uses the same code that the LVA uses to capture microphone audio. Defaults to downmix and accepts left and right as variable for --channel.