Applying the cellular automaton model in speech production process simulation

Abstract

This study investigates the dynamic simulation of the speech production process. When an utterance occurs, the sound pressure and the particle velocity distribution of the vocal tract space changes from word to word. The author examined and visualized this change to illustrate how it happens. The author used the cellular automata as a model to calculate the acoustic characteristics of the vocal tract in each region. This study takes a different approach from the traditional z-transform to the cellular automata to calculate the dynamic sound pressure at each site of the vocal tract space. To simplify the calculation, an acoustic cylindrical tube model was used. Dynamic calculations were made even more manageable by changing the physical model from a three-dimensional to a one-dimensional cellular automaton model. Consequently, this study demonstrates that the relationship between the words and the sound pressure distribution can be examined dynamically using the cellular automata model.