An analysis of /p/, /b/ and /m/ in the speechreading signal.
Item
-
Title
-
An analysis of /p/, /b/ and /m/ in the speechreading signal.
-
Identifier
-
AAI8820900
-
identifier
-
8820900
-
Creator
-
Scheinberg, Judith Carol Schwarz.
-
Contributor
-
Adviser: Harry Levitt
-
Date
-
1988
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Health Sciences, Audiology
-
Abstract
-
Visible articulatory movements of /p/, /b/ and /m/ were examined in a VCV context with the vowels /i/, /a/ and /u/. In the first experiment, the locations for measurement were determined using an interleaving technique in which odd frames of one VCV sequence were alternated with even frames of another VCV sequence. Areas of jitter appeared in those regions of the face that were not in identical positions at the same time during the production of the two utterances. Jitter occurred primarily in the lower lip, chin and cheeks. The areas of jitter indicated the locations of interest for detailed measurement. In the second experiment, VCV productions were recorded with face markers placed at the jitter locations. Automatic tracking of the face markers was accomplished by computer. Plots of face marker movement for the mid upper lip, mid lower lip and chin were analyzed and the major findings were as follows: Both lips and the chin were displaced further upward for the production of /m/ than for /p/ and /b/. The rate of movement away from closure for the lower lip and chin was greater for /m/ than for /p/ or /b/. The rate of movement of the upper lip towards closure was faster for /p/ and /b/ than for /m/. There were only subtle durational differences measured between /p/, /b/ and /m/. In the third experiment, a forced-choice speechreading task tested the two-way discrimination between /m/ and /b/, /m/ and /p/, and /p/ and /b/ in the CV context with the vowels /i/, /a/ and /u/. The results showed that observers were able to discriminate between consonants with the same place of production, with the best performance for the /u/ context. The results suggest that /p/, /b/ and /m/ should be grouped into two visemes: /p,b/ and /m/.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
