Glottal modeling and closed-phase analysis for speaker recognition.