Comparison of some mechanical models of larynx in the synthesis of voiced sounds

Abstract

The process of voiced sounds production can be described as follows: air coming from the lungs is forced through the narrow space between the two vocal folds, which are set in motion in a frequency governed by the tension of their tissues. The vocal folds change the continuous flow that comes from the lungs into a series of pulses. Then, as the flow passes through the oral and nasal cavities it is amplified and changed until it is finally radiated from the mouth. This complex process can be modeled by a system of integral-differential equations. In spite of such complexity, this paper shows that it is possible to obtain synthetic voice sounds of satisfactory realism using simple mathematical models. The perception of a synthetic sound as natural is increased by choosing suitable waveforms for the time-varying subglottal pressure, rather than by augmenting the number of degrees of freedom of the mechanical model. This paper also shows possible ways to adapt the models to voices of men, women and children.