High-frequency transfer function model of the vocal tract sections using FEM matrix condensation and transfer matrix techniques

Abstract

Using numerical tools such as the Finite Element Method (FEM) for acoustical modeling of the vocal tract would require a huge number of nodes and elements. In this paper, the matrix condensation and transfer matrix techniques [A. Craggs, J. Sound Vib. 132, 393–402 (1989)] adapted to a three-dimensional model of the vocal tract section analysis will be introduced and discussed. Using these new FEM techniques, which use only the nodes at the input and output of the structure, the transfer function can be computed with a very small matrix comparing to standard FEM. The second goal of this research is to attempt to discover some geometrical parameters which could influence the transfer function. This investigation could be useful to make a new electrical vocal tract model which would consist of cascade transfer functions controlled by parameters. The study of the transfer functions will be carried out in low and high frequencies. [Work partly supported by Japanese CREST project.]