This may be a stupid question, but what's the flaw? Surely transcribing audio requires a microphone?