ios Core audio: how to get samples from AudioBuffer with interleaved audio

Question

I have read an audio file into AudioBufferList with ExtAudioFileRead function.
This is and ASBD for the audio:

AudioStreamBasicDescription importFormat;

importFormat.mFormatID          = kAudioFormatLinearPCM;
importFormat.mFormatFlags       = kAudioFormatFlagIsSignedInteger | kAudioFormatFlagIsPacked;
importFormat.mBytesPerPacket    = 4;
importFormat.mFramesPerPacket   = 1;
importFormat.mBytesPerFrame     = 4;
importFormat.mChannelsPerFrame  = 2;
importFormat.mBitsPerChannel    = 16;
importFormat.mSampleRate = [[AVAudioSession sharedInstance] sampleRate];

So we got and interleaved audio with 2 channels with 16 bits signed int for each channel
AudioBufferList init:

UInt32 *audioData = (UInt32 *) calloc (totalFramesInFile, sizeof (UInt32));

AudioBufferList *bufferList;
bufferList = (AudioBufferList *) malloc (sizeof (AudioBufferList));

// buffers amount is 1 because audio is interleaved
bufferList->mNumberBuffers = 1;

bufferList->mBuffers[0].mNumberChannels  = 2;
bufferList->mBuffers[0].mDataByteSize    = totalFramesInFile * sizeof(UInt32);
bufferList->mBuffers[0].mData            = audioData;

And reading into buffer:

CheckError(ExtAudioFileRead (
                             audioFileObject,
                             &numberOfPacketsToRead,
                             bufferList), "error ExtAudioFileRead");

audioFileObject is and instance of ExtAudioFileRef which is initiated earlier in code which I did not paste here to save space.
What I am trying to accomplish is to modify audio samples in my render callback.

OSStatus MyCallback (void *inRefCon,
                 AudioUnitRenderActionFlags *ioActionFlags,
                 const AudioTimeStamp *inTimeStamp,
                 UInt32 inBusNumber,
                 UInt32 inNumberFrames,
                 AudioBufferList *ioData){


    ViewController *view = (__bridge ViewController *) inRefCon;

    soundStruct *soundStruct = (soundStruct *) &view->mys;

    SInt64            frameTotalForSound        = soundStruct->frameCount;

    soundStruct->isPlaying = true;

    UInt32 *audioData   = soundStruct->audioData;

    UInt32 sampleNumber = soundStruct->sampleNumber;

    for( int i = 0; i < ioData->mNumberBuffers; i++){

        AudioBuffer buffer = ioData->mBuffers[i];
        UInt32 *frameBuffer = buffer.mData;

        for(UInt32 frame = 0; frame < inNumberFrames; frame++) {

            // here I fill the buffer with my audio data.
            // i need to get left and right channel samples 
            // from  audioData[sampleNumber], modify them
            // and write into frameBuffer 

            frameBuffer[frame] = audioData[sampleNumber];

            sampleNumber++;

            if(sampleNumber > frameTotalForSound) {
                soundStruct->isPlaying = false;
                AudioOutputUnitStop(soundStruct->outputUnit);
            }
        }
    }

    soundStruct->sampleNumber = sampleNumber;

    return noErr;

}

Is it possible to get Sint16 left and right channel samples from UInt32 array of audio data?

Answer 1

@Rhythmic Fistman, Thanks a lot - it helped.
I could not set up frameBuffer to work that way though. The sound was distorted at the output.
I guess it's because the AudioUnit expects both channels data in one frame. Or maybe there is some another explanation.
Here is the code that I modified, hope it will help someone:

audioData init:

SInt16 *audioData = (SInt16 *) malloc (sizeof(SInt16) * totalFramesInFile * 2);

Modified render callback:

OSStatus MyCallback (void *inRefCon,
             AudioUnitRenderActionFlags *ioActionFlags,
             const AudioTimeStamp *inTimeStamp,
             UInt32 inBusNumber,
             UInt32 inNumberFrames,
             AudioBufferList *ioData)
{
    ViewController *view = (__bridge ViewController *) inRefCon;

    soundStruct *soundStruct  = (soundStruct *) &view->mys;

    SInt64 frameTotalForSound = soundStruct->frameCount;

    soundStruct->isPlaying = true;

    SInt16 *audioData   = soundStruct->audioData;

    UInt32 sampleNumber = soundStruct->sampleNumber;

    for( int i = 0; i < ioData->mNumberBuffers; i++){
        AudioBuffer buffer = ioData->mBuffers[i];
        SInt16 *frameBuffer = (SInt16*) ioData->mBuffers[0].mData;

        for(UInt32 frame = 0; frame < inNumberFrames * 2; frame+=2) {

            /* .. some samples modification code .. */

            // left channel
            frameBuffer[frame] = audioData[sampleNumber];
            // right channel
            frameBuffer[frame + 1] = audioData[sampleNumber + 1];

            sampleNumber +=2;

            if(sampleNumber > frameTotalForSound * 2) {
                soundStruct->isPlaying = false;
                AudioOutputUnitStop(soundStruct->outputUnit);
            }
        }
    }

    soundStruct->sampleNumber = sampleNumber;
    return noErr;
}

Answer 2

Let both audioData and frameBuffer be SInt16s:

SInt16 *audioData;
// ...
SInt16 *frameBuffer;

Your buffer size calculations should be n * 2 * sizeof(SInt16) and you'll either need to changesoundStruct` or add type casts.

Then you can access the interleaved samples like so:

frameBuffer[0] = modify(audioData[0]);    // left sample 1
frameBuffer[1] = modify(audioData[1]);    // right sample 1
frameBuffer[2] = modify(audioData[2]);    // left sample 2
frameBuffer[3] = modify(audioData[3]);    // right sample 2
// ...
frameBuffer[2*(n-1)] = modify(audioData[2*(n-1)]);    // left sample n
frameBuffer[2*(n-1)+1] = modify(audioData[2*(n-1)+1]); // right sample n