First International Conference on Spoken Language Processing (ICSLP 90)

Kobe, Japan
November 18-22, 1990

Laryngeal Modeling: Translaryngeal Pressure for a Model With Many Glottal Shapes

Ronald C. Scherer (1,2), Chwen-geng Guo (1)

(1) The Recording and Research Center, The Denver Center for the Performing Arts, Denver, CO, USA
(2) The Wendell Johnson Speech and Hearing Center, The University of Iowa, Iowa City, IA, USA

Voice quality is dependent on the shape and amplitude of the laryngeal airflow signal. Physiological models of the airflow need laryngeal pressure-flow-geometry information. A comprehensive translaryngeal pressure-flow equation is offered. A model of the larynx 7.5 times life size was used with 63 combinations of glottal angles and diameters. The translaryngeal pressure-flow data were nondimensionalized into a pressure coefficient P* and Reynolds number Re such that P* = (A1/Re) + A2. A1 and A2 were empirically determined from the data, and were structured on Poiseuille, diffuser, and optimum pressure recovery considerations. For pressure drops ranging between 3 and 50 cm H2O, the average mean difference between the equation predictions and the empirical data was 4.45% (sd 3.38%). The results can be used in phonatory models or speech synthesis schemes for which volume velocity is dependent upon subglottal pressure and glottal configuration.

Full Paper

Bibliographic reference.  Scherer, Ronald C. / Guo, Chwen-geng (1990): "Laryngeal modeling: translaryngeal pressure for a model with many glottal shapes", In ICSLP-1990, 57-60.