The same idea applies to audio data.