How to set up Splitter Unit in Core Audio - core-audio

Question

From the book "Learning Core Audio",I have learned how to mix the stream 1(voice from mic) and stream 2(voice synthesizer) then connect its output
to outputUnit(speaker).
Now I tried to record its output while monitoring its output from speaker.
So I set up a Splitter unit, connect the output of mixerUnit to its input,
connect its outputScope 0 to outputUnit, set a renderCallback(for recording) to its outputScope 1.(I thought Splitter Unit will split the input stream to bus 0 and bus 1)
The result was that I can hear it from speaker, but the renderCallback never get called due to its "Pull Model", since it didn't connect to any output unit.
But if I set up another Generic Output Unit then connect to it, AUGraph showed me an Error message.(No 2 Output Unit allowed)
Anybody can help me out here ?
CheckError(AUGraphNodeInfo(player->graph,
outputNode,
NULL,
&player->outputUnit),
"AUGraphNodeInfo failed");
CheckError(AUGraphNodeInfo(player->graph,
speechNode,
NULL,
&player->speechUnit),
"AUGraphInfo failed");
AudioUnit mixerUnit;
CheckError(AUGraphNodeInfo(player->graph,
mixerNode,
NULL,
&mixerUnit),
"AUGraphNOdeInfo failed");
AudioUnit splitterUnit;
CheckError(AUGraphNodeInfo(player->graph,
splitterNode,
NULL,
&splitterUnit),
"AUGraphInfo failed");
CheckError(AudioUnitSetProperty(mixerUnit,
kAudioUnitProperty_StreamFormat,
kAudioUnitScope_Input,
0,
&player->streamFormat,
propertySize),
"Couldn't set stream format on mixer unit bus 0");
CheckError(AudioUnitSetProperty(mixerUnit,
kAudioUnitProperty_StreamFormat,
kAudioUnitScope_Input,
1,
&player->streamFormat,
propertySize),
"Couldn't set stream format on mixer unit bus");
CheckError(AudioUnitSetProperty(splitterUnit,
kAudioUnitProperty_StreamFormat,
kAudioUnitScope_Input,
0,
&player->streamFormat,
propertySize),
"Couldn't set stream format on Splitter unit");
CheckError(AUGraphConnectNodeInput(player->graph,
splitterNode,
0,
outputNode,
0),
"Couldn't connect splitter 0 to outputNode");
CheckError(AUGraphConnectNodeInput(player->graph,
speechNode,
0,
mixerNode,
1),
"Couldn't connect speech speechNode to mixer input(1)");
AURenderCallbackStruct callbackStruct;
callbackStruct.inputProc = GraphRenderProc;
callbackStruct.inputProcRefCon = player;
CheckError(AudioUnitSetProperty(mixerUnit,//was outputUnit
kAudioUnitProperty_SetRenderCallback,
kAudioUnitScope_Input,
0,
&callbackStruct,
sizeof(callbackStruct)),
"Couldn't set render callback on mixer unit");
AURenderCallbackStruct recorderCallback;
recorderCallback.inputProc = recordRenderProc;
recorderCallback.inputProcRefCon = player;
CheckError(AudioUnitSetProperty(splitterUnit,
kAudioUnitProperty_SetRenderCallback,
kAudioUnitScope_Output,
0,
&recorderCallback,
sizeof(recorderCallback)),
"Couldn't set render callback on splitter output");

Answer 1

Are you just trying to record your final mixed output to a file?
If so, you can write to the file in your render callback. However, you can't block for very long. You have to completely finish your writing before the next callback. The easiest way to write would probably be to use grand central dispatch to asynchronously write your file.