Are different kinds of acoustic features processed differently for speech and non-speech sounds?