
#endif

185
Frameworks/Opus/Opus/opus/celt/x86/pitch_sse.c Normal file
View file

@ -0,0 +1,185 @@
/* Copyright (c) 2014, Cisco Systems, INC
Written by XiangMingZhu WeiZhou MinPeng YanWang
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "macros.h"
#include "celt_lpc.h"
#include "stack_alloc.h"
#include "mathops.h"
#include "pitch.h"
#if defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT)
#include <xmmintrin.h>
#include "arch.h"
void xcorr_kernel_sse(const opus_val16 *x, const opus_val16 *y, opus_val32 sum[4], int len)
{
int j;
__m128 xsum1, xsum2;
xsum1 = _mm_loadu_ps(sum);
xsum2 = _mm_setzero_ps();
for (j = 0; j < len-3; j += 4)
{
__m128 x0 = _mm_loadu_ps(x+j);
__m128 yj = _mm_loadu_ps(y+j);
__m128 y3 = _mm_loadu_ps(y+j+3);
xsum1 = _mm_add_ps(xsum1,_mm_mul_ps(_mm_shuffle_ps(x0,x0,0x00),yj));
xsum2 = _mm_add_ps(xsum2,_mm_mul_ps(_mm_shuffle_ps(x0,x0,0x55),
_mm_shuffle_ps(yj,y3,0x49)));
xsum1 = _mm_add_ps(xsum1,_mm_mul_ps(_mm_shuffle_ps(x0,x0,0xaa),
_mm_shuffle_ps(yj,y3,0x9e)));
xsum2 = _mm_add_ps(xsum2,_mm_mul_ps(_mm_shuffle_ps(x0,x0,0xff),y3));
}
if (j < len)
{
xsum1 = _mm_add_ps(xsum1,_mm_mul_ps(_mm_load1_ps(x+j),_mm_loadu_ps(y+j)));
if (++j < len)
{
xsum2 = _mm_add_ps(xsum2,_mm_mul_ps(_mm_load1_ps(x+j),_mm_loadu_ps(y+j)));
if (++j < len)
{
xsum1 = _mm_add_ps(xsum1,_mm_mul_ps(_mm_load1_ps(x+j),_mm_loadu_ps(y+j)));
}
}
}
_mm_storeu_ps(sum,_mm_add_ps(xsum1,xsum2));
}
void dual_inner_prod_sse(const opus_val16 *x, const opus_val16 *y01, const opus_val16 *y02,
int N, opus_val32 *xy1, opus_val32 *xy2)
{
int i;
__m128 xsum1, xsum2;
xsum1 = _mm_setzero_ps();
xsum2 = _mm_setzero_ps();
for (i=0;i<N-3;i+=4)
{
__m128 xi = _mm_loadu_ps(x+i);
__m128 y1i = _mm_loadu_ps(y01+i);
__m128 y2i = _mm_loadu_ps(y02+i);
xsum1 = _mm_add_ps(xsum1,_mm_mul_ps(xi, y1i));
xsum2 = _mm_add_ps(xsum2,_mm_mul_ps(xi, y2i));
}
/* Horizontal sum */
xsum1 = _mm_add_ps(xsum1, _mm_movehl_ps(xsum1, xsum1));
xsum1 = _mm_add_ss(xsum1, _mm_shuffle_ps(xsum1, xsum1, 0x55));
_mm_store_ss(xy1, xsum1);
xsum2 = _mm_add_ps(xsum2, _mm_movehl_ps(xsum2, xsum2));
xsum2 = _mm_add_ss(xsum2, _mm_shuffle_ps(xsum2, xsum2, 0x55));
_mm_store_ss(xy2, xsum2);
for (;i<N;i++)
{
*xy1 = MAC16_16(*xy1, x[i], y01[i]);
*xy2 = MAC16_16(*xy2, x[i], y02[i]);
}
}
opus_val32 celt_inner_prod_sse(const opus_val16 *x, const opus_val16 *y,
int N)
{
int i;
float xy;
__m128 sum;
sum = _mm_setzero_ps();
/* FIXME: We should probably go 8-way and use 2 sums. */
for (i=0;i<N-3;i+=4)
{
__m128 xi = _mm_loadu_ps(x+i);
__m128 yi = _mm_loadu_ps(y+i);
sum = _mm_add_ps(sum,_mm_mul_ps(xi, yi));
}
/* Horizontal sum */
sum = _mm_add_ps(sum, _mm_movehl_ps(sum, sum));
sum = _mm_add_ss(sum, _mm_shuffle_ps(sum, sum, 0x55));
_mm_store_ss(&xy, sum);
for (;i<N;i++)
{
xy = MAC16_16(xy, x[i], y[i]);
}
return xy;
}
void comb_filter_const_sse(opus_val32 *y, opus_val32 *x, int T, int N,
opus_val16 g10, opus_val16 g11, opus_val16 g12)
{
int i;
__m128 x0v;
__m128 g10v, g11v, g12v;
g10v = _mm_load1_ps(&g10);
g11v = _mm_load1_ps(&g11);
g12v = _mm_load1_ps(&g12);
x0v = _mm_loadu_ps(&x[-T-2]);
for (i=0;i<N-3;i+=4)
{
__m128 yi, yi2, x1v, x2v, x3v, x4v;
const opus_val32 *xp = &x[i-T-2];
yi = _mm_loadu_ps(x+i);
x4v = _mm_loadu_ps(xp+4);
#if 0
/* Slower version with all loads */
x1v = _mm_loadu_ps(xp+1);
x2v = _mm_loadu_ps(xp+2);
x3v = _mm_loadu_ps(xp+3);
#else
x2v = _mm_shuffle_ps(x0v, x4v, 0x4e);
x1v = _mm_shuffle_ps(x0v, x2v, 0x99);
x3v = _mm_shuffle_ps(x2v, x4v, 0x99);
#endif
yi = _mm_add_ps(yi, _mm_mul_ps(g10v,x2v));
#if 0 /* Set to 1 to make it bit-exact with the non-SSE version */
yi = _mm_add_ps(yi, _mm_mul_ps(g11v,_mm_add_ps(x3v,x1v)));
yi = _mm_add_ps(yi, _mm_mul_ps(g12v,_mm_add_ps(x4v,x0v)));
#else
/* Use partial sums */
yi2 = _mm_add_ps(_mm_mul_ps(g11v,_mm_add_ps(x3v,x1v)),
_mm_mul_ps(g12v,_mm_add_ps(x4v,x0v)));
yi = _mm_add_ps(yi, yi2);
#endif
x0v=x4v;
_mm_storeu_ps(y+i, yi);
}
#ifdef CUSTOM_MODES
for (;i<N;i++)
{
y[i] = x[i]
+ MULT16_32_Q15(g10,x[i-T])
+ MULT16_32_Q15(g11,ADD32(x[i-T+1],x[i-T-1]))
+ MULT16_32_Q15(g12,ADD32(x[i-T+2],x[i-T-2]));
}
#endif
}
#endif

192
Frameworks/Opus/Opus/opus/celt/x86/pitch_sse.h Normal file
View file

@ -0,0 +1,192 @@
/* Copyright (c) 2013 Jean-Marc Valin and John Ridges
Copyright (c) 2014, Cisco Systems, INC MingXiang WeiZhou MinPeng YanWang*/
/**
@file pitch_sse.h
@brief Pitch analysis
*/
/*
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef PITCH_SSE_H
#define PITCH_SSE_H
#if defined(HAVE_CONFIG_H)
#include "config.h"
#endif
#if defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)
void xcorr_kernel_sse4_1(
const opus_int16 *x,
const opus_int16 *y,
opus_val32 sum[4],
int len);
#endif
#if defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT)
void xcorr_kernel_sse(
const opus_val16 *x,
const opus_val16 *y,
opus_val32 sum[4],
int len);
#endif
#if defined(OPUS_X86_PRESUME_SSE4_1) && defined(FIXED_POINT)
#define OVERRIDE_XCORR_KERNEL
#define xcorr_kernel(x, y, sum, len, arch) \
((void)arch, xcorr_kernel_sse4_1(x, y, sum, len))
#elif defined(OPUS_X86_PRESUME_SSE) && !defined(FIXED_POINT)
#define OVERRIDE_XCORR_KERNEL
#define xcorr_kernel(x, y, sum, len, arch) \
((void)arch, xcorr_kernel_sse(x, y, sum, len))
#elif (defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)) || (defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT))
extern void (*const XCORR_KERNEL_IMPL[OPUS_ARCHMASK + 1])(
const opus_val16 *x,
const opus_val16 *y,
opus_val32 sum[4],
int len);
#define OVERRIDE_XCORR_KERNEL
#define xcorr_kernel(x, y, sum, len, arch) \
((*XCORR_KERNEL_IMPL[(arch) & OPUS_ARCHMASK])(x, y, sum, len))
#endif
#if defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)
opus_val32 celt_inner_prod_sse4_1(
const opus_int16 *x,
const opus_int16 *y,
int N);
#endif
#if defined(OPUS_X86_MAY_HAVE_SSE2) && defined(FIXED_POINT)
opus_val32 celt_inner_prod_sse2(
const opus_int16 *x,
const opus_int16 *y,
int N);
#endif
#if defined(OPUS_X86_MAY_HAVE_SSE2) && !defined(FIXED_POINT)
opus_val32 celt_inner_prod_sse(
const opus_val16 *x,
const opus_val16 *y,
int N);
#endif
#if defined(OPUS_X86_PRESUME_SSE4_1) && defined(FIXED_POINT)
#define OVERRIDE_CELT_INNER_PROD
#define celt_inner_prod(x, y, N, arch) \
((void)arch, celt_inner_prod_sse4_1(x, y, N))
#elif defined(OPUS_X86_PRESUME_SSE2) && defined(FIXED_POINT) && !defined(OPUS_X86_MAY_HAVE_SSE4_1)
#define OVERRIDE_CELT_INNER_PROD
#define celt_inner_prod(x, y, N, arch) \
((void)arch, celt_inner_prod_sse2(x, y, N))
#elif defined(OPUS_X86_PRESUME_SSE) && !defined(FIXED_POINT)
#define OVERRIDE_CELT_INNER_PROD
#define celt_inner_prod(x, y, N, arch) \
((void)arch, celt_inner_prod_sse(x, y, N))
#elif ((defined(OPUS_X86_MAY_HAVE_SSE4_1) || defined(OPUS_X86_MAY_HAVE_SSE2)) && defined(FIXED_POINT)) || \
(defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT))
extern opus_val32 (*const CELT_INNER_PROD_IMPL[OPUS_ARCHMASK + 1])(
const opus_val16 *x,
const opus_val16 *y,
int N);
#define OVERRIDE_CELT_INNER_PROD
#define celt_inner_prod(x, y, N, arch) \
((*CELT_INNER_PROD_IMPL[(arch) & OPUS_ARCHMASK])(x, y, N))
#endif
#if defined(OPUS_X86_MAY_HAVE_SSE) && !defined(FIXED_POINT)
#define OVERRIDE_DUAL_INNER_PROD
#define OVERRIDE_COMB_FILTER_CONST
#undef dual_inner_prod
#undef comb_filter_const
void dual_inner_prod_sse(const opus_val16 *x,
const opus_val16 *y01,
const opus_val16 *y02,
int N,
opus_val32 *xy1,
opus_val32 *xy2);
void comb_filter_const_sse(opus_val32 *y,
opus_val32 *x,
int T,
int N,
opus_val16 g10,
opus_val16 g11,
opus_val16 g12);
#if defined(OPUS_X86_PRESUME_SSE)
# define dual_inner_prod(x, y01, y02, N, xy1, xy2, arch) \
((void)(arch),dual_inner_prod_sse(x, y01, y02, N, xy1, xy2))
# define comb_filter_const(y, x, T, N, g10, g11, g12, arch) \
((void)(arch),comb_filter_const_sse(y, x, T, N, g10, g11, g12))
#else
extern void (*const DUAL_INNER_PROD_IMPL[OPUS_ARCHMASK + 1])(
const opus_val16 *x,
const opus_val16 *y01,
const opus_val16 *y02,
int N,
opus_val32 *xy1,
opus_val32 *xy2);
#define dual_inner_prod(x, y01, y02, N, xy1, xy2, arch) \
((*DUAL_INNER_PROD_IMPL[(arch) & OPUS_ARCHMASK])(x, y01, y02, N, xy1, xy2))
extern void (*const COMB_FILTER_CONST_IMPL[OPUS_ARCHMASK + 1])(
opus_val32 *y,
opus_val32 *x,
int T,
int N,
opus_val16 g10,
opus_val16 g11,
opus_val16 g12);
#define comb_filter_const(y, x, T, N, g10, g11, g12, arch) \
((*COMB_FILTER_CONST_IMPL[(arch) & OPUS_ARCHMASK])(y, x, T, N, g10, g11, g12))
#define NON_STATIC_COMB_FILTER_CONST_C
#endif
#endif
#endif

95
Frameworks/Opus/Opus/opus/celt/x86/pitch_sse2.c Normal file
View file

@ -0,0 +1,95 @@
/* Copyright (c) 2014, Cisco Systems, INC
Written by XiangMingZhu WeiZhou MinPeng YanWang
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <xmmintrin.h>
#include <emmintrin.h>
#include "macros.h"
#include "celt_lpc.h"
#include "stack_alloc.h"
#include "mathops.h"
#include "pitch.h"
#if defined(OPUS_X86_MAY_HAVE_SSE2) && defined(FIXED_POINT)
opus_val32 celt_inner_prod_sse2(const opus_val16 *x, const opus_val16 *y,
int N)
{
opus_int i, dataSize16;
opus_int32 sum;
__m128i inVec1_76543210, inVec1_FEDCBA98, acc1;
__m128i inVec2_76543210, inVec2_FEDCBA98, acc2;
sum = 0;
dataSize16 = N & ~15;
acc1 = _mm_setzero_si128();
acc2 = _mm_setzero_si128();
for (i=0;i<dataSize16;i+=16)
{
inVec1_76543210 = _mm_loadu_si128((__m128i *)(&x[i + 0]));
inVec2_76543210 = _mm_loadu_si128((__m128i *)(&y[i + 0]));
inVec1_FEDCBA98 = _mm_loadu_si128((__m128i *)(&x[i + 8]));
inVec2_FEDCBA98 = _mm_loadu_si128((__m128i *)(&y[i + 8]));
inVec1_76543210 = _mm_madd_epi16(inVec1_76543210, inVec2_76543210);
inVec1_FEDCBA98 = _mm_madd_epi16(inVec1_FEDCBA98, inVec2_FEDCBA98);
acc1 = _mm_add_epi32(acc1, inVec1_76543210);
acc2 = _mm_add_epi32(acc2, inVec1_FEDCBA98);
}
acc1 = _mm_add_epi32( acc1, acc2 );
if (N - i >= 8)
{
inVec1_76543210 = _mm_loadu_si128((__m128i *)(&x[i + 0]));
inVec2_76543210 = _mm_loadu_si128((__m128i *)(&y[i + 0]));
inVec1_76543210 = _mm_madd_epi16(inVec1_76543210, inVec2_76543210);
acc1 = _mm_add_epi32(acc1, inVec1_76543210);
i += 8;
}
acc1 = _mm_add_epi32(acc1, _mm_unpackhi_epi64( acc1, acc1));
acc1 = _mm_add_epi32(acc1, _mm_shufflelo_epi16( acc1, 0x0E));
sum += _mm_cvtsi128_si32(acc1);
for (;i<N;i++) {
sum = silk_SMLABB(sum, x[i], y[i]);
}
return sum;
}
#endif

