[jjk166]

I doubt the sampling rate is anywhere near what you would need to make out dialog in a sound recording. You might be able to tell who is speaking though if you had a voice profile.