Fix for Flac to work with 24bit files (thanks to Play).

Updated COMPILE instructions to avoid confusion about scripts directory.

COMPILE

@@ -1,5 +1,5 @@
-To compile Cog, you must first compile all of the projects in the Frameworks and Preferences folders with the Release build configuration.
+To compile Cog, you must first compile all of the projects in the Frameworks, Plugins, and Preferences folders with the Release build configuration.
 
-To make this easier, I have created a bash script named build_dependencies.sh in the Scripts folder. Just run that in a terminal from the project root, and when it is finished, all the dependencies should be built.
+To make this easier, I have created a bash script named build_dependencies.sh in the Scripts folder. Just run that in a terminal from the project root (./Scripts/build_dependencies.sh). When it is finished, all the dependencies should be built.
 
 If you have any problems, email me at vspader@users.sourceforge.net

Plugins/Flac/FlacDecoder.h

@@ -18,7 +18,7 @@
 @interface FlacDecoder : NSObject <CogDecoder>
 {
 	FLAC__FileDecoder *decoder;
-	char buffer[SAMPLE_BUFFER_SIZE];
+	void *buffer;
 	int bufferAmount;
 	
 	int bitsPerSample;

Plugins/Flac/FlacDecoder.m

@@ -14,23 +14,65 @@
 FLAC__StreamDecoderWriteStatus WriteProc(const FLAC__FileDecoder *decoder, const FLAC__Frame *frame, const FLAC__int32 * const sampleBuffer[], void *client_data)
 {
 	FlacDecoder *flacDecoder = (FlacDecoder *)client_data;
-	unsigned wide_samples = frame->header.blocksize;
-	unsigned channels = frame->header.channels;
 	
-	UInt16 *buffer = (UInt16 *)[flacDecoder buffer];
+	void *buffer = [flacDecoder buffer];
-	int i, j, c;
 
-	for (i = j = 0; i < wide_samples; i++)
+	int8_t  *alias8;
-	{
+	int16_t *alias16;
-		for (c = 0; c < channels; c++, j++)
+	int32_t *alias32;
-		{
+	int sample, channel;
-//			buffer[j] = CFSwapInt16LittleToHost(sampleBuffer[c][i]);
+	int32_t	audioSample;
-			buffer[j] = sampleBuffer[c][i];
+
-		}
+    switch(frame->header.bits_per_sample) {
+        case 8:
+            // Interleave the audio (no need for byte swapping)
+            alias8 = buffer;
+            for(sample = 0; sample < frame->header.blocksize; ++sample) {
+                for(channel = 0; channel < frame->header.channels; ++channel) {
+                    *alias8++ = (int8_t)sampleBuffer[channel][sample];
+                }
+            }
+
+            break;
+
+        case 16:
+            // Interleave the audio, converting to big endian byte order
+            alias16 = buffer;
+            for(sample = 0; sample < frame->header.blocksize; ++sample) {
+                for(channel = 0; channel < frame->header.channels; ++channel) {
+                    *alias16++ = (int16_t)OSSwapHostToBigInt16((int16_t)sampleBuffer[channel][sample]);
+                }
+            }
+
+            break;
+
+        case 24:
+            // Interleave the audio (no need for byte swapping)
+            alias8 = buffer;
+            for(sample = 0; sample < frame->header.blocksize; ++sample) {
+                for(channel = 0; channel < frame->header.channels; ++channel) {
+                    audioSample = sampleBuffer[channel][sample];
+                    *alias8++   = (int8_t)(audioSample >> 16);
+                    *alias8++   = (int8_t)(audioSample >> 8);
+                    *alias8++   = (int8_t)audioSample;
+                }
+            }
+
+            break;
+
+        case 32:
+            // Interleave the audio, converting to big endian byte order
+            alias32 = buffer;
+            for(sample = 0; sample < frame->header.blocksize; ++sample) {
+                for(channel = 0; channel < frame->header.channels; ++channel) {
+                    *alias32++ = OSSwapHostToBigInt32(sampleBuffer[channel][sample]);
+                }
+            }
+		default:
+			NSLog(@"Error, unsupported sample size.");
 	}
-	
+
-//	memcpy([flacDecoder buffer], sampleBuffer, wide_samples*[flacDecoder bitsPerSample]/8);
+	[flacDecoder setBufferAmount:frame->header.blocksize * frame->header.channels * (frame->header.bits_per_sample/8)];
-	[flacDecoder setBufferAmount:(wide_samples * channels*2)];
 
 	return FLAC__STREAM_DECODER_WRITE_STATUS_CONTINUE;
 }
@@ -84,6 +126,8 @@ void ErrorProc(const FLAC__FileDecoder *decoder, FLAC__StreamDecoderErrorStatus
 
 	FLAC__file_decoder_process_until_end_of_metadata(decoder);
 
+	buffer = malloc(SAMPLE_BUFFER_SIZE);
+
 	return YES;
 }
 
@@ -117,7 +161,7 @@ void ErrorProc(const FLAC__FileDecoder *decoder, FLAC__StreamDecoderErrorStatus
 	bufferAmount -= count;
 	
 	if (bufferAmount > 0)
-		memmove(buffer, &buffer[count], bufferAmount);
+		memmove((char *)buffer, &((char *)buffer)[count], bufferAmount);
 	
 	if (count < size)
 		numread = [self fillBuffer:(&((char *)buf)[count]) ofSize:(size - count)];
@@ -135,8 +179,13 @@ void ErrorProc(const FLAC__FileDecoder *decoder, FLAC__StreamDecoderErrorStatus
 		FLAC__file_decoder_finish(decoder);
 		FLAC__file_decoder_delete(decoder);
 	}
+	if (buffer)
+	{
+		free(buffer);
+	}
+
 	decoder = NULL;
-	
+	buffer = NULL;
 }
 
 - (double)seekToTime:(double)milliseconds
@@ -172,7 +221,7 @@ void ErrorProc(const FLAC__FileDecoder *decoder, FLAC__StreamDecoderErrorStatus
 		[NSNumber numberWithInt:bitsPerSample],@"bitsPerSample",
 		[NSNumber numberWithFloat:frequency],@"sampleRate",
 		[NSNumber numberWithDouble:length],@"length",
-		@"host",@"endian",
+		@"big",@"endian",
 		nil];
 }