Fluid flow in a dynamic mechanical model of the vocal folds and tract. II. Implications for speech production studies

Abstract

A companion paper [Barney et al., J. Acoust. Soc. Am. 105, 444–455 (1999)] presents measurements in a dynamic mechanical model (the DMM) of the vocal folds and vocal tract: It was shown that closer prediction of the radiated sound pressure was possible when nonacoustic (vortical) as well as acoustic components of the velocity in the duct were included. In this paper, using such a simple geometry to model the vocal tract is justified by comparing acoustic and aerodynamic measurements in the DMM to those made in vivo: sub- and supraglottal pressures, radiated pressure, and hot wire velocities. The DMM produces sound equivalent to weak, low-frequency falsetto. A Rothenberg mask was then placed on the end of the DMM, and two estimates of the glottal waveform were compared to velocities measured near the “glottis.” The results show that the glottal waveform does not resemble any hot wire velocities measured near the shutters; travel times for acoustic and nonacoustic components of the velocity field differ significantly, which may cause problems when using the Rothenberg mask to analyze transients; the mask itself alters the acoustic field in the duct, and the radiated pressure, significantly. The evidence points toward the existence of a vortex train during and caused by phonation, and significant sound generation due to the interaction of that train with tract boundaries; these findings indicate that the models on which inverse filtering are based have been overgeneralized.