Pull audio data from graph kAudioUnitSubType_GenericOutput

Question

I am constructing audio graph, which has: AudioFilePlayer (Generator unit) -> GenericOutput (Output unit)

The basic idea:

AudioFilePlayer is producing audio frames to it's output. Generic output takes data from AudioFilePlayer's output as input. Than I can take that data and save it to other buffer.

My question is how to pull data from generic output to get data for offline rendering.

I did some research on it, and find out that I have to use AudioUnitRender on GenericOutput unit to pull audio data from it.

Uint32 frames = 512;
timestamp.mSampleTime = 0;

... While Loop ...

timestamp.mFlags = kAudioTimeStampSampleTimeValid;

bufferList.mNumberBuffers = 1;
bufferList.mBuffers[0].mNumberChannels = param->inputFormat.mChannelsPerFrame;
bufferList.mBuffers[0].mDataByteSize = frames * param->inputFormat.mChannelsPerFrame * sizeof(float);
bufferList.mBuffers[0].mData = NULL;

AudioUnitRenderActionFlags flags = 0;

OSStatus error = noErr;
if ((error = AudioUnitRender(param->genericOutput, &flags, &timestamp, 0, frames, &bufferList)) != noErr)
{
    printf("Cannot AudioUnitRender: %d\n", error);
}

timestamp.mSampleTime += frames;

But i get error with number -50.

Am I setting input parameters wrong? Do I need to register any render callback functions to GenericOutput unit?

Answer 1

The problem was that I set GenericOutput StreamFormat after calling AUGraphInitialize(). It must be done before that.

I used CAShow(reader.graph) to get log of audio graph, which helped me to notice that format is't changed after trying to set it.