Add a pipeline that supports reading audio from input devices such as microphones. This pipeline should ahave voice activity detection built in.