195
Frameworks/Opus/Opus/opus/celt/x86/pitch_sse4_1.c Normal file
View file

@ -0,0 +1,195 @@
/* Copyright (c) 2014, Cisco Systems, INC
Written by XiangMingZhu WeiZhou MinPeng YanWang
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <xmmintrin.h>
#include <emmintrin.h>
#include "macros.h"
#include "celt_lpc.h"
#include "stack_alloc.h"
#include "mathops.h"
#include "pitch.h"
#if defined(OPUS_X86_MAY_HAVE_SSE4_1) && defined(FIXED_POINT)
#include <smmintrin.h>
#include "x86cpu.h"
opus_val32 celt_inner_prod_sse4_1(const opus_val16 *x, const opus_val16 *y,
int N)
{
opus_int i, dataSize16;
opus_int32 sum;
__m128i inVec1_76543210, inVec1_FEDCBA98, acc1;
__m128i inVec2_76543210, inVec2_FEDCBA98, acc2;
__m128i inVec1_3210, inVec2_3210;
sum = 0;
dataSize16 = N & ~15;
acc1 = _mm_setzero_si128();
acc2 = _mm_setzero_si128();
for (i=0;i<dataSize16;i+=16) {
inVec1_76543210 = _mm_loadu_si128((__m128i *)(&x[i + 0]));
inVec2_76543210 = _mm_loadu_si128((__m128i *)(&y[i + 0]));
inVec1_FEDCBA98 = _mm_loadu_si128((__m128i *)(&x[i + 8]));
inVec2_FEDCBA98 = _mm_loadu_si128((__m128i *)(&y[i + 8]));
inVec1_76543210 = _mm_madd_epi16(inVec1_76543210, inVec2_76543210);
inVec1_FEDCBA98 = _mm_madd_epi16(inVec1_FEDCBA98, inVec2_FEDCBA98);
acc1 = _mm_add_epi32(acc1, inVec1_76543210);
acc2 = _mm_add_epi32(acc2, inVec1_FEDCBA98);
}
acc1 = _mm_add_epi32(acc1, acc2);
if (N - i >= 8)
{
inVec1_76543210 = _mm_loadu_si128((__m128i *)(&x[i + 0]));
inVec2_76543210 = _mm_loadu_si128((__m128i *)(&y[i + 0]));
inVec1_76543210 = _mm_madd_epi16(inVec1_76543210, inVec2_76543210);
acc1 = _mm_add_epi32(acc1, inVec1_76543210);
i += 8;
}
if (N - i >= 4)
{
inVec1_3210 = OP_CVTEPI16_EPI32_M64(&x[i + 0]);
inVec2_3210 = OP_CVTEPI16_EPI32_M64(&y[i + 0]);
inVec1_3210 = _mm_mullo_epi32(inVec1_3210, inVec2_3210);
acc1 = _mm_add_epi32(acc1, inVec1_3210);
i += 4;
}
acc1 = _mm_add_epi32(acc1, _mm_unpackhi_epi64(acc1, acc1));
acc1 = _mm_add_epi32(acc1, _mm_shufflelo_epi16(acc1, 0x0E));
sum += _mm_cvtsi128_si32(acc1);
for (;i<N;i++)
{
sum = silk_SMLABB(sum, x[i], y[i]);
}
return sum;
}
void xcorr_kernel_sse4_1(const opus_val16 * x, const opus_val16 * y, opus_val32 sum[ 4 ], int len)
{
int j;
__m128i vecX, vecX0, vecX1, vecX2, vecX3;
__m128i vecY0, vecY1, vecY2, vecY3;
__m128i sum0, sum1, sum2, sum3, vecSum;
__m128i initSum;
celt_assert(len >= 3);
sum0 = _mm_setzero_si128();
sum1 = _mm_setzero_si128();
sum2 = _mm_setzero_si128();
sum3 = _mm_setzero_si128();
for (j=0;j<(len-7);j+=8)
{
vecX = _mm_loadu_si128((__m128i *)(&x[j + 0]));
vecY0 = _mm_loadu_si128((__m128i *)(&y[j + 0]));
vecY1 = _mm_loadu_si128((__m128i *)(&y[j + 1]));
vecY2 = _mm_loadu_si128((__m128i *)(&y[j + 2]));
vecY3 = _mm_loadu_si128((__m128i *)(&y[j + 3]));
sum0 = _mm_add_epi32(sum0, _mm_madd_epi16(vecX, vecY0));
sum1 = _mm_add_epi32(sum1, _mm_madd_epi16(vecX, vecY1));
sum2 = _mm_add_epi32(sum2, _mm_madd_epi16(vecX, vecY2));
sum3 = _mm_add_epi32(sum3, _mm_madd_epi16(vecX, vecY3));
}
sum0 = _mm_add_epi32(sum0, _mm_unpackhi_epi64( sum0, sum0));
sum0 = _mm_add_epi32(sum0, _mm_shufflelo_epi16( sum0, 0x0E));
sum1 = _mm_add_epi32(sum1, _mm_unpackhi_epi64( sum1, sum1));
sum1 = _mm_add_epi32(sum1, _mm_shufflelo_epi16( sum1, 0x0E));
sum2 = _mm_add_epi32(sum2, _mm_unpackhi_epi64( sum2, sum2));
sum2 = _mm_add_epi32(sum2, _mm_shufflelo_epi16( sum2, 0x0E));
sum3 = _mm_add_epi32(sum3, _mm_unpackhi_epi64( sum3, sum3));
sum3 = _mm_add_epi32(sum3, _mm_shufflelo_epi16( sum3, 0x0E));
vecSum = _mm_unpacklo_epi64(_mm_unpacklo_epi32(sum0, sum1),
_mm_unpacklo_epi32(sum2, sum3));
for (;j<(len-3);j+=4)
{
vecX = OP_CVTEPI16_EPI32_M64(&x[j + 0]);
vecX0 = _mm_shuffle_epi32(vecX, 0x00);
vecX1 = _mm_shuffle_epi32(vecX, 0x55);
vecX2 = _mm_shuffle_epi32(vecX, 0xaa);
vecX3 = _mm_shuffle_epi32(vecX, 0xff);
vecY0 = OP_CVTEPI16_EPI32_M64(&y[j + 0]);
vecY1 = OP_CVTEPI16_EPI32_M64(&y[j + 1]);
vecY2 = OP_CVTEPI16_EPI32_M64(&y[j + 2]);
vecY3 = OP_CVTEPI16_EPI32_M64(&y[j + 3]);
sum0 = _mm_mullo_epi32(vecX0, vecY0);
sum1 = _mm_mullo_epi32(vecX1, vecY1);
sum2 = _mm_mullo_epi32(vecX2, vecY2);
sum3 = _mm_mullo_epi32(vecX3, vecY3);
sum0 = _mm_add_epi32(sum0, sum1);
sum2 = _mm_add_epi32(sum2, sum3);
vecSum = _mm_add_epi32(vecSum, sum0);
vecSum = _mm_add_epi32(vecSum, sum2);
}
for (;j<len;j++)
{
vecX = OP_CVTEPI16_EPI32_M64(&x[j + 0]);
vecX0 = _mm_shuffle_epi32(vecX, 0x00);
vecY0 = OP_CVTEPI16_EPI32_M64(&y[j + 0]);
sum0 = _mm_mullo_epi32(vecX0, vecY0);
vecSum = _mm_add_epi32(vecSum, sum0);
}
initSum = _mm_loadu_si128((__m128i *)(&sum[0]));
initSum = _mm_add_epi32(initSum, vecSum);
_mm_storeu_si128((__m128i *)sum, initSum);
}
#endif

148
Frameworks/Opus/Opus/opus/celt/x86/x86_celt_map.c Normal file
View file

@ -0,0 +1,148 @@
/* Copyright (c) 2014, Cisco Systems, INC
Written by XiangMingZhu WeiZhou MinPeng YanWang
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#if defined(HAVE_CONFIG_H)
#include "config.h"
#endif
#include "x86/x86cpu.h"
#include "celt_lpc.h"
#include "pitch.h"
#include "pitch_sse.h"
#if defined(OPUS_HAVE_RTCD)
# if defined(FIXED_POINT)
#if defined(OPUS_X86_MAY_HAVE_SSE4_1) && !defined(OPUS_X86_PRESUME_SSE4_1)
void (*const CELT_FIR_IMPL[OPUS_ARCHMASK + 1])(
const opus_val16 *x,
const opus_val16 *num,
opus_val16 *y,
int N,
int ord,
opus_val16 *mem,
int arch
) = {
celt_fir_c, /* non-sse */
celt_fir_c,
celt_fir_c,
MAY_HAVE_SSE4_1(celt_fir), /* sse4.1 */
};
void (*const XCORR_KERNEL_IMPL[OPUS_ARCHMASK + 1])(
const opus_val16 *x,
const opus_val16 *y,
opus_val32 sum[4],
int len
) = {
xcorr_kernel_c, /* non-sse */
xcorr_kernel_c,
xcorr_kernel_c,
MAY_HAVE_SSE4_1(xcorr_kernel), /* sse4.1 */
};
#endif
#if (defined(OPUS_X86_MAY_HAVE_SSE4_1) && !defined(OPUS_X86_PRESUME_SSE4_1)) || \
(!defined(OPUS_X86_MAY_HAVE_SSE_4_1) && defined(OPUS_X86_MAY_HAVE_SSE2) && !defined(OPUS_X86_PRESUME_SSE2))
opus_val32 (*const CELT_INNER_PROD_IMPL[OPUS_ARCHMASK + 1])(
const opus_val16 *x,
const opus_val16 *y,
int N
) = {
celt_inner_prod_c, /* non-sse */
celt_inner_prod_c,
MAY_HAVE_SSE2(celt_inner_prod),
MAY_HAVE_SSE4_1(celt_inner_prod), /* sse4.1 */
};
#endif
# else
#if defined(OPUS_X86_MAY_HAVE_SSE) && !defined(OPUS_X86_PRESUME_SSE)
void (*const XCORR_KERNEL_IMPL[OPUS_ARCHMASK + 1])(
const opus_val16 *x,
const opus_val16 *y,
opus_val32 sum[4],
int len
) = {
xcorr_kernel_c, /* non-sse */
MAY_HAVE_SSE(xcorr_kernel),
MAY_HAVE_SSE(xcorr_kernel),
MAY_HAVE_SSE(xcorr_kernel),
};
opus_val32 (*const CELT_INNER_PROD_IMPL[OPUS_ARCHMASK + 1])(
const opus_val16 *x,
const opus_val16 *y,
int N
) = {
celt_inner_prod_c, /* non-sse */
MAY_HAVE_SSE(celt_inner_prod),
MAY_HAVE_SSE(celt_inner_prod),
MAY_HAVE_SSE(celt_inner_prod),
};
void (*const DUAL_INNER_PROD_IMPL[OPUS_ARCHMASK + 1])(
const opus_val16 *x,
const opus_val16 *y01,
const opus_val16 *y02,
int N,
opus_val32 *xy1,
opus_val32 *xy2
) = {
dual_inner_prod_c, /* non-sse */
MAY_HAVE_SSE(dual_inner_prod),
MAY_HAVE_SSE(dual_inner_prod),
MAY_HAVE_SSE(dual_inner_prod),
};
void (*const COMB_FILTER_CONST_IMPL[OPUS_ARCHMASK + 1])(
opus_val32 *y,
opus_val32 *x,
int T,
int N,
opus_val16 g10,
opus_val16 g11,
opus_val16 g12
) = {
comb_filter_const_c, /* non-sse */
MAY_HAVE_SSE(comb_filter_const),
MAY_HAVE_SSE(comb_filter_const),
MAY_HAVE_SSE(comb_filter_const),
};
#endif
#endif
#endif

146
Frameworks/Opus/Opus/opus/celt/x86/x86cpu.c Normal file
View file

@ -0,0 +1,146 @@
/* Copyright (c) 2014, Cisco Systems, INC
Written by XiangMingZhu WeiZhou MinPeng YanWang
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include "cpu_support.h"
#include "macros.h"
#include "main.h"
#include "pitch.h"
#include "x86cpu.h"
#if (defined(OPUS_X86_MAY_HAVE_SSE) && !defined(OPUS_X86_PRESUME_SSE)) || \
(defined(OPUS_X86_MAY_HAVE_SSE2) && !defined(OPUS_X86_PRESUME_SSE2)) || \
(defined(OPUS_X86_MAY_HAVE_SSE4_1) && !defined(OPUS_X86_PRESUME_SSE4_1))
#if defined(_MSC_VER)
#include <intrin.h>
static _inline void cpuid(unsigned int CPUInfo[4], unsigned int InfoType)
{
__cpuid((int*)CPUInfo, InfoType);
}
#else
#if defined(CPU_INFO_BY_C)
#include <cpuid.h>
#endif
static void cpuid(unsigned int CPUInfo[4], unsigned int InfoType)
{
#if defined(CPU_INFO_BY_ASM)
#if defined(__i386__) && defined(__PIC__)
/* %ebx is PIC register in 32-bit, so mustn't clobber it. */
__asm__ __volatile__ (
"xchg %%ebx, %1\n"
"cpuid\n"
"xchg %%ebx, %1\n":
"=a" (CPUInfo[0]),
"=r" (CPUInfo[1]),
"=c" (CPUInfo[2]),
"=d" (CPUInfo[3]) :
"0" (InfoType)
);
#else
__asm__ __volatile__ (
"cpuid":
"=a" (CPUInfo[0]),
"=b" (CPUInfo[1]),
"=c" (CPUInfo[2]),
"=d" (CPUInfo[3]) :
"0" (InfoType)
);
#endif
#elif defined(CPU_INFO_BY_C)
__get_cpuid(InfoType, &(CPUInfo[0]), &(CPUInfo[1]), &(CPUInfo[2]), &(CPUInfo[3]));
#endif
}
#endif
typedef struct CPU_Feature{
/* SIMD: 128-bit */
int HW_SSE;
int HW_SSE2;
int HW_SSE41;
} CPU_Feature;
static void opus_cpu_feature_check(CPU_Feature *cpu_feature)
{
unsigned int info[4] = {0};
unsigned int nIds = 0;
cpuid(info, 0);
nIds = info[0];
if (nIds >= 1){
cpuid(info, 1);
cpu_feature->HW_SSE = (info[3] & (1 << 25)) != 0;
cpu_feature->HW_SSE2 = (info[3] & (1 << 26)) != 0;
cpu_feature->HW_SSE41 = (info[2] & (1 << 19)) != 0;
}
else {
cpu_feature->HW_SSE = 0;
cpu_feature->HW_SSE2 = 0;
cpu_feature->HW_SSE41 = 0;
}
}
int opus_select_arch(void)
{
CPU_Feature cpu_feature;
int arch;
opus_cpu_feature_check(&cpu_feature);
arch = 0;
if (!cpu_feature.HW_SSE)
{
return arch;
}
arch++;
if (!cpu_feature.HW_SSE2)
{
return arch;
}
arch++;
if (!cpu_feature.HW_SSE41)
{
return arch;
}
arch++;
return arch;
}
#endif

87
Frameworks/Opus/Opus/opus/celt/x86/x86cpu.h Normal file
View file

@ -0,0 +1,87 @@
/* Copyright (c) 2014, Cisco Systems, INC
Written by XiangMingZhu WeiZhou MinPeng YanWang
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
- Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#if !defined(X86CPU_H)
# define X86CPU_H
# if defined(OPUS_X86_MAY_HAVE_SSE)
# define MAY_HAVE_SSE(name) name ## _sse
# else
# define MAY_HAVE_SSE(name) name ## _c
# endif
# if defined(OPUS_X86_MAY_HAVE_SSE2)
# define MAY_HAVE_SSE2(name) name ## _sse2
# else
# define MAY_HAVE_SSE2(name) name ## _c
# endif
# if defined(OPUS_X86_MAY_HAVE_SSE4_1)
# define MAY_HAVE_SSE4_1(name) name ## _sse4_1
# else
# define MAY_HAVE_SSE4_1(name) name ## _c
# endif
# if defined(OPUS_HAVE_RTCD)
int opus_select_arch(void);
# endif
/*gcc appears to emit MOVDQA's to load the argument of an _mm_cvtepi8_epi32()
or _mm_cvtepi16_epi32() when optimizations are disabled, even though the
actual PMOVSXWD instruction takes an m32 or m64. Unlike a normal memory
reference, these require 16-byte alignment and load a full 16 bytes (instead
of 4 or 8), possibly reading out of bounds.
We can insert an explicit MOVD or MOVQ using _mm_cvtsi32_si128() or
_mm_loadl_epi64(), which should have the same semantics as an m32 or m64
reference in the PMOVSXWD instruction itself, but gcc is not smart enough to
optimize this out when optimizations ARE enabled.
Clang, in contrast, requires us to do this always for _mm_cvtepi8_epi32
(which is fair, since technically the compiler is always allowed to do the
dereference before invoking the function implementing the intrinsic).
However, it is smart enough to eliminate the extra MOVD instruction.
For _mm_cvtepi16_epi32, it does the right thing, though does *not* optimize out
the extra MOVQ if it's specified explicitly */
# if defined(__clang__) || !defined(__OPTIMIZE__)
# define OP_CVTEPI8_EPI32_M32(x) \
(_mm_cvtepi8_epi32(_mm_cvtsi32_si128(*(int *)(x))))
# else
# define OP_CVTEPI8_EPI32_M32(x) \
(_mm_cvtepi8_epi32(*(__m128i *)(x)))
#endif
# if !defined(__OPTIMIZE__)
# define OP_CVTEPI16_EPI32_M64(x) \
(_mm_cvtepi16_epi32(_mm_loadl_epi64((__m128i *)(x))))
# else
# define OP_CVTEPI16_EPI32_M64(x) \
(_mm_cvtepi16_epi32(*(__m128i *)(x)))
# endif
#endif

22
Frameworks/Opus/Opus/opus/celt_headers.mk
View file

@ -2,6 +2,7 @@ CELT_HEAD = \
celt/arch.h \
celt/bands.h \
celt/celt.h \
celt/cpu_support.h \
include/opus_types.h \
include/opus_defines.h \
include/opus_custom.h \
@ -23,9 +24,28 @@ celt/modes.h \
celt/os_support.h \
celt/pitch.h \
celt/celt_lpc.h \
celt/x86/celt_lpc_sse.h \
celt/quant_bands.h \
celt/rate.h \
celt/stack_alloc.h \
celt/vq.h \
celt/static_modes_float.h \
celt/static_modes_fixed.h
celt/static_modes_fixed.h \
celt/static_modes_float_arm_ne10.h \
celt/static_modes_fixed_arm_ne10.h \
celt/arm/armcpu.h \
celt/arm/fixed_armv4.h \
celt/arm/fixed_armv5e.h \
celt/arm/kiss_fft_armv4.h \
celt/arm/kiss_fft_armv5e.h \
celt/arm/pitch_arm.h \
celt/arm/fft_arm.h \
celt/arm/mdct_arm.h \
celt/mips/celt_mipsr1.h \
celt/mips/fixed_generic_mipsr1.h \
celt/mips/kiss_fft_mipsr1.h \
celt/mips/mdct_mipsr1.h \
celt/mips/pitch_mipsr1.h \
celt/mips/vq_mipsr1.h \
celt/x86/pitch_sse.h \
celt/x86/x86cpu.h

