Improvement of speaker recognition by combining residual and prosodic features with acoustic features | IEEE Conference Publication | IEEE Xplore