When capturing audio, it's very useful to be able to choose the number of channels and sample rate for the recorded data. This commit adds that posibility.