Stability of vowel formants based on a simple acoustic tube model and a tongue model

Abstract

Mapping from articulatory parameters to acoustic parameters for vowel(s) was examined based on a simple acoustic tube model of the vocal tract and on a three-dimensional model of the tongue. In the acoustic tube model, two articulatory parameters, i.e., the location (L) and size (S) of the constriction, were selected. The second formant frequency showed an interesting folded pattern, as L varies, in the vicinity of the formant frequency for the vowel /i/. However, only using L and S, no other quantal characteristics in Stevens' sense [K. N. Stevens, Human Communication, A Unified View, edited by P. B. Denes and E. E. David, Jr. (McGraw-Hill, New York, 1972), pp. 51–66] were observed. On the other hand, the tongue model demonstrated stable formant frequencies for the vowel /i/ as reported in our previous paper [O. Fujumura and Y. Kakita, Frontiers of Speech Communication Research, edited by B. Lindblom and S. Ohman (Academic, London, 1979), pp. 17– 24]. The stabilizing effect of muscle contraction on formant frequencies is further discussed.