Music models for music-speech separation | IEEE Conference Publication | IEEE Xplore