A Study on the Preferred Level of Late Reverberation in Speech and Music