28
Frameworks/Opus/Opus/opus/celt_sources.mk
View file

@ -1,5 +1,7 @@
CELT_SOURCES = celt/bands.c \
celt/celt.c \
celt/celt_encoder.c \
celt/celt_decoder.c \
celt/cwrs.c \
celt/entcode.c \
celt/entdec.c \
@ -14,3 +16,29 @@ celt/celt_lpc.c \
celt/quant_bands.c \
celt/rate.c \
celt/vq.c
CELT_SOURCES_SSE = celt/x86/x86cpu.c \
celt/x86/x86_celt_map.c \
celt/x86/pitch_sse.c
CELT_SOURCES_SSE2 = celt/x86/pitch_sse2.c
CELT_SOURCES_SSE4_1 = celt/x86/celt_lpc_sse.c \
celt/x86/pitch_sse4_1.c
CELT_SOURCES_ARM = \
celt/arm/armcpu.c \
celt/arm/arm_celt_map.c
CELT_SOURCES_ARM_ASM = \
celt/arm/celt_pitch_xcorr_arm.s
CELT_AM_SOURCES_ARM_ASM = \
celt/arm/armopts.s.in
CELT_SOURCES_ARM_NEON_INTR = \
celt/arm/celt_neon_intr.c
CELT_SOURCES_ARM_NE10= \
celt/arm/celt_ne10_fft.c \
celt/arm/celt_ne10_mdct.c

10
Frameworks/Opus/Opus/opus/compile
View file

@ -1,10 +1,9 @@
#! /bin/sh
# Wrapper for compilers which do not understand '-c -o'.
scriptversion=2012-03-05.13; # UTC
scriptversion=2012-10-14.11; # UTC
# Copyright (C) 1999, 2000, 2003, 2004, 2005, 2009, 2010, 2012 Free
# Software Foundation, Inc.
# Copyright (C) 1999-2014 Free Software Foundation, Inc.
# Written by Tom Tromey <tromey@cygnus.com>.
#
# This program is free software; you can redistribute it and/or modify
@ -113,6 +112,11 @@ func_cl_dashl ()
lib=$dir/$lib.lib
break
fi
if test -f "$dir/lib$lib.a"; then
found=yes
lib=$dir/lib$lib.a
break
fi
done
IFS=$save_IFS

337
Frameworks/Opus/Opus/opus/config.guess vendored
View file

@ -1,14 +1,12 @@
#! /bin/sh
# Attempt to guess a canonical system name.
# Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
# 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010,
# 2011, 2012 Free Software Foundation, Inc.
# Copyright 1992-2015 Free Software Foundation, Inc.
timestamp='2012-02-10'
timestamp='2015-01-01'
# This file is free software; you can redistribute it and/or modify it
# under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful, but
@ -22,19 +20,17 @@ timestamp='2012-02-10'
# As a special exception to the GNU General Public License, if you
# distribute this file as part of a program that contains a
# configuration script generated by Autoconf, you may include it under
# the same distribution terms that you use for the rest of that program.
# Originally written by Per Bothner. Please send patches (context
# diff format) to <config-patches@gnu.org> and include a ChangeLog
# entry.
# the same distribution terms that you use for the rest of that
# program. This Exception is an additional permission under section 7
# of the GNU General Public License, version 3 ("GPLv3").
#
# This script attempts to guess a canonical system name similar to
# config.sub. If it succeeds, it prints the system name on stdout, and
# exits with 0. Otherwise, it exits with 1.
# Originally written by Per Bothner; maintained since 2000 by Ben Elliston.
#
# You can get the latest version of this script from:
# http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD
#
# Please send patches to <config-patches@gnu.org>.
me=`echo "$0" | sed -e 's,.*/,,'`
@ -54,9 +50,7 @@ version="\
GNU config.guess ($timestamp)
Originally written by Per Bothner.
Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012
Free Software Foundation, Inc.
Copyright 1992-2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
@ -138,6 +132,27 @@ UNAME_RELEASE=`(uname -r) 2>/dev/null` || UNAME_RELEASE=unknown
UNAME_SYSTEM=`(uname -s) 2>/dev/null` || UNAME_SYSTEM=unknown
UNAME_VERSION=`(uname -v) 2>/dev/null` || UNAME_VERSION=unknown
case "${UNAME_SYSTEM}" in
Linux|GNU|GNU/*)
# If the system lacks a compiler, then just pick glibc.
# We could probably try harder.
LIBC=gnu
eval $set_cc_for_build
cat <<-EOF > $dummy.c
#include <features.h>
#if defined(__UCLIBC__)
LIBC=uclibc
#elif defined(__dietlibc__)
LIBC=dietlibc
#else
LIBC=gnu
#endif
EOF
eval `$CC_FOR_BUILD -E $dummy.c 2>/dev/null | grep '^LIBC' | sed 's, ,,g'`
;;
esac
# Note: order is significant - the case branches are not exclusive.
case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
@ -200,6 +215,10 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
# CPU_TYPE-MANUFACTURER-OPERATING_SYSTEM is used.
echo "${machine}-${os}${release}"
exit ;;
*:Bitrig:*:*)
UNAME_MACHINE_ARCH=`arch | sed 's/Bitrig.//'`
echo ${UNAME_MACHINE_ARCH}-unknown-bitrig${UNAME_RELEASE}
exit ;;
*:OpenBSD:*:*)
UNAME_MACHINE_ARCH=`arch | sed 's/OpenBSD.//'`
echo ${UNAME_MACHINE_ARCH}-unknown-openbsd${UNAME_RELEASE}
@ -302,7 +321,7 @@ case "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" in
arm:RISC*:1.[012]*:*|arm:riscix:1.[012]*:*)
echo arm-acorn-riscix${UNAME_RELEASE}
exit ;;
arm:riscos:*:*|arm:RISCOS:*:*)
arm*:riscos:*:*|arm*:RISCOS:*:*)
echo arm-unknown-riscos
exit ;;
SR2?01:HI-UX/MPP:*:* | SR8000:HI-UX/MPP:*:*)
@ -560,8 +579,9 @@ EOF
else
IBM_ARCH=powerpc
fi
if [ -x /usr/bin/oslevel ] ; then
IBM_REV=`/usr/bin/oslevel`
if [ -x /usr/bin/lslpp ] ; then
IBM_REV=`/usr/bin/lslpp -Lqc bos.rte.libc |
awk -F: '{ print $3 }' | sed s/[0-9]*$/0/`
else
IBM_REV=${UNAME_VERSION}.${UNAME_RELEASE}
fi
@ -801,10 +821,13 @@ EOF
i*:CYGWIN*:*)
echo ${UNAME_MACHINE}-pc-cygwin
exit ;;
*:MINGW64*:*)
echo ${UNAME_MACHINE}-pc-mingw64
exit ;;
*:MINGW*:*)
echo ${UNAME_MACHINE}-pc-mingw32
exit ;;
i*:MSYS*:*)
*:MSYS*:*)
echo ${UNAME_MACHINE}-pc-msys
exit ;;
i*:windows32*:*)
@ -852,21 +875,21 @@ EOF
exit ;;
*:GNU:*:*)
# the GNU system
echo `echo ${UNAME_MACHINE}|sed -e 's,[-/].*$,,'`-unknown-gnu`echo ${UNAME_RELEASE}|sed -e 's,/.*$,,'`
echo `echo ${UNAME_MACHINE}|sed -e 's,[-/].*$,,'`-unknown-${LIBC}`echo ${UNAME_RELEASE}|sed -e 's,/.*$,,'`
exit ;;
*:GNU/*:*:*)
# other systems with GNU libc and userland
echo ${UNAME_MACHINE}-unknown-`echo ${UNAME_SYSTEM} | sed 's,^[^/]*/,,' | tr '[A-Z]' '[a-z]'``echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'`-gnu
echo ${UNAME_MACHINE}-unknown-`echo ${UNAME_SYSTEM} | sed 's,^[^/]*/,,' | tr '[A-Z]' '[a-z]'``echo ${UNAME_RELEASE}|sed -e 's/[-(].*//'`-${LIBC}
exit ;;
i*86:Minix:*:*)
echo ${UNAME_MACHINE}-pc-minix
exit ;;
aarch64:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
aarch64_be:Linux:*:*)
UNAME_MACHINE=aarch64_be
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
alpha:Linux:*:*)
case `sed -n '/^cpu model/s/^.*: \(.*\)/\1/p' < /proc/cpuinfo` in
@ -879,59 +902,54 @@ EOF
EV68*) UNAME_MACHINE=alphaev68 ;;
esac
objdump --private-headers /bin/sh | grep -q ld.so.1
if test "$?" = 0 ; then LIBC="libc1" ; else LIBC="" ; fi
echo ${UNAME_MACHINE}-unknown-linux-gnu${LIBC}
if test "$?" = 0 ; then LIBC="gnulibc1" ; fi
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
arc:Linux:*:* | arceb:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
arm*:Linux:*:*)
eval $set_cc_for_build
if echo __ARM_EABI__ | $CC_FOR_BUILD -E - 2>/dev/null \
| grep -q __ARM_EABI__
then
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
else
if echo __ARM_PCS_VFP | $CC_FOR_BUILD -E - 2>/dev/null \
| grep -q __ARM_PCS_VFP
then
echo ${UNAME_MACHINE}-unknown-linux-gnueabi
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}eabi
else
echo ${UNAME_MACHINE}-unknown-linux-gnueabihf
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}eabihf
fi
fi
exit ;;
avr32*:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
cris:Linux:*:*)
echo ${UNAME_MACHINE}-axis-linux-gnu
echo ${UNAME_MACHINE}-axis-linux-${LIBC}
exit ;;
crisv32:Linux:*:*)
echo ${UNAME_MACHINE}-axis-linux-gnu
echo ${UNAME_MACHINE}-axis-linux-${LIBC}
exit ;;
frv:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
hexagon:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
i*86:Linux:*:*)
LIBC=gnu
eval $set_cc_for_build
sed 's/^ //' << EOF >$dummy.c
#ifdef __dietlibc__
LIBC=dietlibc
#endif
EOF
eval `$CC_FOR_BUILD -E $dummy.c 2>/dev/null | grep '^LIBC'`
echo "${UNAME_MACHINE}-pc-linux-${LIBC}"
echo ${UNAME_MACHINE}-pc-linux-${LIBC}
exit ;;
ia64:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
m32r*:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
m68*:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
mips:Linux:*:* | mips64:Linux:*:*)
eval $set_cc_for_build
@ -950,54 +968,63 @@ EOF
#endif
EOF
eval `$CC_FOR_BUILD -E $dummy.c 2>/dev/null | grep '^CPU'`
test x"${CPU}" != x && { echo "${CPU}-unknown-linux-gnu"; exit; }
test x"${CPU}" != x && { echo "${CPU}-unknown-linux-${LIBC}"; exit; }
;;
or32:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
openrisc*:Linux:*:*)
echo or1k-unknown-linux-${LIBC}
exit ;;
or32:Linux:*:* | or1k*:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
padre:Linux:*:*)
echo sparc-unknown-linux-gnu
echo sparc-unknown-linux-${LIBC}
exit ;;
parisc64:Linux:*:* | hppa64:Linux:*:*)
echo hppa64-unknown-linux-gnu
echo hppa64-unknown-linux-${LIBC}
exit ;;
parisc:Linux:*:* | hppa:Linux:*:*)
# Look for CPU level
case `grep '^cpu[^a-z]*:' /proc/cpuinfo 2>/dev/null | cut -d' ' -f2` in
PA7*) echo hppa1.1-unknown-linux-gnu ;;
PA8*) echo hppa2.0-unknown-linux-gnu ;;
*) echo hppa-unknown-linux-gnu ;;
PA7*) echo hppa1.1-unknown-linux-${LIBC} ;;
PA8*) echo hppa2.0-unknown-linux-${LIBC} ;;
*) echo hppa-unknown-linux-${LIBC} ;;
esac
exit ;;
ppc64:Linux:*:*)
echo powerpc64-unknown-linux-gnu
echo powerpc64-unknown-linux-${LIBC}
exit ;;
ppc:Linux:*:*)
echo powerpc-unknown-linux-gnu
echo powerpc-unknown-linux-${LIBC}
exit ;;
ppc64le:Linux:*:*)
echo powerpc64le-unknown-linux-${LIBC}
exit ;;
ppcle:Linux:*:*)
echo powerpcle-unknown-linux-${LIBC}
exit ;;
s390:Linux:*:* | s390x:Linux:*:*)
echo ${UNAME_MACHINE}-ibm-linux
echo ${UNAME_MACHINE}-ibm-linux-${LIBC}
exit ;;
sh64*:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
sh*:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
sparc:Linux:*:* | sparc64:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
tile*:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
vax:Linux:*:*)
echo ${UNAME_MACHINE}-dec-linux-gnu
echo ${UNAME_MACHINE}-dec-linux-${LIBC}
exit ;;
x86_64:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
xtensa*:Linux:*:*)
echo ${UNAME_MACHINE}-unknown-linux-gnu
echo ${UNAME_MACHINE}-unknown-linux-${LIBC}
exit ;;
i*86:DYNIX/ptx:4*:*)
# ptx 4.0 does uname -s correctly, with DYNIX/ptx in there.
@ -1201,6 +1228,9 @@ EOF
BePC:Haiku:*:*) # Haiku running on Intel PC compatible.
echo i586-pc-haiku
exit ;;
x86_64:Haiku:*:*)
echo x86_64-unknown-haiku
exit ;;
SX-4:SUPER-UX:*:*)
echo sx4-nec-superux${UNAME_RELEASE}
exit ;;
@ -1227,19 +1257,31 @@ EOF
exit ;;
*:Darwin:*:*)
UNAME_PROCESSOR=`uname -p` || UNAME_PROCESSOR=unknown
case $UNAME_PROCESSOR in
i386)
eval $set_cc_for_build
if test "$UNAME_PROCESSOR" = unknown ; then
UNAME_PROCESSOR=powerpc
fi
if test `echo "$UNAME_RELEASE" | sed -e 's/\..*//'` -le 10 ; then
if [ "$CC_FOR_BUILD" != 'no_compiler_found' ]; then
if (echo '#ifdef __LP64__'; echo IS_64BIT_ARCH; echo '#endif') | \
(CCOPTS= $CC_FOR_BUILD -E - 2>/dev/null) | \
grep IS_64BIT_ARCH >/dev/null
then
UNAME_PROCESSOR="x86_64"
fi
fi ;;
unknown) UNAME_PROCESSOR=powerpc ;;
case $UNAME_PROCESSOR in
i386) UNAME_PROCESSOR=x86_64 ;;
powerpc) UNAME_PROCESSOR=powerpc64 ;;
esac
fi
fi
elif test "$UNAME_PROCESSOR" = i386 ; then
# Avoid executing cc on OS X 10.9, as it ships with a stub
# that puts up a graphical alert prompting to install
# developer tools. Any system running Mac OS X 10.7 or
# later (Darwin 11 and later) is required to have a 64-bit
# processor. This is not true of the ARM version of Darwin
# that Apple uses in portable devices.
UNAME_PROCESSOR=x86_64
fi
echo ${UNAME_PROCESSOR}-apple-darwin${UNAME_RELEASE}
exit ;;
*:procnto*:*:* | *:QNX:[0123456789]*:*)
@ -1256,7 +1298,7 @@ EOF
NEO-?:NONSTOP_KERNEL:*:*)
echo neo-tandem-nsk${UNAME_RELEASE}
exit ;;
NSE-?:NONSTOP_KERNEL:*:*)
NSE-*:NONSTOP_KERNEL:*:*)
echo nse-tandem-nsk${UNAME_RELEASE}
exit ;;
NSR-?:NONSTOP_KERNEL:*:*)
@ -1330,157 +1372,6 @@ EOF
exit ;;
esac
#echo '(No uname command or uname output not recognized.)' 1>&2
#echo "${UNAME_MACHINE}:${UNAME_SYSTEM}:${UNAME_RELEASE}:${UNAME_VERSION}" 1>&2
eval $set_cc_for_build
cat >$dummy.c <<EOF
#ifdef _SEQUENT_
# include <sys/types.h>
# include <sys/utsname.h>
#endif
main ()
{
#if defined (sony)
#if defined (MIPSEB)
/* BFD wants "bsd" instead of "newsos". Perhaps BFD should be changed,
I don't know.... */
printf ("mips-sony-bsd\n"); exit (0);
#else
#include <sys/param.h>
printf ("m68k-sony-newsos%s\n",
#ifdef NEWSOS4
"4"
#else
""
#endif
); exit (0);
#endif
#endif
#if defined (__arm) && defined (__acorn) && defined (__unix)
printf ("arm-acorn-riscix\n"); exit (0);
#endif
#if defined (hp300) && !defined (hpux)
printf ("m68k-hp-bsd\n"); exit (0);
#endif
#if defined (NeXT)
#if !defined (__ARCHITECTURE__)
#define __ARCHITECTURE__ "m68k"
#endif
int version;
version=`(hostinfo | sed -n 's/.*NeXT Mach \([0-9]*\).*/\1/p') 2>/dev/null`;
if (version < 4)
printf ("%s-next-nextstep%d\n", __ARCHITECTURE__, version);
else
printf ("%s-next-openstep%d\n", __ARCHITECTURE__, version);
exit (0);
#endif
#if defined (MULTIMAX) || defined (n16)
#if defined (UMAXV)
printf ("ns32k-encore-sysv\n"); exit (0);
#else
#if defined (CMU)
printf ("ns32k-encore-mach\n"); exit (0);
#else
printf ("ns32k-encore-bsd\n"); exit (0);
#endif
#endif
#endif
#if defined (__386BSD__)
printf ("i386-pc-bsd\n"); exit (0);
#endif
#if defined (sequent)
#if defined (i386)
printf ("i386-sequent-dynix\n"); exit (0);
#endif
#if defined (ns32000)
printf ("ns32k-sequent-dynix\n"); exit (0);
#endif
#endif
#if defined (_SEQUENT_)
struct utsname un;
uname(&un);
if (strncmp(un.version, "V2", 2) == 0) {
printf ("i386-sequent-ptx2\n"); exit (0);
}
if (strncmp(un.version, "V1", 2) == 0) { /* XXX is V1 correct? */
printf ("i386-sequent-ptx1\n"); exit (0);
}
printf ("i386-sequent-ptx\n"); exit (0);
#endif
#if defined (vax)
# if !defined (ultrix)
# include <sys/param.h>
# if defined (BSD)
# if BSD == 43
printf ("vax-dec-bsd4.3\n"); exit (0);
# else
# if BSD == 199006
printf ("vax-dec-bsd4.3reno\n"); exit (0);
# else
printf ("vax-dec-bsd\n"); exit (0);
# endif
# endif
# else
printf ("vax-dec-bsd\n"); exit (0);
# endif
# else
printf ("vax-dec-ultrix\n"); exit (0);
# endif
#endif
#if defined (alliant) && defined (i860)
printf ("i860-alliant-bsd\n"); exit (0);
#endif
exit (1);
}
EOF
$CC_FOR_BUILD -o $dummy $dummy.c 2>/dev/null && SYSTEM_NAME=`$dummy` &&
{ echo "$SYSTEM_NAME"; exit; }
# Apollos put the system type in the environment.
test -d /usr/apollo && { echo ${ISP}-apollo-${SYSTYPE}; exit; }
# Convex versions that predate uname can use getsysinfo(1)
if [ -x /usr/convex/getsysinfo ]
then
case `getsysinfo -f cpu_type` in
c1*)
echo c1-convex-bsd
exit ;;
c2*)
if getsysinfo -f scalar_acc
then echo c32-convex-bsd
else echo c2-convex-bsd
fi
exit ;;
c34*)
echo c34-convex-bsd
exit ;;
c38*)
echo c38-convex-bsd
exit ;;
c4*)
echo c4-convex-bsd
exit ;;
esac
fi
cat >&2 <<EOF
$0: unable to guess system type

