Hi Kevin,
Thanks for the tip on memsetting the buffers to 0 so that I play silence instead of rubbish if there is a glitch.
My problem is somewhat resolved - the sample type being returned by the AudioUnit is SInt32 (AudioSampleType). I was using SInt16,
I had to do some conversion to convert from SInt32 to SInt16.
My audio was mono so it was ok to take from one of the buffers.. ioData.mBuffer[0]. Worked.
I have 2 additional questions here -
1) Am I able to get my mixer AudioUnit to return SInt16 instead? (so that I don't need to convert it)
2) Am I able to get AudioUnit to return mono (one buffer)?
2) Is converting from SInt32 to SInt16 within the render callback considered expensive?
My alternative is to use SInt32 (AudioSampleType) for my buffers, but it'll take up more memory in that case, at not much audible improvement imo.
Thanks for all the help.