Let's put it back in if/when we have a clearer view of voice/music detection