5
Frameworks/Opus/Opus/opus/config.h
View file

@ -41,4 +41,9 @@ POSSIBILITY OF SUCH DAMAGE.
#define __MACOSX__
#define OPUS_HAVE_RTCD 1
#define OPUS_X86_MAY_HAVE_SSE 1
#define OPUS_X86_MAY_HAVE_SSE2 1
#define OPUS_X86_MAY_HAVE_SSE4_1 1
#endif CONFIG_H

88
Frameworks/Opus/Opus/opus/config.h.in
View file

@ -1,8 +1,17 @@
/* config.h.in. Generated from configure.ac by autoheader. */
/* Get CPU Info by asm method */
#undef CPU_INFO_BY_ASM
/* Get CPU Info by c method */
#undef CPU_INFO_BY_C
/* Custom modes */
#undef CUSTOM_MODES
/* Do not build the float API */
#undef DISABLE_FLOAT_API
/* Assertions */
#undef ENABLE_ASSERTIONS
@ -21,12 +30,12 @@
/* Define to 1 if you have the <alloca.h> header file. */
#undef HAVE_ALLOCA_H
/* NE10 library is installed on host. Make sure it is on target! */
#undef HAVE_ARM_NE10
/* Define to 1 if you have the <dlfcn.h> header file. */
#undef HAVE_DLFCN_H
/* Define to 1 if you have the <getopt.h> header file. */
#undef HAVE_GETOPT_H
/* Define to 1 if you have the <inttypes.h> header file. */
#undef HAVE_INTTYPES_H
@ -67,12 +76,69 @@
*/
#undef LT_OBJDIR
/* Define to 1 if your C compiler doesn't accept -c and -o together. */
#undef NO_MINUS_C_MINUS_O
/* Make use of ARM asm optimization */
#undef OPUS_ARM_ASM
/* Use generic ARMv4 inline asm optimizations */
#undef OPUS_ARM_INLINE_ASM
/* Use ARMv5E inline asm optimizations */
#undef OPUS_ARM_INLINE_EDSP
/* Use ARMv6 inline asm optimizations */
#undef OPUS_ARM_INLINE_MEDIA
/* Use ARM NEON inline asm optimizations */
#undef OPUS_ARM_INLINE_NEON
/* Define if assembler supports EDSP instructions */
#undef OPUS_ARM_MAY_HAVE_EDSP
/* Define if assembler supports ARMv6 media instructions */
#undef OPUS_ARM_MAY_HAVE_MEDIA
/* Define if compiler supports NEON instructions */
#undef OPUS_ARM_MAY_HAVE_NEON
/* Compiler supports ARMv7 Neon Intrinsics */
#undef OPUS_ARM_MAY_HAVE_NEON_INTR
/* Define if binary requires EDSP instruction support */
#undef OPUS_ARM_PRESUME_EDSP
/* Define if binary requires ARMv6 media instruction support */
#undef OPUS_ARM_PRESUME_MEDIA
/* Define if binary requires NEON instruction support */
#undef OPUS_ARM_PRESUME_NEON
/* Define if binary requires NEON intrinsics support */
#undef OPUS_ARM_PRESUME_NEON_INTR
/* This is a build of OPUS */
#undef OPUS_BUILD
/* Use run-time CPU capabilities detection */
#undef OPUS_HAVE_RTCD
/* Compiler supports X86 SSE Intrinsics */
#undef OPUS_X86_MAY_HAVE_SSE
/* Compiler supports X86 SSE2 Intrinsics */
#undef OPUS_X86_MAY_HAVE_SSE2
/* Compiler supports X86 SSE4.1 Intrinsics */
#undef OPUS_X86_MAY_HAVE_SSE4_1
/* Define if binary requires SSE intrinsics support */
#undef OPUS_X86_PRESUME_SSE
/* Define if binary requires SSE2 intrinsics support */
#undef OPUS_X86_PRESUME_SSE2
/* Define if binary requires SSE4.1 intrinsics support */
#undef OPUS_X86_PRESUME_SSE4_1
/* Define to the address where bug reports for this package should be sent. */
#undef PACKAGE_BUGREPORT
@ -91,18 +157,6 @@
/* Define to the version of this package. */
#undef PACKAGE_VERSION
/* The size of `int', as computed by sizeof. */
#undef SIZEOF_INT
/* The size of `long', as computed by sizeof. */
#undef SIZEOF_LONG
/* The size of `long long', as computed by sizeof. */
#undef SIZEOF_LONG_LONG
/* The size of `short', as computed by sizeof. */
#undef SIZEOF_SHORT
/* Define to 1 if you have the ANSI C header files. */
#undef STDC_HEADERS

132
Frameworks/Opus/Opus/opus/config.sub vendored
View file

@ -1,24 +1,18 @@
#! /bin/sh
# Configuration validation subroutine script.
# Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
# 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010,
# 2011, 2012 Free Software Foundation, Inc.
# Copyright 1992-2015 Free Software Foundation, Inc.
timestamp='2012-02-10'
timestamp='2015-01-01'
# This file is (in principle) common to ALL GNU software.
# The presence of a machine in this file suggests that SOME GNU software
# can handle that machine. It does not imply ALL GNU software can.
#
# This file is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# This file is free software; you can redistribute it and/or modify it
# under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
# General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, see <http://www.gnu.org/licenses/>.
@ -26,11 +20,12 @@ timestamp='2012-02-10'
# As a special exception to the GNU General Public License, if you
# distribute this file as part of a program that contains a
# configuration script generated by Autoconf, you may include it under
# the same distribution terms that you use for the rest of that program.
# the same distribution terms that you use for the rest of that
# program. This Exception is an additional permission under section 7
# of the GNU General Public License, version 3 ("GPLv3").
# Please send patches to <config-patches@gnu.org>. Submit a context
# diff and a properly formatted GNU ChangeLog entry.
# Please send patches to <config-patches@gnu.org>.
#
# Configuration subroutine to validate and canonicalize a configuration type.
# Supply the specified configuration type as an argument.
@ -73,9 +68,7 @@ Report bugs and patches to <config-patches@gnu.org>."
version="\
GNU config.sub ($timestamp)
Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012
Free Software Foundation, Inc.
Copyright 1992-2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
@ -123,7 +116,7 @@ esac
maybe_os=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\2/'`
case $maybe_os in
nto-qnx* | linux-gnu* | linux-android* | linux-dietlibc | linux-newlib* | \
linux-uclibc* | uclinux-uclibc* | uclinux-gnu* | kfreebsd*-gnu* | \
linux-musl* | linux-uclibc* | uclinux-uclibc* | uclinux-gnu* | kfreebsd*-gnu* | \
knetbsd*-gnu* | netbsd*-gnu* | \
kopensolaris*-gnu* | \
storm-chaos* | os2-emx* | rtmk-nova*)
@ -156,7 +149,7 @@ case $os in
-convergent* | -ncr* | -news | -32* | -3600* | -3100* | -hitachi* |\
-c[123]* | -convex* | -sun | -crds | -omron* | -dg | -ultra | -tti* | \
-harris | -dolphin | -highlevel | -gould | -cbm | -ns | -masscomp | \
-apple | -axis | -knuth | -cray | -microblaze)
-apple | -axis | -knuth | -cray | -microblaze*)
os=
basic_machine=$1
;;
@ -225,6 +218,12 @@ case $os in
-isc*)
basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
;;
-lynx*178)
os=-lynxos178
;;
-lynx*5)
os=-lynxos5
;;
-lynx*)
os=-lynxos
;;
@ -253,21 +252,24 @@ case $basic_machine in
| alpha | alphaev[4-8] | alphaev56 | alphaev6[78] | alphapca5[67] \
| alpha64 | alpha64ev[4-8] | alpha64ev56 | alpha64ev6[78] | alpha64pca5[67] \
| am33_2.0 \
| arc | arm | arm[bl]e | arme[lb] | armv[2345] | armv[345][lb] | avr | avr32 \
| arc | arceb \
| arm | arm[bl]e | arme[lb] | armv[2-8] | armv[3-8][lb] | armv7[arm] \
| avr | avr32 \
| be32 | be64 \
| bfin \
| c4x | clipper \
| c4x | c8051 | clipper \
| d10v | d30v | dlx | dsp16xx \
| epiphany \
| fido | fr30 | frv \
| fido | fr30 | frv | ft32 \
| h8300 | h8500 | hppa | hppa1.[01] | hppa2.0 | hppa2.0[nw] | hppa64 \
| hexagon \
| i370 | i860 | i960 | ia64 \
| ip2k | iq2000 \
| k1om \
| le32 | le64 \
| lm32 \
| m32c | m32r | m32rle | m68000 | m68k | m88k \
| maxq | mb | microblaze | mcore | mep | metag \
| maxq | mb | microblaze | microblazeel | mcore | mep | metag \
| mips | mipsbe | mipseb | mipsel | mipsle \
| mips16 \
| mips64 | mips64el \
@ -281,23 +283,26 @@ case $basic_machine in
| mips64vr5900 | mips64vr5900el \
| mipsisa32 | mipsisa32el \
| mipsisa32r2 | mipsisa32r2el \
| mipsisa32r6 | mipsisa32r6el \
| mipsisa64 | mipsisa64el \
| mipsisa64r2 | mipsisa64r2el \
| mipsisa64r6 | mipsisa64r6el \
| mipsisa64sb1 | mipsisa64sb1el \
| mipsisa64sr71k | mipsisa64sr71kel \
| mipsr5900 | mipsr5900el \
| mipstx39 | mipstx39el \
| mn10200 | mn10300 \
| moxie \
| mt \
| msp430 \
| nds32 | nds32le | nds32be \
| nios | nios2 \
| nios | nios2 | nios2eb | nios2el \
| ns16k | ns32k \
| open8 \
| or32 \
| open8 | or1k | or1knd | or32 \
| pdp10 | pdp11 | pj | pjl \
| powerpc | powerpc64 | powerpc64le | powerpcle \
| pyramid \
| riscv32 | riscv64 \
| rl78 | rx \
| score \
| sh | sh[1234] | sh[24]a | sh[24]aeb | sh[23]e | sh[34]eb | sheb | shbe | shle | sh[1234]le | sh3ele \
@ -308,6 +313,7 @@ case $basic_machine in
| tahoe | tic4x | tic54x | tic55x | tic6x | tic80 | tron \
| ubicom32 \
| v850 | v850e | v850e1 | v850e2 | v850es | v850e2v3 \
| visium \
| we32k \
| x86 | xc16x | xstormy16 | xtensa \
| z8k | z80)
@ -322,7 +328,10 @@ case $basic_machine in
c6x)
basic_machine=tic6x-unknown
;;
m6811 | m68hc11 | m6812 | m68hc12 | m68hcs12x | picochip)
leon|leon[3-9])
basic_machine=sparc-$basic_machine
;;
m6811 | m68hc11 | m6812 | m68hc12 | m68hcs12x | nvptx | picochip)
basic_machine=$basic_machine-unknown
os=-none
;;
@ -364,13 +373,13 @@ case $basic_machine in
| aarch64-* | aarch64_be-* \
| alpha-* | alphaev[4-8]-* | alphaev56-* | alphaev6[78]-* \
| alpha64-* | alpha64ev[4-8]-* | alpha64ev56-* | alpha64ev6[78]-* \
| alphapca5[67]-* | alpha64pca5[67]-* | arc-* \
| alphapca5[67]-* | alpha64pca5[67]-* | arc-* | arceb-* \
| arm-* | armbe-* | armle-* | armeb-* | armv*-* \
| avr-* | avr32-* \
| be32-* | be64-* \
| bfin-* | bs2000-* \
| c[123]* | c30-* | [cjt]90-* | c4x-* \
| clipper-* | craynv-* | cydra-* \
| c8051-* | clipper-* | craynv-* | cydra-* \
| d10v-* | d30v-* | dlx-* \
| elxsi-* \
| f30[01]-* | f700-* | fido-* | fr30-* | frv-* | fx80-* \
@ -379,11 +388,13 @@ case $basic_machine in
| hexagon-* \
| i*86-* | i860-* | i960-* | ia64-* \
| ip2k-* | iq2000-* \
| k1om-* \
| le32-* | le64-* \
| lm32-* \
| m32c-* | m32r-* | m32rle-* \
| m68000-* | m680[012346]0-* | m68360-* | m683?2-* | m68k-* \
| m88110-* | m88k-* | maxq-* | mcore-* | metag-* | microblaze-* \
| m88110-* | m88k-* | maxq-* | mcore-* | metag-* \
| microblaze-* | microblazeel-* \
| mips-* | mipsbe-* | mipseb-* | mipsel-* | mipsle-* \
| mips16-* \
| mips64-* | mips64el-* \
@ -397,18 +408,22 @@ case $basic_machine in
| mips64vr5900-* | mips64vr5900el-* \
| mipsisa32-* | mipsisa32el-* \
| mipsisa32r2-* | mipsisa32r2el-* \
| mipsisa32r6-* | mipsisa32r6el-* \
| mipsisa64-* | mipsisa64el-* \
| mipsisa64r2-* | mipsisa64r2el-* \
| mipsisa64r6-* | mipsisa64r6el-* \
| mipsisa64sb1-* | mipsisa64sb1el-* \
| mipsisa64sr71k-* | mipsisa64sr71kel-* \
| mipsr5900-* | mipsr5900el-* \
| mipstx39-* | mipstx39el-* \
| mmix-* \
| mt-* \
| msp430-* \
| nds32-* | nds32le-* | nds32be-* \
| nios-* | nios2-* \
| nios-* | nios2-* | nios2eb-* | nios2el-* \
| none-* | np1-* | ns16k-* | ns32k-* \
| open8-* \
| or1k*-* \
| orion-* \
| pdp10-* | pdp11-* | pj-* | pjl-* | pn-* | power-* \
| powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* \
@ -426,6 +441,7 @@ case $basic_machine in
| ubicom32-* \
| v850-* | v850e-* | v850e1-* | v850es-* | v850e2-* | v850e2v3-* \
| vax-* \
| visium-* \
| we32k-* \
| x86-* | x86_64-* | xc16x-* | xps100-* \
| xstormy16-* | xtensa*-* \
@ -763,6 +779,9 @@ case $basic_machine in
basic_machine=m68k-isi
os=-sysv
;;
leon-*|leon[3-9]-*)
basic_machine=sparc-`echo $basic_machine | sed 's/-.*//'`
;;
m68knommu)
basic_machine=m68k-unknown
os=-linux
@ -782,11 +801,15 @@ case $basic_machine in
basic_machine=ns32k-utek
os=-sysv
;;
microblaze)
microblaze*)
basic_machine=microblaze-xilinx
;;
mingw64)
basic_machine=x86_64-pc
os=-mingw64
;;
mingw32)
basic_machine=i386-pc
basic_machine=i686-pc
os=-mingw32
;;
mingw32ce)
@ -814,6 +837,10 @@ case $basic_machine in
basic_machine=powerpc-unknown
os=-morphos
;;
moxiebox)
basic_machine=moxie-unknown
os=-moxiebox
;;
msdos)
basic_machine=i386-pc
os=-msdos
@ -822,7 +849,7 @@ case $basic_machine in
basic_machine=`echo $basic_machine | sed -e 's/ms1-/mt-/'`
;;
msys)
basic_machine=i386-pc
basic_machine=i686-pc
os=-msys
;;
mvs)
@ -998,7 +1025,7 @@ case $basic_machine in
;;
ppc64) basic_machine=powerpc64-unknown
;;
ppc64-*) basic_machine=powerpc64-`echo $basic_machine | sed 's/^[^-]*-//'`
ppc64-* | ppc64p7-*) basic_machine=powerpc64-`echo $basic_machine | sed 's/^[^-]*-//'`
;;
ppc64le | powerpc64little | ppc64-le | powerpc64-little)
basic_machine=powerpc64le-unknown
@ -1013,7 +1040,11 @@ case $basic_machine in
basic_machine=i586-unknown
os=-pw32
;;
rdos)
rdos | rdos64)
basic_machine=x86_64-pc
os=-rdos
;;
rdos32)
basic_machine=i386-pc
os=-rdos
;;
@ -1340,29 +1371,29 @@ case $os in
-gnu* | -bsd* | -mach* | -minix* | -genix* | -ultrix* | -irix* \
| -*vms* | -sco* | -esix* | -isc* | -aix* | -cnk* | -sunos | -sunos[34]*\
| -hpux* | -unos* | -osf* | -luna* | -dgux* | -auroraux* | -solaris* \
| -sym* | -kopensolaris* \
| -sym* | -kopensolaris* | -plan9* \
| -amigaos* | -amigados* | -msdos* | -newsos* | -unicos* | -aof* \
| -aos* | -aros* \
| -nindy* | -vxsim* | -vxworks* | -ebmon* | -hms* | -mvs* \
| -clix* | -riscos* | -uniplus* | -iris* | -rtu* | -xenix* \
| -hiux* | -386bsd* | -knetbsd* | -mirbsd* | -netbsd* \
| -openbsd* | -solidbsd* \
| -bitrig* | -openbsd* | -solidbsd* \
| -ekkobsd* | -kfreebsd* | -freebsd* | -riscix* | -lynxos* \
| -bosx* | -nextstep* | -cxux* | -aout* | -elf* | -oabi* \
| -ptx* | -coff* | -ecoff* | -winnt* | -domain* | -vsta* \
| -udi* | -eabi* | -lites* | -ieee* | -go32* | -aux* \
| -chorusos* | -chorusrdb* | -cegcc* \
| -cygwin* | -msys* | -pe* | -psos* | -moss* | -proelf* | -rtems* \
| -mingw32* | -linux-gnu* | -linux-android* \
| -linux-newlib* | -linux-uclibc* \
| -uxpv* | -beos* | -mpeix* | -udk* \
| -mingw32* | -mingw64* | -linux-gnu* | -linux-android* \
| -linux-newlib* | -linux-musl* | -linux-uclibc* \
| -uxpv* | -beos* | -mpeix* | -udk* | -moxiebox* \
| -interix* | -uwin* | -mks* | -rhapsody* | -darwin* | -opened* \
| -openstep* | -oskit* | -conix* | -pw32* | -nonstopux* \
| -storm-chaos* | -tops10* | -tenex* | -tops20* | -its* \
| -os2* | -vos* | -palmos* | -uclinux* | -nucleus* \
| -morphos* | -superux* | -rtmk* | -rtmk-nova* | -windiss* \
| -powermax* | -dnix* | -nx6 | -nx7 | -sei* | -dragonfly* \
| -skyos* | -haiku* | -rdos* | -toppers* | -drops* | -es*)
| -skyos* | -haiku* | -rdos* | -toppers* | -drops* | -es* | -tirtos*)
# Remember, each alternative MUST END IN *, to match a version number.
;;
-qnx*)
@ -1486,9 +1517,6 @@ case $os in
-aros*)
os=-aros
;;
-kaos*)
os=-kaos
;;
-zvmoe)
os=-zvmoe
;;
@ -1537,6 +1565,12 @@ case $basic_machine in
c4x-* | tic4x-*)
os=-coff
;;
c8051-*)
os=-elf
;;
hexagon-*)
os=-elf
;;
tic54x-*)
os=-coff
;;

