iOS Audio Unit - Creating Stereo Sine Waves

Question

Over the weekend I hit a stumbling block learning how to program audio synthesis on iOS. I have been developing on iOS for several years, but I am just getting into the audio synthesis aspect. Right now, I am just programming demo apps to help me learn the concepts. I have currently been able to build and stack sine waves in a playback renderer for Audio Units without a problem. But, I want to understand what is going on in the renderer so I can render 2 separate sine waves in each Left and Right Channel. Currently, I assume that in my init audio section I would need to make the following changes:

From:

AudioStreamBasicDescription audioFormat;
    audioFormat.mSampleRate = kSampleRate;
    audioFormat.mFormatID = kAudioFormatLinearPCM;
    audioFormat.mFormatFlags = kAudioFormatFlagIsSignedInteger | kAudioFormatFlagIsPacked;
    audioFormat.mFramesPerPacket = 1;
    audioFormat.mChannelsPerFrame = 1;
    audioFormat.mBitsPerChannel = 16;
    audioFormat.mBytesPerPacket = 2;
    audioFormat.mBytesPerFrame = 2;

To:

AudioStreamBasicDescription audioFormat;
    audioFormat.mSampleRate = kSampleRate;
    audioFormat.mFormatID = kAudioFormatLinearPCM;
    audioFormat.mFormatFlags = kAudioFormatFlagIsSignedInteger | kAudioFormatFlagIsPacked;
    audioFormat.mFramesPerPacket = 1;
    audioFormat.mChannelsPerFrame = 2;
    audioFormat.mBitsPerChannel = 16;
    audioFormat.mBytesPerPacket = 4;
    audioFormat.mBytesPerFrame = 4;

But, the renderer is somewhat greek to me. I have been working off of any tutorial or sample code I can find. I can make things work for the given context of a mono signal, but I cannot make the renderer generate stereo signals. All I want is one distinct frequency in a left channel and a different frequency in a right channel - but I honestly don't understand the renderer enough to get it working. I have attempted the memcpy function into mBuffers[0] and mbuffers[1], but that crashes the app. My render is below (it currently contains stacked sine waves, but for the stereo example I can just use one wave of a set frequency in each channel though).

#define kOutputBus 0
#define kSampleRate 44100
//44100.0f
#define kWaveform (M_PI * 2.0f / kSampleRate)

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

        HomeViewController *me = (HomeViewController *)inRefCon;

    static int phase = 1;
    static int phase1 = 1;

    for(UInt32 i = 0; i < ioData->mNumberBuffers; i++) {

        int samples = ioData->mBuffers[i].mDataByteSize / sizeof(SInt16);

        SInt16 values[samples];

        float waves;
        float volume=.5;
        float wave1;

        for(int j = 0; j < samples; j++) {


            waves = 0;
            wave1 = 0;

            MyManager *sharedManager = [MyManager sharedManager];


            wave1 = sin(kWaveform * sharedManager.globalFr1 * phase1)*sharedManager.globalVol1;
            if (0.000001f > wave1) {
                [me setFr1:sharedManager.globalFr1];
                phase1 = 0;
                //NSLog(@"switch");
            }

            waves += wave1;
            waves += sin(kWaveform * sharedManager.globalFr2 * phase)*sharedManager.globalVol2;
            waves += sin(kWaveform * sharedManager.globalFr3 * phase)*sharedManager.globalVol3;
            waves += sin(kWaveform * sharedManager.globalFr4 * phase)*sharedManager.globalVol4;
            waves += sin(kWaveform * sharedManager.globalFr5 * phase)*sharedManager.globalVol5;
            waves += sin(kWaveform * sharedManager.globalFr6 * phase)*sharedManager.globalVol6;
            waves += sin(kWaveform * sharedManager.globalFr7 * phase)*sharedManager.globalVol7;
            waves += sin(kWaveform * sharedManager.globalFr8 * phase)*sharedManager.globalVol8;
            waves += sin(kWaveform * sharedManager.globalFr9 * phase)*sharedManager.globalVol9;
            waves *= 32767 / 9; // <--------- make sure to divide by how many waves you're stacking

            values[j] = (SInt16)waves;
            values[j] += values[j]<<16;

            phase++;
            phase1++;

        }

        memcpy(ioData->mBuffers[i].mData, values, samples * sizeof(SInt16));

    }


    return noErr;

}

Thanks in advance for any help!

Answer 1

The OP seems to have solved his problem, but I thought posting an explicit answer would be helpful to the rest of us.

I had the same question of wanting to direct the tones to left and right channels independently. It's easiest to describe in terms of Matt Gallagher's now-standard An iOS tone generator (an introduction to AudioUnits).

The first change to make is to set (following @jwkerr) streamFormat.mChannelsPerFrame = 2; (instead of streamFormat.mChannelsPerFrame = 1;) in the createToneUnit method. Once that is done and you have two channels/buffers in each frame, you need to fill the left and right buffers independently in RenderTone():

// Set the left and right buffers independently
Float32 tmp;
Float32 *buffer0 = (Float32 *)ioData->mBuffers[0].mData;
Float32 *buffer1 = (Float32 *)ioData->mBuffers[1].mData;

// Generate the samples
for (UInt32 frame = 0; frame < inNumberFrames; frame++) {
    tmp = sin(theta) * amplitude;

    if (channelLR[0]) buffer0[frame] = tmp; else buffer0[frame] = 0;
    if (channelLR[1]) buffer1[frame] = tmp; else buffer1[frame] = 0;

    theta += theta_increment;
    if (theta > 2.0 * M_PI) theta -= 2.0 * M_PI;
}

Of course channelLR[2] is a bool array whose elements you set to indicate whether the respective channel is audible. Notice that the program needs to explicitly set the frames of silent channels to zero, otherwise you get some funny tones out.