集中系モデルを用いた音声生成解析

抄録

Speech is produced by self-excited oscillation of vocal cords. In previous works, the vocal cords were modeled to perform by time history analysis of self-excited oscillation in various ways. However, eigenvalue analysis of coupled acoustic and vocal cords vibration has been not performed. In this paper, we propose a concentrated mass model to analyze a speech production and a self-excited oscillation in the vocal cords, this model can be applied to eigenvalue analysis. This model consists of mass points, connecting springs and dampers. And, we simulate a speech production by the concentrated mass model. Then, we conduct an experiment by using model of vocal organ. We confirm the validity of the concentrated mass model by comparing numerical results and measurement results. According to these result, it is concluded that proposed model is valid.