3109
Frameworks/Opus/Opus/opus/configure vendored
View file

File diff suppressed because it is too large Load diff

839
Frameworks/Opus/Opus/opus/configure.ac
View file

@ -4,7 +4,8 @@ dnl The package_version file will be automatically synced to the git revision
dnl by the update_version script when configured in the repository, but will
dnl remain constant in tarball releases unless it is manually edited.
m4_define([CURRENT_VERSION],
m4_esyscmd([ if test -e package_version || ./update_version; then
m4_esyscmd([ ./update_version 2>/dev/null || true
if test -e package_version; then
. ./package_version
printf "$PACKAGE_VERSION"
else
@ -14,16 +15,16 @@ m4_define([CURRENT_VERSION],
AC_INIT([opus],[CURRENT_VERSION],[opus@xiph.org])
AC_CONFIG_SRCDIR(src/opus_encoder.c)
AC_CONFIG_MACRO_DIR([m4])
dnl enable silent rules on automake 1.11 and later
m4_ifdef([AM_SILENT_RULES], [AM_SILENT_RULES([yes])])
# For libtool.
dnl Please update these for releases.
OPUS_LT_CURRENT=4
OPUS_LT_REVISION=0
OPUS_LT_AGE=4
OPUS_LT_CURRENT=5
OPUS_LT_REVISION=1
OPUS_LT_AGE=5
AC_SUBST(OPUS_LT_CURRENT)
AC_SUBST(OPUS_LT_REVISION)
@ -41,8 +42,14 @@ AC_PROG_CC_C99
AC_C_CONST
AC_C_INLINE
AM_PROG_AS
AC_DEFINE([OPUS_BUILD], [], [This is a build of OPUS])
#Use a hacked up version of autoconf's AC_C_RESTRICT because it's not
#strong enough a test to detect old buggy versions of GCC (e.g. 2.95.3)
#Note: Both this and the test for variable-size arrays below are also
# done by AC_PROG_CC_C99, but not thoroughly enough apparently.
AC_CACHE_CHECK([for C/C++ restrict keyword], ac_cv_c_restrict,
[ac_cv_c_restrict=no
# The order here caters to the fact that C++ does not require restrict.
@ -60,7 +67,8 @@ AC_CACHE_CHECK([for C/C++ restrict keyword], ac_cv_c_restrict,
test "$ac_cv_c_restrict" != no && break
done
])
AH_VERBATIM([restrict],
AH_VERBATIM([restrict],
[/* Define to the equivalent of the C99 'restrict' keyword, or to
nothing if this is not supported. Do not define if restrict is
supported directly. */
@ -74,43 +82,85 @@ AC_CACHE_CHECK([for C/C++ restrict keyword], ac_cv_c_restrict,
# define _Restrict
# define __restrict__
#endif])
case $ac_cv_c_restrict in
case $ac_cv_c_restrict in
restrict) ;;
no) AC_DEFINE([restrict], []) ;;
*) AC_DEFINE_UNQUOTED([restrict], [$ac_cv_c_restrict]) ;;
esac
AC_DEFINE([OPUS_BUILD], [], [This is a build of OPUS])
esac
AC_MSG_CHECKING(for C99 variable-size arrays)
AC_TRY_COMPILE( [], [static int x; char a[++x]; a[sizeof a - 1] = 0; int N; return a[0];],
[has_var_arrays=yes;AC_DEFINE([VAR_ARRAYS], [], [Use C99 variable-size arrays])
],
has_var_arrays=no
)
AC_MSG_RESULT($has_var_arrays)
AC_COMPILE_IFELSE([AC_LANG_PROGRAM([],
[[static int x; char a[++x]; a[sizeof a - 1] = 0; int N; return a[0];]])],
[ has_var_arrays=yes
use_alloca="no (using var arrays)"
AC_DEFINE([VAR_ARRAYS], [1], [Use C99 variable-size arrays])
],[
has_var_arrays=no
])
AC_MSG_RESULT([$has_var_arrays])
AC_CHECK_HEADERS([alloca.h getopt.h])
AC_MSG_CHECKING(for alloca)
AC_TRY_COMPILE( [#include <alloca.h>], [
int foo=10;
int *array = alloca(foo);
],
[
has_alloca=yes;
if test x$has_var_arrays = "xno" ; then
AC_DEFINE([USE_ALLOCA], [], [Make use of alloca])
fi
],
has_alloca=no
)
AC_MSG_RESULT($has_alloca)
AS_IF([test "$has_var_arrays" = "no"],
[
AC_CHECK_HEADERS([alloca.h])
AC_MSG_CHECKING(for alloca)
AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[#include <alloca.h>]],
[[int foo=10; int *array = alloca(foo);]])],
[ use_alloca=yes;
AC_DEFINE([USE_ALLOCA], [], [Make use of alloca])
],[
use_alloca=no
])
AC_MSG_RESULT([$use_alloca])
])
AC_CHECK_FUNC(exp,[fp_libm_not_needed=yes;LIBM=],[fp_libm_not_needed=dunno])
if test x"$fp_libm_not_needed" = xdunno; then
AC_CHECK_LIB([m], [exp], [LIBS="-lm $LIBS"; LIBM="-lm"],[LIBM=])
fi
AC_SUBST([LIBM])
LT_LIB_M
AC_ARG_ENABLE([fixed-point],
[AS_HELP_STRING([--enable-fixed-point],
[compile without floating point (for machines without a fast enough FPU)])],,
[enable_fixed_point=no])
AS_IF([test "$enable_fixed_point" = "yes"],[
enable_float="no"
AC_DEFINE([FIXED_POINT], [1], [Compile as fixed-point (for machines without a fast enough FPU)])
PC_BUILD="fixed-point"
],[
enable_float="yes";
PC_BUILD="floating-point"
])
AM_CONDITIONAL([FIXED_POINT], [test "$enable_fixed_point" = "yes"])
AC_ARG_ENABLE([fixed-point-debug],
[AS_HELP_STRING([--enable-fixed-point-debug], [debug fixed-point implementation])],,
[enable_fixed_point_debug=no])
AS_IF([test "$enable_fixed_point_debug" = "yes"],[
AC_DEFINE([FIXED_DEBUG], [1], [Debug fixed-point implementation])
])
AC_ARG_ENABLE([float_api],
[AS_HELP_STRING([--disable-float-api],
[compile without the floating point API (for machines with no float library)])],,
[enable_float_api=yes])
AM_CONDITIONAL([DISABLE_FLOAT_API], [test "$enable_float_api" = "no"])
AS_IF([test "$enable_float_api" = "no"],[
AC_DEFINE([DISABLE_FLOAT_API], [1], [Do not build the float API])
])
AC_ARG_ENABLE([custom-modes],
[AS_HELP_STRING([--enable-custom-modes], [enable non-Opus modes, e.g. 44.1 kHz & 2^n frames])],,
[enable_custom_modes=no])
AS_IF([test "$enable_custom_modes" = "yes"],[
AC_DEFINE([CUSTOM_MODES], [1], [Custom modes])
PC_BUILD="$PC_BUILD, custom modes"
])
AM_CONDITIONAL([CUSTOM_MODES], [test "$enable_custom_modes" = "yes"])
has_float_approx=no
#case "$host_cpu" in
@ -119,153 +169,600 @@ has_float_approx=no
# ;;
#esac
ac_enable_fixed="no";
ac_enable_float="yes";
AC_ARG_ENABLE(fixed-point, [ --enable-fixed-point compile without floating point (for machines without a fast enough FPU)],
[if test "$enableval" = yes; then
ac_enable_fixed="yes";
ac_enable_float="no";
AC_DEFINE([FIXED_POINT], [1], [Compile as fixed-point (for machines without a fast enough FPU)])
fi])
ac_enable_fixed_debug="no"
AC_ARG_ENABLE(fixed-point-debug, [ --enable-fixed-point-debug debug fixed-point implementation],
[if test "$enableval" = yes; then
ac_enable_fixed_debug="yes"
AC_DEFINE([FIXED_DEBUG], , [Debug fixed-point implementation])
fi])
ac_enable_custom_modes="no"
AC_ARG_ENABLE(custom-modes, [ --enable-custom-modes enable non-Opus modes, e.g. 44.1 kHz & 2^n frames],
[if test "$enableval" = yes; then
ac_enable_custom_modes="yes"
AC_DEFINE([CUSTOM_MODES], , [Custom modes])
fi])
float_approx=$has_float_approx
AC_ARG_ENABLE(float-approx, [ --enable-float-approx enable fast approximations for floating point],
[ if test "$enableval" = yes; then
AC_ARG_ENABLE([float-approx],
[AS_HELP_STRING([--enable-float-approx], [enable fast approximations for floating point])],
[if test "$enable_float_approx" = "yes"; then
AC_WARN([Floating point approximations are not supported on all platforms.])
float_approx=yes
else
float_approx=no
fi], [ float_approx=$has_float_approx ])
fi
],
[enable_float_approx=$has_float_approx])
if test "x${float_approx}" = "xyes"; then
AC_DEFINE([FLOAT_APPROX], , [Float approximations])
fi
AS_IF([test "$enable_float_approx" = "yes"],[
AC_DEFINE([FLOAT_APPROX], [1], [Float approximations])
])
ac_enable_assertions="no"
AC_ARG_ENABLE(assertions, [ --enable-assertions enable additional software error checking],
[if test "$enableval" = yes; then
ac_enable_assertions="yes"
AC_DEFINE([ENABLE_ASSERTIONS], , [Assertions])
fi])
AC_ARG_ENABLE([asm],
[AS_HELP_STRING([--disable-asm], [Disable assembly optimizations])],,
[enable_asm=yes])
ac_enable_fuzzing="no"
AC_ARG_ENABLE(fuzzing, [ --enable-fuzzing causes the encoder to make random decisions],
[if test "$enableval" = yes; then
ac_enable_fuzzing="yes"
AC_DEFINE([FUZZING], , [Fuzzing])
fi])
AC_ARG_ENABLE([rtcd],
[AS_HELP_STRING([--disable-rtcd], [Disable run-time CPU capabilities detection])],,
[enable_rtcd=yes])
AC_ARG_ENABLE([intrinsics],
[AS_HELP_STRING([--enable-intrinsics], [Enable intrinsics optimizations for ARM(float) X86(fixed)])],,
[enable_intrinsics=no])
rtcd_support=no
cpu_arm=no
AS_IF([test x"${enable_asm}" = x"yes"],[
inline_optimization="No inline ASM for your platform, please send patches"
case $host_cpu in
arm*)
dnl Currently we only have asm for fixed-point
AS_IF([test "$enable_float" != "yes"],[
cpu_arm=yes
AC_DEFINE([OPUS_ARM_ASM], [], [Make use of ARM asm optimization])
AS_GCC_INLINE_ASSEMBLY(
[inline_optimization="ARM"],
[inline_optimization="disabled"]
)
AS_ASM_ARM_EDSP([OPUS_ARM_INLINE_EDSP=1],[OPUS_ARM_INLINE_EDSP=0])
AS_ASM_ARM_MEDIA([OPUS_ARM_INLINE_MEDIA=1],
[OPUS_ARM_INLINE_MEDIA=0])
AS_ASM_ARM_NEON([OPUS_ARM_INLINE_NEON=1],[OPUS_ARM_INLINE_NEON=0])
AS_IF([test x"$inline_optimization" = x"ARM"],[
AM_CONDITIONAL([OPUS_ARM_INLINE_ASM],[true])
AC_DEFINE([OPUS_ARM_INLINE_ASM], 1,
[Use generic ARMv4 inline asm optimizations])
AS_IF([test x"$OPUS_ARM_INLINE_EDSP" = x"1"],[
AC_DEFINE([OPUS_ARM_INLINE_EDSP], [1],
[Use ARMv5E inline asm optimizations])
inline_optimization="$inline_optimization (EDSP)"
])
AS_IF([test x"$OPUS_ARM_INLINE_MEDIA" = x"1"],[
AC_DEFINE([OPUS_ARM_INLINE_MEDIA], [1],
[Use ARMv6 inline asm optimizations])
inline_optimization="$inline_optimization (Media)"
])
AS_IF([test x"$OPUS_ARM_INLINE_NEON" = x"1"],[
AC_DEFINE([OPUS_ARM_INLINE_NEON], 1,
[Use ARM NEON inline asm optimizations])
inline_optimization="$inline_optimization (NEON)"
])
])
dnl We need Perl to translate RVCT-syntax asm to gas syntax.
AC_CHECK_PROG([HAVE_PERL], perl, yes, no)
AS_IF([test x"$HAVE_PERL" = x"yes"],[
AM_CONDITIONAL([OPUS_ARM_EXTERNAL_ASM],[true])
