How to process audio from the user's microphone