使用不同格式将音频从文件渲染到语音处理音频单元

Question

Using Core Audio, I am trying to manually play a music as an exercise to play around with CoreAudio and learn. As my purpose is for learning, I would like to understand why the following does not work.

Everything works fine when I simply create those two units without connecting them together (neither manually or with AUGraph) :

• outputAU (kAudioUnitSubType_VoiceProcessingIO)
• fileGeneratorUnit (kAudioUnitType_Generator)

Instead I configured a rendering callback on outputAU so that inside I am able to call in order to manually populate ioData.

var err = AudioUnitRender(outputAU,&flags,&ts,inBusNumber,inNumberFrames,&ioData!.pointee)

This works well my outputUnit is able to play the song.

However it stops working if I change the format of my outputUnit to this one :

var format = AudioStreamBasicDescription(mSampleRate: 44100,
                                         mFormatID: kAudioFormatLinearPCM,
                                         mFormatFlags: kLinearPCMFormatFlagIsSignedInteger | kLinearPCMFormatFlagIsPacked,
                                         mBytesPerPacket: bytePerSample, // bytePerSample = 2
                                         mFramesPerPacket: 1, 
                                         mBytesPerFrame: bytePerSample,
                                         mChannelsPerFrame: 1,
                                         mBitsPerChannel: 8 * bytePerSample,
                                         mReserved: 0)  

I set this format on both output and input scopes like this :

error = AudioUnitSetProperty(outputAU,
                             kAudioUnitProperty_StreamFormat,
                             kAudioUnitScope_Output,
                             input,
                             &format,
                             UInt32(MemoryLayout<AudioStreamBasicDescription>.size))

guard error == noErr else {
    print("An error occured \(error)")
    return
}

error = AudioUnitSetProperty(outputAU,
                             kAudioUnitProperty_StreamFormat,
                             kAudioUnitScope_Input,
                             output,
                             &format,
                             UInt32(MemoryLayout<AudioStreamBasicDescription>.size))

guard error == noErr else {
    print("An error occured \(error)")
    return
}

My AudioUnitRender call is now returning me constantly -50. And I am unable to populate the buffer. I am interested to understand what is going on there and how I can fix this issue.

Thanks!