asm_optimization="ARM"
AS_IF([test x"$OPUS_ARM_INLINE_EDSP" = x"1"], [
OPUS_ARM_PRESUME_EDSP=1
OPUS_ARM_MAY_HAVE_EDSP=1
],
[
OPUS_ARM_PRESUME_EDSP=0
OPUS_ARM_MAY_HAVE_EDSP=0
])
AS_IF([test x"$OPUS_ARM_INLINE_MEDIA" = x"1"], [
OPUS_ARM_PRESUME_MEDIA=1
OPUS_ARM_MAY_HAVE_MEDIA=1
],
[
OPUS_ARM_PRESUME_MEDIA=0
OPUS_ARM_MAY_HAVE_MEDIA=0
])
AS_IF([test x"$OPUS_ARM_INLINE_NEON" = x"1"], [
OPUS_ARM_PRESUME_NEON=1
OPUS_ARM_MAY_HAVE_NEON=1
],
[
OPUS_ARM_PRESUME_NEON=0
OPUS_ARM_MAY_HAVE_NEON=0
])
AS_IF([test x"$enable_rtcd" = x"yes"],[
AS_IF([test x"$OPUS_ARM_MAY_HAVE_EDSP" != x"1"],[
AC_MSG_NOTICE(
[Trying to force-enable armv5e EDSP instructions...])
AS_ASM_ARM_EDSP_FORCE([OPUS_ARM_MAY_HAVE_EDSP=1])
])
AS_IF([test x"$OPUS_ARM_MAY_HAVE_MEDIA" != x"1"],[
AC_MSG_NOTICE(
[Trying to force-enable ARMv6 media instructions...])
AS_ASM_ARM_MEDIA_FORCE([OPUS_ARM_MAY_HAVE_MEDIA=1])
])
AS_IF([test x"$OPUS_ARM_MAY_HAVE_NEON" != x"1"],[
AC_MSG_NOTICE(
[Trying to force-enable NEON instructions...])
AS_ASM_ARM_NEON_FORCE([OPUS_ARM_MAY_HAVE_NEON=1])
])
])
rtcd_support=
AS_IF([test x"$OPUS_ARM_MAY_HAVE_EDSP" = x"1"],[
AC_DEFINE(OPUS_ARM_MAY_HAVE_EDSP, 1,
[Define if assembler supports EDSP instructions])
AS_IF([test x"$OPUS_ARM_PRESUME_EDSP" = x"1"],[
AC_DEFINE(OPUS_ARM_PRESUME_EDSP, 1,
[Define if binary requires EDSP instruction support])
asm_optimization="$asm_optimization (EDSP)"
],
[rtcd_support="$rtcd_support (EDSP)"]
)
])
AC_SUBST(OPUS_ARM_MAY_HAVE_EDSP)
AS_IF([test x"$OPUS_ARM_MAY_HAVE_MEDIA" = x"1"],[
AC_DEFINE(OPUS_ARM_MAY_HAVE_MEDIA, 1,
[Define if assembler supports ARMv6 media instructions])
AS_IF([test x"$OPUS_ARM_PRESUME_MEDIA" = x"1"],[
AC_DEFINE(OPUS_ARM_PRESUME_MEDIA, 1,
[Define if binary requires ARMv6 media instruction support])
asm_optimization="$asm_optimization (Media)"
],
[rtcd_support="$rtcd_support (Media)"]
)
])
AC_SUBST(OPUS_ARM_MAY_HAVE_MEDIA)
AS_IF([test x"$OPUS_ARM_MAY_HAVE_NEON" = x"1"],[
AC_DEFINE(OPUS_ARM_MAY_HAVE_NEON, 1,
[Define if compiler supports NEON instructions])
AS_IF([test x"$OPUS_ARM_PRESUME_NEON" = x"1"], [
AC_DEFINE(OPUS_ARM_PRESUME_NEON, 1,
[Define if binary requires NEON instruction support])
asm_optimization="$asm_optimization (NEON)"
],
[rtcd_support="$rtcd_support (NEON)"]
)
])
AC_SUBST(OPUS_ARM_MAY_HAVE_NEON)
dnl Make sure turning on RTCD gets us at least one
dnl instruction set.
AS_IF([test x"$rtcd_support" != x""],
[rtcd_support=ARM"$rtcd_support"],
[rtcd_support="no"]
)
AC_MSG_CHECKING([for apple style tools])
AC_PREPROC_IFELSE([AC_LANG_PROGRAM([
#ifndef __APPLE__
#error 1
#endif],[])],
[AC_MSG_RESULT([yes]); ARM2GNU_PARAMS="--apple"],
[AC_MSG_RESULT([no]); ARM2GNU_PARAMS=""])
AC_SUBST(ARM2GNU_PARAMS)
],
[
AC_MSG_WARN(
[*** ARM assembly requires perl -- disabling optimizations])
asm_optimization="(missing perl dependency for ARM)"
])
])
;;
esac
],[
inline_optimization="disabled"
asm_optimization="disabled"
])
AM_CONDITIONAL([OPUS_ARM_INLINE_ASM],
[test x"${inline_optimization%% *}" = x"ARM"])
AM_CONDITIONAL([OPUS_ARM_EXTERNAL_ASM],
[test x"${asm_optimization%% *}" = x"ARM"])
AM_CONDITIONAL([HAVE_SSE], [false])
AM_CONDITIONAL([HAVE_SSE2], [false])
AM_CONDITIONAL([HAVE_SSE4_1], [false])
m4_define([DEFAULT_X86_SSE_CFLAGS], [-msse])
m4_define([DEFAULT_X86_SSE2_CFLAGS], [-msse2])
m4_define([DEFAULT_X86_SSE4_1_CFLAGS], [-msse4.1])
m4_define([DEFAULT_ARM_NEON_INTR_CFLAGS], [-mfpu=neon])
# With GCC on ARM32 softfp architectures (e.g. Android, or older Ubuntu) you need to specify
# -mfloat-abi=softfp for -mfpu=neon to work. However, on ARM32 hardfp architectures (e.g. newer Ubuntu),
# this option will break things.
# As a heuristic, if host matches arm*eabi* but not arm*hf*, it's probably soft-float.
m4_define([DEFAULT_ARM_NEON_SOFTFP_INTR_CFLAGS], [-mfpu=neon -mfloat-abi=softfp])
AS_CASE([$host],
[arm*hf*], [AS_VAR_SET([RESOLVED_DEFAULT_ARM_NEON_INTR_CFLAGS], "DEFAULT_ARM_NEON_INTR_CFLAGS")],
[arm*eabi*], [AS_VAR_SET([RESOLVED_DEFAULT_ARM_NEON_INTR_CFLAGS], "DEFAULT_ARM_NEON_SOFTFP_INTR_CFLAGS")],
[AS_VAR_SET([RESOLVED_DEFAULT_ARM_NEON_INTR_CFLAGS], "DEFAULT_ARM_NEON_INTR_CFLAGS")])
AC_ARG_VAR([X86_SSE_CFLAGS], [C compiler flags to compile SSE intrinsics @<:@default=]DEFAULT_X86_SSE_CFLAGS[@:>@])
AC_ARG_VAR([X86_SSE2_CFLAGS], [C compiler flags to compile SSE2 intrinsics @<:@default=]DEFAULT_X86_SSE2_CFLAGS[@:>@])
AC_ARG_VAR([X86_SSE4_1_CFLAGS], [C compiler flags to compile SSE4.1 intrinsics @<:@default=]DEFAULT_X86_SSE4_1_CFLAGS[@:>@])
AC_ARG_VAR([ARM_NEON_INTR_CFLAGS], [C compiler flags to compile ARM NEON intrinsics @<:@default=]DEFAULT_ARM_NEON_INTR_CFLAGS / DEFAULT_ARM_NEON_SOFTFP_INTR_CFLAGS[@:>@])
AS_VAR_SET_IF([X86_SSE_CFLAGS], [], [AS_VAR_SET([X86_SSE_CFLAGS], "DEFAULT_X86_SSE_CFLAGS")])
AS_VAR_SET_IF([X86_SSE2_CFLAGS], [], [AS_VAR_SET([X86_SSE2_CFLAGS], "DEFAULT_X86_SSE2_CFLAGS")])
AS_VAR_SET_IF([X86_SSE4_1_CFLAGS], [], [AS_VAR_SET([X86_SSE4_1_CFLAGS], "DEFAULT_X86_SSE4_1_CFLAGS")])
AS_VAR_SET_IF([ARM_NEON_INTR_CFLAGS], [], [AS_VAR_SET([ARM_NEON_INTR_CFLAGS], ["$RESOLVED_DEFAULT_ARM_NEON_INTR_CFLAGS"])])
AC_DEFUN([OPUS_PATH_NE10],
[
AC_ARG_WITH(NE10,
AC_HELP_STRING([--with-NE10=PFX],[Prefix where libNE10 is installed (optional)]),
NE10_prefix="$withval", NE10_prefix="")
AC_ARG_WITH(NE10-libraries,
AC_HELP_STRING([--with-NE10-libraries=DIR],
[Directory where libNE10 library is installed (optional)]),
NE10_libraries="$withval", NE10_libraries="")
AC_ARG_WITH(NE10-includes,
AC_HELP_STRING([--with-NE10-includes=DIR],
[Directory where libNE10 header files are installed (optional)]),
NE10_includes="$withval", NE10_includes="")
if test "x$NE10_libraries" != "x" ; then
NE10_LIBS="-L$NE10_libraries"
elif test "x$NE10_prefix" = "xno" || test "x$NE10_prefix" = "xyes" ; then
NE10_LIBS=""
elif test "x$NE10_prefix" != "x" ; then
NE10_LIBS="-L$NE10_prefix/lib"
elif test "x$prefix" != "xNONE" ; then
NE10_LIBS="-L$prefix/lib"
fi
if test "x$NE10_prefix" != "xno" ; then
NE10_LIBS="$NE10_LIBS -lNE10"
fi
if test "x$NE10_includes" != "x" ; then
NE10_CFLAGS="-I$NE10_includes"
elif test "x$NE10_prefix" = "xno" || test "x$NE10_prefix" = "xyes" ; then
NE10_CFLAGS=""
elif test "x$ogg_prefix" != "x" ; then
NE10_CFLAGS="-I$NE10_prefix/include"
elif test "x$prefix" != "xNONE"; then
NE10_CFLAGS="-I$prefix/include"
fi
AC_MSG_CHECKING(for NE10)
save_CFLAGS="$CFLAGS"; CFLAGS="$NE10_CFLAGS"
save_LIBS="$LIBS"; LIBS="$NE10_LIBS $LIBM"
AC_LINK_IFELSE(
[
AC_LANG_PROGRAM(
[[#include <NE10_init.h>
]],
[[
ne10_fft_cfg_float32_t cfg;
cfg = ne10_fft_alloc_c2c_float32_neon(480);
]]
)
],[
HAVE_ARM_NE10=1
AC_MSG_RESULT([yes])
],[
HAVE_ARM_NE10=0
AC_MSG_RESULT([no])
NE10_CFLAGS=""
NE10_LIBS=""
]
)
CFLAGS="$save_CFLAGS"; LIBS="$save_LIBS"
#Now we know if libNE10 is installed or not
AS_IF([test x"$HAVE_ARM_NE10" = x"1"],
[
AC_DEFINE([HAVE_ARM_NE10], 1, [NE10 library is installed on host. Make sure it is on target!])
AC_SUBST(HAVE_ARM_NE10)
AC_SUBST(NE10_CFLAGS)
AC_SUBST(NE10_LIBS)
]
)
]
)
AS_IF([test x"$enable_intrinsics" = x"yes"],[
intrinsics_support=""
AS_CASE([$host_cpu],
[arm*],
[
cpu_arm=yes
OPUS_CHECK_INTRINSICS(
[ARM Neon],
[$ARM_NEON_INTR_CFLAGS],
[OPUS_ARM_MAY_HAVE_NEON_INTR],
[OPUS_ARM_PRESUME_NEON_INTR],
[[#include <arm_neon.h>
]],
[[
static float32x4_t A0, A1, SUMM;
SUMM = vmlaq_f32(SUMM, A0, A1);
]]
)
AS_IF([test x"$OPUS_ARM_MAY_HAVE_NEON_INTR" = x"1" && test x"$OPUS_ARM_PRESUME_NEON_INTR" != x"1"],
[
OPUS_ARM_NEON_INTR_CFLAGS="$ARM_NEON_INTR_CFLAGS"
AC_SUBST([OPUS_ARM_NEON_INTR_CFLAGS])
]
)
AS_IF([test x"$OPUS_ARM_MAY_HAVE_NEON_INTR" = x"1"],
[
AC_DEFINE([OPUS_ARM_MAY_HAVE_NEON_INTR], 1, [Compiler supports ARMv7 Neon Intrinsics])
intrinsics_support="$intrinsics_support (Neon_Intrinsics)"
AS_IF([test x"enable_rtcd" != x"" && test x"$OPUS_ARM_PRESUME_NEON_INTR" != x"1"],
[rtcd_support="$rtcd_support (ARMv7_Neon_Intrinsics)"])
AS_IF([test x"$OPUS_ARM_PRESUME_NEON_INTR" = x"1"],
[AC_DEFINE([OPUS_ARM_PRESUME_NEON_INTR], 1, [Define if binary requires NEON intrinsics support])])
OPUS_PATH_NE10()
AS_IF([test x"$NE10_LIBS" != x""],
[
intrinsics_support="$intrinsics_support (NE10)"
AS_IF([test x"enable_rtcd" != x"" \
&& test x"$OPUS_ARM_PRESUME_NEON_INTR" != x"1"],
[rtcd_support="$rtcd_support (NE10)"])
])
AS_IF([test x"$rtcd_support" = x""],
[rtcd_support=no])
AS_IF([test x"$intrinsics_support" = x""],
[intrinsics_support=no],
[intrinsics_support="arm$intrinsics_support"])
],
[
AC_MSG_WARN([Compiler does not support ARM intrinsics])
intrinsics_support=no
])
],
[i?86|x86_64],
[
OPUS_CHECK_INTRINSICS(
[SSE],
[$X86_SSE_CFLAGS],
[OPUS_X86_MAY_HAVE_SSE],
[OPUS_X86_PRESUME_SSE],
[[#include <xmmintrin.h>
]],
[[
static __m128 mtest;
mtest = _mm_setzero_ps();
]]
)
AS_IF([test x"$OPUS_X86_MAY_HAVE_SSE" = x"1" && test x"$OPUS_X86_PRESUME_SSE" != x"1"],
[
OPUS_X86_SSE_CFLAGS="$X86_SSE_CFLAGS"
AC_SUBST([OPUS_X86_SSE_CFLAGS])
]
)
OPUS_CHECK_INTRINSICS(
[SSE2],
[$X86_SSE2_CFLAGS],
[OPUS_X86_MAY_HAVE_SSE2],
[OPUS_X86_PRESUME_SSE2],
[[#include <emmintrin.h>
]],
[[
static __m128i mtest;
mtest = _mm_setzero_si128();
]]
)
AS_IF([test x"$OPUS_X86_MAY_HAVE_SSE2" = x"1" && test x"$OPUS_X86_PRESUME_SSE2" != x"1"],
[
OPUS_X86_SSE2_CFLAGS="$X86_SSE2_CFLAGS"
AC_SUBST([OPUS_X86_SSE2_CFLAGS])
]
)
OPUS_CHECK_INTRINSICS(
[SSE4.1],
[$X86_SSE4_1_CFLAGS],
[OPUS_X86_MAY_HAVE_SSE4_1],
[OPUS_X86_PRESUME_SSE4_1],
[[#include <smmintrin.h>
]],
[[
