A Simplified Early Auditory Model with Application in Speech/Music Classification