885095944

THURSDAY AFTERNOON, 2 MAY 1991

INTERNATIONAL B, 1.00 TO 5:00 P.M.

Session 7SP

Speech Communication: Issues in Production and Perception

Maureen L. Stone, Chair

National Institutes of Health, Department of Rehabilitation, Building 10, Room 6S235, Bethesda,

Maryland 20892

Contrihuted Papers

1:00

7SP1. Effects of typicality and interstimulus interval on the discrimination of speech stimuli. Minoru Tsuzaki (ATR Auditory and Visual Perception Res. Labs., Seika-cho, Soraku-gun, Kyoto 619-02, Japan) and Jorge A. Gurlekian (Escuela de Salud Publica, Buenos Aires, Argentina)

To investigate the effects of processing time and retention inlcrval on the perceptual judgment for speech sounds, iapanese listeners were tested with the AX discrimination task in various interstimulus inter-vals. Speech stimuli were synthesized to comprisc both typical and atyp-ical stimulus sets. In the typical set, the standard stimulus had charac-terislics of the typical Japanese /aba/ sound. In the atypical set, the standard was not a good exemplar of the Japanese /aba/ sound al-though it was usually recognized as /aba/. The performance of the subjects was analyzcd in terms of both the center and width of discrim-inal processes. At relatively short intervals, there was a slight tendency for the typical standard to have a smaller width score than the atypical standard. At long intervals, there was noeffect of the standard stimulus. Although the center shifted to the direction of morę plosive counter-parts at longer intervałs, there was no effect of the standard on the center of discriminal process. The results suggest that the typical standard is represented with less noise, but this advantage decreases with inereases in the interstimulus interval.

1:30

7SP3. Models for voicing and aspiration in stop consonants. Kenneth N. Stevens (Res. Lab. of Electron, and Dept. of Elcctrical Eng. and Comput. Sci.. MIT. Cambridge. MA 02139)

A theorelical analysis of sound production for voiceless and voiced stop consonants has bccn carried out. The analysis includes the effects of activc and passive (in response to intraoral pressure changes) expansion or contraction of the pharyngeal volume, active and passive changes in glottal configuration, the generation of periodic glottal vibration and aspiration noise, and the generation of transient and frication noise sources at the consonantal rełease. Both monopole and dipole components of the turbulence noise sources are considered. The model is used to calculate the absolute sound-pressure levels and spectra for the var-ious components of the radiated sound for labial, alveolar, and velar stops. The predictions of the model are in reasonable agreement (within a few dB) with data from spoken syllabłes. Procedures for estimating the characteristics of turbulence noise sources still need to be refined, particularly the distribution of the sources downstream from a glottal or supraglottal constriction. Tłie model can be used to infer details of the time course of supraglottal and laryngcal constrictions based on data from hne-grained acoustic analysis of utterances. [Research supported in part by NIH Grant DC00075.]

1:15

7SP2. The effects of attention on selective adaptation. Joan E. Sussman, Sandra Lane, and Valerie I-auekner (Dept. of Commun. Disord. and Sci., State Univ. of New York at Buffalo, 122 Park Hall, Buffalo, NY 14260)

Ten young adults participated in four conditions of selcctivc adaptation prcccded by baseline labeling: (1) end point-[ba], (2) “focused” (subvocal repetition of the adaptor) end point-[ba], (3) end point-[da], and (4) “distracted” end point -[da] with adaptor and test stimuli pre-sented to the right ear and synthetic [si] and [/i] syllabłes presented to the left ear at 40 dB SPL for periodic labeling during presentation of the adaptor. The purpose of the investigation was to determine: (1) if the amount of adaptation could be inereased by focusing subjects’ attention to the [ba] adaptor and (2) if the amount of adaptation could be de-creased by focusing subjects’ attention away from the [da] adaptor. Results showed that (1) no signifieant differences between focused attention and [ba] selcctive adaptation (5.4% vs 3.7% shift of [ba] labels) and (2) surprisingly, an inereased amount of selective adaptation for the “distractor” condition compared to the end point-[da] condition (19.3% vs 10.3% shift of [ba] labels). Results will be related to auditory versus higher-Ievel cognitive determinants during selective adaptation labeling tasks. [Work supported by NSF.]

1978 J. Acoust. Soc. Am., Vol. 89. No. 4. Pt. 2. April 1991

1:45

7SP4. A two-dimensional model of laryngeal flow. Fabriborz Alipour (Voice Acoust. and Biomechanics Lab., Dept. of Speech Pathology and Audiology, Univ. of Iowa, Iowa City. IA 52242) and Virendra C Patel (Dept. of Mech. Eng., Univ. of Iowa, Iowa City, IA 52242)

Laryngeal flow was visualized by numerical simulation. Using com-putational fluid dynamics, a two-dimensional model of laryngeal flow was built and aerodynamic properties were calculated for steady-state laminar regime. Thrcc configurations of voca1 folds with convergent. rectangular, and divergent glottis were used to study the effects of glot tal shapc on the airflow. Navier-Stokes cquations were solved with numerical method of Patel et ai A boundary fitted coordinate was used in discretizing the flow domain into exponential grids. The govcrning cquations and coordinates were traasformed into computational domain and solved with finite analytic method. Results were reported on the velocity components, pressure distributions, and wali friction coefficient at Reynolds numbers of 100-900 for different configurations. Results were compared with one-dimensional Bernoulli Solutions and experi-mental data. It was found that flow separation exits even at Iow Reynolds numbers and the separation point moves toward downstream as the Reynolds number inereases. [Work supported by NINCDS Grant No. NS 16320-08.]

121st Meeting: Acoustical Society of America 1978