Cog Audio: Keep track of last resampler quality used so we don't reinitialize when not necessary, and also keep track of DSD decimated sample rate separately, so the input format remains unmodified

Audio/Chain/ConverterNode.h

@@ -62,8 +62,12 @@
     AudioStreamBasicDescription rememberedInputFormat;
     RefillNode *refillNode;
     id __weak originalPreviousNode;
+    
+    NSString *outputResampling;
 }
 
+@property AudioStreamBasicDescription inputFormat;
+
 - (id)initWithController:(id)c previous:(id)p;
 
 - (BOOL)setupWithInputFormat:(AudioStreamBasicDescription)inputFormat outputFormat:(AudioStreamBasicDescription)outputFormat;

Audio/Chain/ConverterNode.m

@@ -35,6 +35,8 @@ void PrintStreamDesc (AudioStreamBasicDescription *inDesc)
 
 @implementation ConverterNode
 
+@synthesize inputFormat;
+
 - (id)initWithController:(id)c previous:(id)p
 {
     self = [super initWithController:c previous:p];
@@ -69,6 +71,8 @@ void PrintStreamDesc (AudioStreamBasicDescription *inDesc)
         dsd2pcm = NULL;
         dsd2pcmCount = 0;
         
+        outputResampling = @"";
+
         [[NSUserDefaultsController sharedUserDefaultsController] addObserver:self forKeyPath:@"values.volumeScaling"		options:0 context:nil];
         [[NSUserDefaultsController sharedUserDefaultsController] addObserver:self forKeyPath:@"values.outputResampling" options:0 context:nil];
     }
@@ -1038,8 +1042,11 @@ tryagain:
     }
     else if ([keyPath isEqual:@"values.outputResampling"]) {
         // Reset resampler
-        if (resampler && resampler_data)
+        if (resampler && resampler_data) {
-            [self inputFormatDidChange:inputFormat];
+            NSString *value = [[NSUserDefaults standardUserDefaults] stringForKey:@"outputResampling"];
+            if (![value isEqualToString:outputResampling])
+                [self inputFormatDidChange:inputFormat];
+        }
     }
 }
 
@@ -1104,10 +1111,18 @@ static float db_to_scale(float db)
         || (isFloat && !(inputFormat.mBitsPerChannel == 32 || inputFormat.mBitsPerChannel == 64)))
         return NO;
     
+    // These are really placeholders, as we're doing everything internally now
+    
+    floatFormat = inputFormat;
+    floatFormat.mFormatFlags = kAudioFormatFlagsNativeFloatPacked;
+    floatFormat.mBitsPerChannel = 32;
+    floatFormat.mBytesPerFrame = (32/8)*floatFormat.mChannelsPerFrame;
+    floatFormat.mBytesPerPacket = floatFormat.mBytesPerFrame * floatFormat.mFramesPerPacket;
+    
     if (inputFormat.mBitsPerChannel == 1) {
         // Decimate this for speed
-        inputFormat.mSampleRate *= 1.0 / 8.0;
+        floatFormat.mSampleRate *= 1.0 / 8.0;
-        dsd2pcmCount = inputFormat.mChannelsPerFrame;
+        dsd2pcmCount = floatFormat.mChannelsPerFrame;
         dsd2pcm = calloc(dsd2pcmCount, sizeof(void*));
         dsd2pcm[0] = dsd2pcm_alloc();
         dsd2pcmLatency = dsd2pcm_latency(dsd2pcm[0]);
@@ -1117,14 +1132,6 @@ static float db_to_scale(float db)
         }
     }
     
-    // These are really placeholders, as we're doing everything internally now
-    
-    floatFormat = inputFormat;
-    floatFormat.mFormatFlags = kAudioFormatFlagsNativeFloatPacked;
-    floatFormat.mBitsPerChannel = 32;
-    floatFormat.mBytesPerFrame = (32/8)*floatFormat.mChannelsPerFrame;
-    floatFormat.mBytesPerPacket = floatFormat.mBytesPerFrame * floatFormat.mFramesPerPacket;
-    
     inpOffset = 0;
     inpSize = 0;
     
@@ -1142,9 +1149,9 @@ static float db_to_scale(float db)
     convert_s16_to_float_init_simd();
     convert_s32_to_float_init_simd();
     
-    skipResampler = outputFormat.mSampleRate == inputFormat.mSampleRate;
+    skipResampler = outputFormat.mSampleRate == floatFormat.mSampleRate;
     
-    sampleRatio = (double)outputFormat.mSampleRate / (double)inputFormat.mSampleRate;
+    sampleRatio = (double)outputFormat.mSampleRate / (double)floatFormat.mSampleRate;
     
     if (!skipResampler)
     {
@@ -1162,6 +1169,8 @@ static float db_to_scale(float db)
         else if ([resampling isEqualToString:@"highest"])
             quality = RESAMPLER_QUALITY_HIGHEST;
         
+        outputResampling = resampling;
+        
         if (!retro_resampler_realloc(&resampler_data, &resampler, "sinc", quality, inputFormat.mChannelsPerFrame, sampleRatio))
         {
             return NO;
@@ -1211,7 +1220,8 @@ static float db_to_scale(float db)
 	DLog(@"FORMAT CHANGED");
     paused = YES;
 	[self cleanUp];
-    if (outputFormatChanged && ![buffer isEmpty])
+    if (outputFormatChanged && ![buffer isEmpty] &&
+        memcmp(&outputFormat, &previousOutputFormat, sizeof(outputFormat)) != 0)
     {
         // Transfer previously buffered data, remember input format
         rememberedInputFormat = format;

Audio/Chain/InputNode.m

@@ -102,7 +102,10 @@
         DLog(@"Input format changed");
         // Converter doesn't need resetting for this, as output format hasn't changed
         ConverterNode *converter = [[[controller controller] bufferChain] converter];
-        [converter inputFormatDidChange:[[[controller controller] bufferChain] inputFormat]];
+        AudioStreamBasicDescription newInputFormat = [[[controller controller] bufferChain] inputFormat];
+        AudioStreamBasicDescription oldInputFormat = [converter inputFormat];
+        if (memcmp(&oldInputFormat, &newInputFormat, sizeof(oldInputFormat)) != 0)
+            [converter inputFormatDidChange:newInputFormat];
 	}
 	else if ([keyPath isEqual:@"metadata"]) {
 		//Inform something of metadata change