static __m128i mtest;
mtest = _mm_setzero_si128();
mtest = _mm_cmpeq_epi64(mtest, mtest);
]]
)
AS_IF([test x"$OPUS_X86_MAY_HAVE_SSE4_1" = x"1" && test x"$OPUS_X86_PRESUME_SSE4_1" != x"1"],
[
OPUS_X86_SSE4_1_CFLAGS="$X86_SSE4_1_CFLAGS"
AC_SUBST([OPUS_X86_SSE4_1_CFLAGS])
]
)
AS_IF([test x"$rtcd_support" = x"no"], [rtcd_support=""])
AS_IF([test x"$OPUS_X86_MAY_HAVE_SSE" = x"1"],
[
AC_DEFINE([OPUS_X86_MAY_HAVE_SSE], 1, [Compiler supports X86 SSE Intrinsics])
intrinsics_support="$intrinsics_support SSE"
AS_IF([test x"$OPUS_X86_PRESUME_SSE" = x"1"],
[AC_DEFINE([OPUS_X86_PRESUME_SSE], 1, [Define if binary requires SSE intrinsics support])],
[rtcd_support="$rtcd_support SSE"])
],
[
AC_MSG_WARN([Compiler does not support SSE intrinsics])
])
AS_IF([test x"$OPUS_X86_MAY_HAVE_SSE2" = x"1"],
[
AC_DEFINE([OPUS_X86_MAY_HAVE_SSE2], 1, [Compiler supports X86 SSE2 Intrinsics])
intrinsics_support="$intrinsics_support SSE2"
AS_IF([test x"$OPUS_X86_PRESUME_SSE2" = x"1"],
[AC_DEFINE([OPUS_X86_PRESUME_SSE2], 1, [Define if binary requires SSE2 intrinsics support])],
[rtcd_support="$rtcd_support SSE2"])
],
[
AC_MSG_WARN([Compiler does not support SSE2 intrinsics])
])
AS_IF([test x"$OPUS_X86_MAY_HAVE_SSE4_1" = x"1"],
[
AC_DEFINE([OPUS_X86_MAY_HAVE_SSE4_1], 1, [Compiler supports X86 SSE4.1 Intrinsics])
intrinsics_support="$intrinsics_support SSE4.1"
AS_IF([test x"$OPUS_X86_PRESUME_SSE4_1" = x"1"],
[AC_DEFINE([OPUS_X86_PRESUME_SSE4_1], 1, [Define if binary requires SSE4.1 intrinsics support])],
[rtcd_support="$rtcd_support SSE4.1"])
],
[
AC_MSG_WARN([Compiler does not support SSE4.1 intrinsics])
])
AS_IF([test x"$intrinsics_support" = x""],
[intrinsics_support=no],
[intrinsics_support="x86$intrinsics_support"]
)
AS_IF([test x"$rtcd_support" = x""],
[rtcd_support=no],
[rtcd_support="x86$rtcd_support"],
)
AS_IF([test x"$enable_rtcd" = x"yes" && test x"$rtcd_support" != x""],[
get_cpuid_by_asm="no"
AC_MSG_CHECKING([How to get X86 CPU Info])
AC_LINK_IFELSE([AC_LANG_PROGRAM([[
#include <stdio.h>
]],[[
unsigned int CPUInfo0;
unsigned int CPUInfo1;
unsigned int CPUInfo2;
unsigned int CPUInfo3;
unsigned int InfoType;
__asm__ __volatile__ (
"cpuid":
"=a" (CPUInfo0),
"=b" (CPUInfo1),
"=c" (CPUInfo2),
"=d" (CPUInfo3) :
"a" (InfoType), "c" (0)
);
]])],
[get_cpuid_by_asm="yes"
AC_MSG_RESULT([Inline Assembly])
AC_DEFINE([CPU_INFO_BY_ASM], [1], [Get CPU Info by asm method])],
[AC_LINK_IFELSE([AC_LANG_PROGRAM([[
#include <cpuid.h>
]],[[
unsigned int CPUInfo0;
unsigned int CPUInfo1;
unsigned int CPUInfo2;
unsigned int CPUInfo3;
unsigned int InfoType;
__get_cpuid(InfoType, &CPUInfo0, &CPUInfo1, &CPUInfo2, &CPUInfo3);
]])],
[AC_MSG_RESULT([C method])
AC_DEFINE([CPU_INFO_BY_C], [1], [Get CPU Info by c method])],
[AC_MSG_ERROR([no supported Get CPU Info method, please disable intrinsics])])])])
],
[
AC_MSG_WARN([No intrinsics support for your architecture])
intrinsics_support="no"
])
],
[
intrinsics_support="no"
])
AM_CONDITIONAL([CPU_ARM], [test "$cpu_arm" = "yes"])
AM_CONDITIONAL([OPUS_ARM_NEON_INTR],
[test x"$OPUS_ARM_MAY_HAVE_NEON_INTR" = x"1"])
AM_CONDITIONAL([HAVE_ARM_NE10],
[test x"$HAVE_ARM_NE10" = x"1"])
AM_CONDITIONAL([HAVE_SSE],
[test x"$OPUS_X86_MAY_HAVE_SSE" = x"1"])
AM_CONDITIONAL([HAVE_SSE2],
[test x"$OPUS_X86_MAY_HAVE_SSE2" = x"1"])
AM_CONDITIONAL([HAVE_SSE4_1],
[test x"$OPUS_X86_MAY_HAVE_SSE4_1" = x"1"])
AS_IF([test x"$enable_rtcd" = x"yes"],[
AS_IF([test x"$rtcd_support" != x"no"],[
AC_DEFINE([OPUS_HAVE_RTCD], [1],
[Use run-time CPU capabilities detection])
OPUS_HAVE_RTCD=1
AC_SUBST(OPUS_HAVE_RTCD)
])
],[
rtcd_support="disabled"
])
AC_ARG_ENABLE([assertions],
[AS_HELP_STRING([--enable-assertions],[enable additional software error checking])],,
[enable_assertions=no])
AS_IF([test "$enable_assertions" = "yes"], [
AC_DEFINE([ENABLE_ASSERTIONS], [1], [Assertions])
])
AC_ARG_ENABLE([fuzzing],
[AS_HELP_STRING([--enable-fuzzing],[causes the encoder to make random decisions])],,
[enable_fuzzing=no])
AS_IF([test "$enable_fuzzing" = "yes"], [
AC_DEFINE([FUZZING], [1], [Fuzzing])
])
ac_enable_doc="yes"
AC_ARG_ENABLE([doc],
AS_HELP_STRING([--disable-doc], [Do not build API documentation]),
[ac_enable_doc=$enableval])
AC_CHECK_PROG(HAVE_DOXYGEN, [doxygen], [yes], [no])
if test "$HAVE_DOXYGEN" != "yes" -o "$ac_enable_doc" != "yes"; then
HAVE_DOXYGEN="false"
ac_enable_doc="no"
fi
AM_CONDITIONAL(HAVE_DOXYGEN, [test $HAVE_DOXYGEN = yes])
[AS_HELP_STRING([--disable-doc], [Do not build API documentation])],,
[enable_doc=yes])
AS_IF([test "$enable_doc" = "yes"], [
AC_CHECK_PROG(HAVE_DOXYGEN, [doxygen], [yes], [no])
],[
HAVE_DOXYGEN=no
])
AM_CONDITIONAL([HAVE_DOXYGEN], [test "$HAVE_DOXYGEN" = "yes"])
AC_ARG_ENABLE([extra-programs],
[AS_HELP_STRING([--disable-extra-programs], [Do not build extra programs (demo and tests)])],,
[enable_extra_programs=yes])
AM_CONDITIONAL([EXTRA_PROGRAMS], [test "$enable_extra_programs" = "yes"])
saved_CFLAGS="$CFLAGS"
CFLAGS="$CFLAGS -fvisibility=hidden"
AC_MSG_CHECKING([if ${CC} supports -fvisibility=hidden])
AC_COMPILE_IFELSE([AC_LANG_SOURCE([char foo;])],
[ AC_MSG_RESULT([yes])
SYMBOL_VISIBILITY="-fvisibility=hidden" ],
AC_MSG_RESULT([no]))
CFLAGS="$saved_CFLAGS $SYMBOL_VISIBILITY"
AC_SUBST(SYMBOL_VISIBILITY)
AC_COMPILE_IFELSE([AC_LANG_SOURCE([[char foo;]])],
[ AC_MSG_RESULT([yes]) ],
[ AC_MSG_RESULT([no])
CFLAGS="$saved_CFLAGS"
])
CFLAGS="$CFLAGS -W"
warn_CFLAGS="-Wall -Wextra -Wcast-align -Wnested-externs -Wshadow -Wstrict-prototypes"
saved_CFLAGS="$CFLAGS"
CFLAGS="$CFLAGS -Wall -Wextra -Wcast-align -Wnested-externs -Wshadow -Wstrict-prototypes"
AC_MSG_CHECKING([if ${CC} supports -Wall -Wextra -Wcast-align -Wnested-externs -Wshadow -Wstrict-prototypes])
AC_COMPILE_IFELSE([AC_LANG_SOURCE([char foo;])],
[ AC_MSG_RESULT([yes])
EXTRA_WARNS="-Wall -Wextra -Wcast-align -Wnested-externs -Wshadow -Wstrict-prototypes" ],
AC_MSG_RESULT([no]))
CFLAGS="$saved_CFLAGS $EXTRA_WARNS"
CFLAGS="$CFLAGS $warn_CFLAGS"
AC_MSG_CHECKING([if ${CC} supports ${warn_CFLAGS}])
AC_COMPILE_IFELSE([AC_LANG_SOURCE([[char foo;]])],
[ AC_MSG_RESULT([yes]) ],
[ AC_MSG_RESULT([no])
CFLAGS="$saved_CFLAGS"
])
saved_LIBS="$LIBS"
LIBS="$LIBS $LIBM"
AC_CHECK_FUNCS([lrintf])
AC_CHECK_FUNCS([lrint])
LIBS="$saved_LIBS"
AC_CHECK_FUNCS([__malloc_hook])
AC_CHECK_SIZEOF(short)
AC_CHECK_SIZEOF(int)
AC_CHECK_SIZEOF(long)
AC_CHECK_SIZEOF(long long)
if test x$has_char16 = "xyes" ; then
case 1 in
$ac_cv_sizeof_short) SIZE16="short";;
$ac_cv_sizeof_int) SIZE16="int";;
esac
else
case 2 in
$ac_cv_sizeof_short) SIZE16="short";;
$ac_cv_sizeof_int) SIZE16="int";;
esac
fi
if test x$has_char16 = "xyes" ; then
case 2 in
$ac_cv_sizeof_int) SIZE32="int";;
$ac_cv_sizeof_long) SIZE32="long";;
$ac_cv_sizeof_short) SIZE32="short";;
esac
else
case 4 in
$ac_cv_sizeof_int) SIZE32="int";;
$ac_cv_sizeof_long) SIZE32="long";;
$ac_cv_sizeof_short) SIZE32="short";;
esac
fi
AC_SUBST(SIZE16)
AC_SUBST(SIZE32)
AM_CONDITIONAL([FIXED_POINT], [test x$ac_enable_fixed = xyes])
AM_CONDITIONAL([CUSTOM_MODES], [test x$ac_enable_custom_modes = xyes])
dnl subsitutions for the pkg-config files
if test x$ac_enable_float = xyes; then
PC_BUILD="floating-point"
PC_LIBM=$LIBM
else
PC_BUILD="fixed-point"
PC_LIBM=
fi
dnl opus_custom requires libm as well
if test x$ac_enable_custom_modes = xyes; then
PC_BUILD="${PC_BUILD}, custom modes"
PC_LIBM=$LIBM
fi
AC_SUBST([PC_BUILD])
AC_SUBST([PC_LIBM])
AC_CONFIG_FILES([Makefile opus.pc opus-uninstalled.pc
doc/Makefile doc/Doxyfile])
AC_CONFIG_FILES([
Makefile
opus.pc
opus-uninstalled.pc
celt/arm/armopts.s
doc/Makefile
doc/Doxyfile
])
AC_CONFIG_HEADERS([config.h])
AC_OUTPUT
AC_MSG_RESULT([
AC_MSG_NOTICE([
------------------------------------------------------------------------
$PACKAGE_NAME $PACKAGE_VERSION: Automatic configuration OK.
@ -273,20 +770,26 @@ AC_MSG_RESULT([
C99 var arrays: ................ ${has_var_arrays}
C99 lrintf: .................... ${ac_cv_func_lrintf}
Alloca: ........................ ${has_alloca}
Use alloca: .................... ${use_alloca}
General configuration:
Floating point support: ........ ${ac_enable_float}
Fast float approximations: ..... ${float_approx}
Fixed point debugging: ......... ${ac_enable_fixed_debug}
Custom modes: .................. ${ac_enable_custom_modes}
Assertion checking: ............ ${ac_enable_assertions}
Fuzzing: ....................... ${ac_enable_fuzzing}
Floating point support: ........ ${enable_float}
Fast float approximations: ..... ${enable_float_approx}
Fixed point debugging: ......... ${enable_fixed_point_debug}
Inline Assembly Optimizations: . ${inline_optimization}
External Assembly Optimizations: ${asm_optimization}
Intrinsics Optimizations.......: ${intrinsics_support}
Run-time CPU detection: ........ ${rtcd_support}
Custom modes: .................. ${enable_custom_modes}
Assertion checking: ............ ${enable_assertions}
Fuzzing: ....................... ${enable_fuzzing}
API documentation: ............. ${ac_enable_doc}
API documentation: ............. ${enable_doc}
Extra programs: ................ ${enable_extra_programs}
------------------------------------------------------------------------
Type "make; make install" to compile and install
Type "make check" to run the test suite
])
echo "Type \"make; make install\" to compile and install";
echo "Type \"make check\" to run the test suite";

