I'm completely making this up; I'd guess most speaker generated audio would be from a compressed source. Audio compression generally cuts off frequencies that

I'm completely making this up; I'd guess most speaker generated audio would be from a compressed source. Audio compression generally cuts off frequencies that we can not hear. When we speak though, we must be generating a lot of inaudible frequencies. It could be determined by checking if those exist or not.