This is required so we can do different spatialization logic depending
on the channel count of the audio source.