Acoustic coupling in phonation and its effect on inverse filtering of oral airflow and neck surface acceleration

Matias Zanartu Salas, Purdue University

Abstract

Many aspects of the coupling between vocal fold motion, glottal airflow, and resulting sound waves remain unexplored, particularly during incomplete glottal closure. In addition, this coupling has been associated with distortions in the estimates of glottal parameters obtained from neck acceleration measurements, thus affecting the ambulatory assessment of vocal function. This acoustic coupling and its effects on inverse filtering of oral airflow and neck surface acceleration during vowel production were investigated using numerical and experimental methods. A lumped element self-oscillating model was extended to include different types of incomplete glottal closure, which provided an enhanced simulation tool that mimicked normal and hyperfunctional voices. Novel in vivo measurements and analyses of vocal fold tissue instabilities due to acoustic coupling were obtained using synchronous laryngeal high-speed videoendoscopy recordings. These observations suggested that instabilities differed at the vocal fold tissue level from those where no strong coupling was present, but classified each as subcritical Hopf bifurcations. These findings supported the concept of describing interactions within the voice production system based on acoustic impedance representations. In addition, an impedance-based inverse filtering approach with time-varying glottal coupling demonstrated how source-tract interactions skewed and altered the glottal pulses. Based on these observations, a subglottal inverse filtering scheme for neck surface acceleration was proposed and evaluated, yielding estimates of glottal aerodynamic parameters with accuracy comparable to that of standard pneumotachograph measurements and closed-phase inverse filtering.

Degree

Ph.D.

Advisors

Wodicka, Purdue University.

Subject Area

Speech therapy|Biomedical engineering|Electrical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS