Two-Mass Model of the Vocal Cords
D. E. Dudgeon · The Journal of the Acoustical Society of America · 1970
Early vocal system models consisted essentially of a pulsating source, which was used to drive a time-varying filter representing the vocal tract. A later model attempted to parallel the human vocal system more closely by including a mechanical (mass-spring) oscillator, representing the vocal cords, driven by feedback within the vocal system model. This provided a self-excited source for driving the vocal tract analog, bringing speech synthesis and human speech production closer together. The two-mass vocal cord model is a slightly more complex version of the mechanical oscillator, and is designed to incorporate more phenomena observed in human vocal cords by high-speed motion picture techniques. This slight increase in complexity enables even closer analogies between synthetic and human vocal systems. In particular, experiments with this model show that it can sustain phase differences in the displacements of the upper and lower edges of the vocal cords as do the human vocal cords. Other experiments with the two-mass model indicate that there are a range of frequencies at which the vocal cords cannot be driven by the feedback system. These frequencies are related to the natural frequencies of the vocal tract configuration.