341
Frameworks/Opus/Opus/opus/depcomp
View file

@ -1,10 +1,9 @@
#! /bin/sh
# depcomp - compile a program generating dependencies as side-effects
scriptversion=2012-03-27.16; # UTC
scriptversion=2013-05-30.07; # UTC
# Copyright (C) 1999, 2000, 2003, 2004, 2005, 2006, 2007, 2009, 2010,
# 2011, 2012 Free Software Foundation, Inc.
# Copyright (C) 1999-2014 Free Software Foundation, Inc.
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
@ -57,11 +56,65 @@ EOF
;;
esac
# Get the directory component of the given path, and save it in the
# global variables '$dir'. Note that this directory component will
# be either empty or ending with a '/' character. This is deliberate.
set_dir_from ()
{
case $1 in
*/*) dir=`echo "$1" | sed -e 's|/[^/]*$|/|'`;;
*) dir=;;
esac
}
# Get the suffix-stripped basename of the given path, and save it the
# global variable '$base'.
set_base_from ()
{
base=`echo "$1" | sed -e 's|^.*/||' -e 's/\.[^.]*$//'`
}
# If no dependency file was actually created by the compiler invocation,
# we still have to create a dummy depfile, to avoid errors with the
# Makefile "include basename.Plo" scheme.
make_dummy_depfile ()
{
echo "#dummy" > "$depfile"
}
# Factor out some common post-processing of the generated depfile.
# Requires the auxiliary global variable '$tmpdepfile' to be set.
aix_post_process_depfile ()
{
# If the compiler actually managed to produce a dependency file,
# post-process it.
if test -f "$tmpdepfile"; then
# Each line is of the form 'foo.o: dependency.h'.
# Do two passes, one to just change these to
# $object: dependency.h
# and one to simply output
# dependency.h:
# which is needed to avoid the deleted-header problem.
{ sed -e "s,^.*\.[$lower]*:,$object:," < "$tmpdepfile"
sed -e "s,^.*\.[$lower]*:[$tab ]*,," -e 's,$,:,' < "$tmpdepfile"
} > "$depfile"
rm -f "$tmpdepfile"
else
make_dummy_depfile
fi
}
# A tabulation character.
tab=' '
# A newline character.
nl='
'
# Character ranges might be problematic outside the C locale.
# These definitions help.
upper=ABCDEFGHIJKLMNOPQRSTUVWXYZ
lower=abcdefghijklmnopqrstuvwxyz
digits=0123456789
alpha=${upper}${lower}
if test -z "$depmode" || test -z "$source" || test -z "$object"; then
echo "depcomp: Variables source, object and depmode must be set" 1>&2
@ -75,6 +128,9 @@ tmpdepfile=${tmpdepfile-`echo "$depfile" | sed 's/\.\([^.]*\)$/.T\1/'`}
rm -f "$tmpdepfile"
# Avoid interferences from the environment.
gccflag= dashmflag=
# Some modes work just like other modes, but use different flags. We
# parameterize here, but still list the modes in the big case below,
# to make depend.m4 easier to write. Note that we *cannot* use a case
@ -109,7 +165,7 @@ if test "$depmode" = msvc7msys; then
fi
if test "$depmode" = xlc; then
# IBM C/C++ Compilers xlc/xlC can output gcc-like dependency informations.
# IBM C/C++ Compilers xlc/xlC can output gcc-like dependency information.
gccflag=-qmakedep=gcc,-MF
depmode=gcc
fi
@ -134,8 +190,7 @@ gcc3)
done
"$@"
stat=$?
if test $stat -eq 0; then :
else
if test $stat -ne 0; then
rm -f "$tmpdepfile"
exit $stat
fi
@ -143,13 +198,17 @@ gcc3)
;;
gcc)
## Note that this doesn't just cater to obsosete pre-3.x GCC compilers.
## but also to in-use compilers like IMB xlc/xlC and the HP C compiler.
## (see the conditional assignment to $gccflag above).
## There are various ways to get dependency output from gcc. Here's
## why we pick this rather obscure method:
## - Don't want to use -MD because we'd like the dependencies to end
## up in a subdir. Having to rename by hand is ugly.
## (We might end up doing this anyway to support other compilers.)
## - The DEPENDENCIES_OUTPUT environment variable makes gcc act like
## -MM, not -M (despite what the docs say).
## -MM, not -M (despite what the docs say). Also, it might not be
## supported by the other compilers which use the 'gcc' depmode.
## - Using -M directly means running the compiler twice (even worse
## than renaming).
if test -z "$gccflag"; then
@ -157,15 +216,14 @@ gcc)
fi
"$@" -Wp,"$gccflag$tmpdepfile"
stat=$?
if test $stat -eq 0; then :
else
if test $stat -ne 0; then
rm -f "$tmpdepfile"
exit $stat
fi
rm -f "$depfile"
echo "$object : \\" > "$depfile"
alpha=ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz
## The second -e expression handles DOS-style file names with drive letters.
# The second -e expression handles DOS-style file names with drive
# letters.
sed -e 's/^[^:]*: / /' \
-e 's/^['$alpha']:\/[^:]*: / /' < "$tmpdepfile" >> "$depfile"
## This next piece of magic avoids the "deleted header file" problem.
@ -174,14 +232,14 @@ gcc)
## typically no way to rebuild the header). We avoid this by adding
## dummy dependencies for each header file. Too bad gcc doesn't do
## this for us directly.
tr ' ' "$nl" < "$tmpdepfile" |
## Some versions of gcc put a space before the ':'. On the theory
## that the space means something, we add a space to the output as
## well. hp depmode also adds that space, but also prefixes the VPATH
## to the object. Take care to not repeat it in the output.
## Some versions of the HPUX 10.20 sed can't process this invocation
## correctly. Breaking it into two sed invocations is a workaround.
sed -e 's/^\\$//' -e '/^$/d' -e "s|.*$object$||" -e '/:$/d' \
tr ' ' "$nl" < "$tmpdepfile" \
| sed -e 's/^\\$//' -e '/^$/d' -e "s|.*$object$||" -e '/:$/d' \
| sed -e 's/$/ :/' >> "$depfile"
rm -f "$tmpdepfile"
;;
@ -200,8 +258,7 @@ sgi)
"$@" -MDupdate "$tmpdepfile"
fi
stat=$?
if test $stat -eq 0; then :
else
if test $stat -ne 0; then
rm -f "$tmpdepfile"
exit $stat
fi
@ -209,7 +266,6 @@ sgi)
if test -f "$tmpdepfile"; then # yes, the sourcefile depend on other files
echo "$object : \\" > "$depfile"
# Clip off the initial element (the dependent). Don't try to be
# clever and replace this with sed code, as IRIX sed won't handle
# lines with more than a fixed number of characters (4096 in
@ -217,19 +273,15 @@ sgi)
# the IRIX cc adds comments like '#:fec' to the end of the
# dependency line.
tr ' ' "$nl" < "$tmpdepfile" \
| sed -e 's/^.*\.o://' -e 's/#.*$//' -e '/^$/ d' | \
tr "$nl" ' ' >> "$depfile"
| sed -e 's/^.*\.o://' -e 's/#.*$//' -e '/^$/ d' \
| tr "$nl" ' ' >> "$depfile"
echo >> "$depfile"
# The second pass generates a dummy entry for each header file.
tr ' ' "$nl" < "$tmpdepfile" \
| sed -e 's/^.*\.o://' -e 's/#.*$//' -e '/^$/ d' -e 's/$/:/' \
>> "$depfile"
else
# The sourcefile does not contain any dependencies, so just
# store a dummy comment line, to avoid errors with the Makefile
# "include basename.Plo" scheme.
echo "#dummy" > "$depfile"
make_dummy_depfile
fi
rm -f "$tmpdepfile"
;;
@ -247,9 +299,8 @@ aix)
# current directory. Also, the AIX compiler puts '$object:' at the
# start of each line; $object doesn't have directory information.
# Version 6 uses the directory in both cases.
dir=`echo "$object" | sed -e 's|/[^/]*$|/|'`
test "x$dir" = "x$object" && dir=
base=`echo "$object" | sed -e 's|^.*/||' -e 's/\.o$//' -e 's/\.lo$//'`
set_dir_from "$object"
set_base_from "$object"
if test "$libtool" = yes; then
tmpdepfile1=$dir$base.u
tmpdepfile2=$base.u
@ -262,9 +313,7 @@ aix)
"$@" -M
fi
stat=$?
if test $stat -eq 0; then :
else
if test $stat -ne 0; then
rm -f "$tmpdepfile1" "$tmpdepfile2" "$tmpdepfile3"
exit $stat
fi
@ -273,65 +322,113 @@ aix)
do
test -f "$tmpdepfile" && break
done
if test -f "$tmpdepfile"; then
# Each line is of the form 'foo.o: dependent.h'.
# Do two passes, one to just change these to
# '$object: dependent.h' and one to simply 'dependent.h:'.
sed -e "s,^.*\.[a-z]*:,$object:," < "$tmpdepfile" > "$depfile"
sed -e 's,^.*\.[a-z]*:['"$tab"' ]*,,' -e 's,$,:,' < "$tmpdepfile" >> "$depfile"
else
# The sourcefile does not contain any dependencies, so just
# store a dummy comment line, to avoid errors with the Makefile
# "include basename.Plo" scheme.
echo "#dummy" > "$depfile"
fi
rm -f "$tmpdepfile"
aix_post_process_depfile
;;
icc)
# Intel's C compiler anf tcc (Tiny C Compiler) understand '-MD -MF file'.
# However on
# $CC -MD -MF foo.d -c -o sub/foo.o sub/foo.c
# ICC 7.0 will fill foo.d with something like
# foo.o: sub/foo.c
# foo.o: sub/foo.h
# which is wrong. We want
# sub/foo.o: sub/foo.c
# sub/foo.o: sub/foo.h
# sub/foo.c:
# sub/foo.h:
# ICC 7.1 will output
# foo.o: sub/foo.c sub/foo.h
# and will wrap long lines using '\':
# foo.o: sub/foo.c ... \
# sub/foo.h ... \
# ...
# tcc 0.9.26 (FIXME still under development at the moment of writing)
# will emit a similar output, but also prepend the continuation lines
# with horizontal tabulation characters.
tcc)
# tcc (Tiny C Compiler) understand '-MD -MF file' since version 0.9.26
# FIXME: That version still under development at the moment of writing.
# Make that this statement remains true also for stable, released
# versions.
# It will wrap lines (doesn't matter whether long or short) with a
# trailing '\', as in:
#
# foo.o : \
# foo.c \
# foo.h \
#
# It will put a trailing '\' even on the last line, and will use leading
# spaces rather than leading tabs (at least since its commit 0394caf7
# "Emit spaces for -MD").
"$@" -MD -MF "$tmpdepfile"
stat=$?
if test $stat -eq 0; then :
else
if test $stat -ne 0; then
rm -f "$tmpdepfile"
exit $stat
fi
rm -f "$depfile"
# Each line is of the form 'foo.o: dependent.h',
# or 'foo.o: dep1.h dep2.h \', or ' dep3.h dep4.h \'.
# Each non-empty line is of the form 'foo.o : \' or ' dep.h \'.
# We have to change lines of the first kind to '$object: \'.
sed -e "s|.*:|$object :|" < "$tmpdepfile" > "$depfile"
# And for each line of the second kind, we have to emit a 'dep.h:'
# dummy dependency, to avoid the deleted-header problem.
sed -n -e 's|^ *\(.*\) *\\$|\1:|p' < "$tmpdepfile" >> "$depfile"
rm -f "$tmpdepfile"
;;
## The order of this option in the case statement is important, since the
## shell code in configure will try each of these formats in the order
## listed in this file. A plain '-MD' option would be understood by many
## compilers, so we must ensure this comes after the gcc and icc options.
pgcc)
# Portland's C compiler understands '-MD'.
# Will always output deps to 'file.d' where file is the root name of the
# source file under compilation, even if file resides in a subdirectory.
# The object file name does not affect the name of the '.d' file.
# pgcc 10.2 will output
# foo.o: sub/foo.c sub/foo.h
# and will wrap long lines using '\' :
# foo.o: sub/foo.c ... \
# sub/foo.h ... \
# ...
set_dir_from "$object"
# Use the source, not the object, to determine the base name, since
# that's sadly what pgcc will do too.
set_base_from "$source"
tmpdepfile=$base.d
# For projects that build the same source file twice into different object
# files, the pgcc approach of using the *source* file root name can cause
# problems in parallel builds. Use a locking strategy to avoid stomping on
# the same $tmpdepfile.
lockdir=$base.d-lock
trap "
echo '$0: caught signal, cleaning up...' >&2
rmdir '$lockdir'
exit 1
" 1 2 13 15
numtries=100
i=$numtries
while test $i -gt 0; do
# mkdir is a portable test-and-set.
if mkdir "$lockdir" 2>/dev/null; then
# This process acquired the lock.
"$@" -MD
stat=$?
# Release the lock.
rmdir "$lockdir"
break
else
# If the lock is being held by a different process, wait
# until the winning process is done or we timeout.
while test -d "$lockdir" && test $i -gt 0; do
sleep 1
i=`expr $i - 1`
done
fi
i=`expr $i - 1`
done
trap - 1 2 13 15
if test $i -le 0; then
echo "$0: failed to acquire lock after $numtries attempts" >&2
echo "$0: check lockdir '$lockdir'" >&2
exit 1
fi
if test $stat -ne 0; then
rm -f "$tmpdepfile"
exit $stat
fi
rm -f "$depfile"
# Each line is of the form `foo.o: dependent.h',
# or `foo.o: dep1.h dep2.h \', or ` dep3.h dep4.h \'.
# Do two passes, one to just change these to
# '$object: dependent.h' and one to simply 'dependent.h:'.
sed -e "s/^[ $tab][ $tab]*/ /" -e "s,^[^:]*:,$object :," \
< "$tmpdepfile" > "$depfile"
sed '
s/[ '"$tab"'][ '"$tab"']*/ /g
s/^ *//
s/ *\\*$//
s/^[^:]*: *//
/^$/d
/:$/d
s/$/ :/
' < "$tmpdepfile" >> "$depfile"
# `$object: dependent.h' and one to simply `dependent.h:'.
sed "s,^[^:]*:,$object :," < "$tmpdepfile" > "$depfile"
# Some versions of the HPUX 10.20 sed can't process this invocation
# correctly. Breaking it into two sed invocations is a workaround.
sed 's,^[^:]*: \(.*\)$,\1,;s/^\\$//;/^$/d;/:$/d' < "$tmpdepfile" \
| sed -e 's/$/ :/' >> "$depfile"
rm -f "$tmpdepfile"
;;
@ -342,9 +439,8 @@ hp2)
# 'foo.d', which lands next to the object file, wherever that
# happens to be.
# Much of this is similar to the tru64 case; see comments there.
dir=`echo "$object" | sed -e 's|/[^/]*$|/|'`
test "x$dir" = "x$object" && dir=
base=`echo "$object" | sed -e 's|^.*/||' -e 's/\.o$//' -e 's/\.lo$//'`
set_dir_from "$object"
set_base_from "$object"
if test "$libtool" = yes; then
tmpdepfile1=$dir$base.d
tmpdepfile2=$dir.libs/$base.d
@ -355,8 +451,7 @@ hp2)
"$@" +Maked
fi
stat=$?
if test $stat -eq 0; then :
else
if test $stat -ne 0; then
rm -f "$tmpdepfile1" "$tmpdepfile2"
exit $stat
fi
@ -366,7 +461,7 @@ hp2)
test -f "$tmpdepfile" && break
done
if test -f "$tmpdepfile"; then
sed -e "s,^.*\.[a-z]*:,$object:," "$tmpdepfile" > "$depfile"
sed -e "s,^.*\.[$lower]*:,$object:," "$tmpdepfile" > "$depfile"
# Add 'dependent.h:' lines.
sed -ne '2,${
s/^ *//
@ -375,7 +470,7 @@ hp2)
p
}' "$tmpdepfile" >> "$depfile"
else
echo "#dummy" > "$depfile"
make_dummy_depfile
fi
rm -f "$tmpdepfile" "$tmpdepfile2"
;;
@ -386,55 +481,40 @@ tru64)
# At least on Alpha/Redhat 6.1, Compaq CCC V6.2-504 seems to put
# dependencies in 'foo.d' instead, so we check for that too.
# Subdirectories are respected.
dir=`echo "$object" | sed -e 's|/[^/]*$|/|'`
test "x$dir" = "x$object" && dir=
base=`echo "$object" | sed -e 's|^.*/||' -e 's/\.o$//' -e 's/\.lo$//'`
set_dir_from "$object"
set_base_from "$object"
if test "$libtool" = yes; then
# With Tru64 cc, shared objects can also be used to make a
# static library. This mechanism is used in libtool 1.4 series to
# handle both shared and static libraries in a single compilation.
# With libtool 1.4, dependencies were output in $dir.libs/$base.lo.d.
#
# With libtool 1.5 this exception was removed, and libtool now
# generates 2 separate objects for the 2 libraries. These two
# compilations output dependencies in $dir.libs/$base.o.d and
# Libtool generates 2 separate objects for the 2 libraries. These
# two compilations output dependencies in $dir.libs/$base.o.d and
# in $dir$base.o.d. We have to check for both files, because
# one of the two compilations can be disabled. We should prefer
# $dir$base.o.d over $dir.libs/$base.o.d because the latter is
# automatically cleaned when .libs/ is deleted, while ignoring
# the former would cause a distcleancheck panic.
tmpdepfile1=$dir.libs/$base.lo.d # libtool 1.4
tmpdepfile2=$dir$base.o.d # libtool 1.5
tmpdepfile3=$dir.libs/$base.o.d # libtool 1.5
tmpdepfile4=$dir.libs/$base.d # Compaq CCC V6.2-504
tmpdepfile1=$dir$base.o.d # libtool 1.5
tmpdepfile2=$dir.libs/$base.o.d # Likewise.
tmpdepfile3=$dir.libs/$base.d # Compaq CCC V6.2-504
"$@" -Wc,-MD
else
tmpdepfile1=$dir$base.o.d
tmpdepfile1=$dir$base.d
tmpdepfile2=$dir$base.d
tmpdepfile3=$dir$base.d
tmpdepfile4=$dir$base.d
"$@" -MD
fi
stat=$?
if test $stat -eq 0; then :
else
rm -f "$tmpdepfile1" "$tmpdepfile2" "$tmpdepfile3" "$tmpdepfile4"
if test $stat -ne 0; then
rm -f "$tmpdepfile1" "$tmpdepfile2" "$tmpdepfile3"
exit $stat
fi
for tmpdepfile in "$tmpdepfile1" "$tmpdepfile2" "$tmpdepfile3" "$tmpdepfile4"
for tmpdepfile in "$tmpdepfile1" "$tmpdepfile2" "$tmpdepfile3"
do
test -f "$tmpdepfile" && break
done
if test -f "$tmpdepfile"; then
sed -e "s,^.*\.[a-z]*:,$object:," < "$tmpdepfile" > "$depfile"
sed -e 's,^.*\.[a-z]*:['"$tab"' ]*,,' -e 's,$,:,' < "$tmpdepfile" >> "$depfile"
else
echo "#dummy" > "$depfile"
fi
rm -f "$tmpdepfile"
# Same post-processing that is required for AIX mode.
aix_post_process_depfile
;;
msvc7)
@ -446,8 +526,7 @@ msvc7)
"$@" $showIncludes > "$tmpdepfile"
stat=$?
grep -v '^Note: including file: ' "$tmpdepfile"
if test "$stat" = 0; then :
else
if test $stat -ne 0; then
rm -f "$tmpdepfile"
exit $stat
fi
@ -473,6 +552,7 @@ $ {
G
p
}' >> "$depfile"
echo >> "$depfile" # make sure the fragment doesn't end with a backslash
rm -f "$tmpdepfile"
;;
@ -524,13 +604,14 @@ dashmstdout)
# in the target name. This is to cope with DOS-style filenames:
# a dependency such as 'c:/foo/bar' could be seen as target 'c' otherwise.
"$@" $dashmflag |
sed 's:^['"$tab"' ]*[^:'"$tab"' ][^:][^:]*\:['"$tab"' ]*:'"$object"'\: :' > "$tmpdepfile"
sed "s|^[$tab ]*[^:$tab ][^:][^:]*:[$tab ]*|$object: |" > "$tmpdepfile"
rm -f "$depfile"
cat < "$tmpdepfile" > "$depfile"
tr ' ' "$nl" < "$tmpdepfile" | \
## Some versions of the HPUX 10.20 sed can't process this invocation
## correctly. Breaking it into two sed invocations is a workaround.
sed -e 's/^\\$//' -e '/^$/d' -e '/:$/d' | sed -e 's/$/ :/' >> "$depfile"
# Some versions of the HPUX 10.20 sed can't process this sed invocation
# correctly. Breaking it into two sed invocations is a workaround.
tr ' ' "$nl" < "$tmpdepfile" \
| sed -e 's/^\\$//' -e '/^$/d' -e '/:$/d' \
| sed -e 's/$/ :/' >> "$depfile"
rm -f "$tmpdepfile"
;;
@ -583,10 +664,12 @@ makedepend)
# makedepend may prepend the VPATH from the source file name to the object.
# No need to regex-escape $object, excess matching of '.' is harmless.
sed "s|^.*\($object *:\)|\1|" "$tmpdepfile" > "$depfile"
sed '1,2d' "$tmpdepfile" | tr ' ' "$nl" | \
## Some versions of the HPUX 10.20 sed can't process this invocation
## correctly. Breaking it into two sed invocations is a workaround.
sed -e 's/^\\$//' -e '/^$/d' -e '/:$/d' | sed -e 's/$/ :/' >> "$depfile"
# Some versions of the HPUX 10.20 sed can't process the last invocation
# correctly. Breaking it into two sed invocations is a workaround.
sed '1,2d' "$tmpdepfile" \
| tr ' ' "$nl" \
| sed -e 's/^\\$//' -e '/^$/d' -e '/:$/d' \
| sed -e 's/$/ :/' >> "$depfile"
rm -f "$tmpdepfile" "$tmpdepfile".bak
;;
@ -622,10 +705,10 @@ cpp)
esac
done
"$@" -E |
sed -n -e '/^# [0-9][0-9]* "\([^"]*\)".*/ s:: \1 \\:p' \
-e '/^#line [0-9][0-9]* "\([^"]*\)".*/ s:: \1 \\:p' |
sed '$ s: \\$::' > "$tmpdepfile"
"$@" -E \
| sed -n -e '/^# [0-9][0-9]* "\([^"]*\)".*/ s:: \1 \\:p' \
-e '/^#line [0-9][0-9]* "\([^"]*\)".*/ s:: \1 \\:p' \
| sed '$ s: \\$::' > "$tmpdepfile"
rm -f "$depfile"
echo "$object : \\" > "$depfile"
cat < "$tmpdepfile" >> "$depfile"

2
Frameworks/Opus/Opus/opus/doc/Doxyfile.in
View file

@ -37,7 +37,7 @@ PROJECT_NUMBER = @VERSION@
# for a project that appears at the top of each page and should give viewer
# a quick idea about the purpose of the project. Keep the description short.
PROJECT_BRIEF = "IETF low-delay audio codec: API and operations manual"
PROJECT_BRIEF = "Opus audio codec (RFC 6716): API and operations manual"
# With the PROJECT_LOGO tag one can specify an logo or icon that is
# included in the documentation. The maximum height of the logo should not

Some files were not shown because too many files have changed in this diff Show more