Throat singing synthesis by a laryngeal voice model based on vocal fold and false vocal fold vibrations

摘要

In this paper, the authors estimated laryngeal voices of throat singing using an inverse-filtering method and proposed a laryngeal voice model based on vocal fold and false vocal fold vibrations. In the previous paper, the physiological observations on laryngeal movements of throat singing by means of the simultaneous recording of singing sounds, EGG, and high-speed digital images were reported. From the physiological observations and estimated laryngeal voices, we assumed (1) in pressed voice production, vocal folds and false vocal folds vibrated in almost opposite phase; (2) in karygraa voice production, two closed phases of vocal folds were appeared in one period of a glottal volume flow waveform, and at one of the two closed phases, vocal folds were incompletely closed. The authors also proposed a four-mass model as a self-oscillating model of vocal fold and false vocal fold vibrations which is developed from a two-mass model. The simulation of laryngeal movements using a four-mass model agreed with the above assumption for two laryngeal movement patterns of throat singing. Furthermore, the authors also proposed a laryngeal voice model for throat singing and synthesized throat singing sounds.