The suspending audio-source object allows an application to temporarily suspect the audio going into speech recognition. While suspended, the wave-in/out device will be freed, so an application can play a wave file or do other recording.