You are here

Speechreading's Benefit to the Recognition of Sentences as a Function of Signal-to-Noise Ratio

Title: Speechreading's Benefit to the Recognition of Sentences as a Function of Signal-to-Noise Ratio.
47 views
18 downloads
Name(s): Southard, Stuart D., Jr., author
Morris, Richard J., professor directing dissertation
Kamata, Akihito, outside committee member
Walker, Virginia, committee member
Goldstein, Howard, committee member
Terrio, Lee, committee member
School of Communication, degree granting department
Florida State University, degree granting institution
Type of Resource: text
Genre: Text
Issuance: monographic
Date Issued: 2003
Publisher: Florida State University
Place of Publication: Tallahassee, Florida
Physical Form: computer
online resource
Extent: 1 online resource
Language(s): English
Abstract/Description: Despite attempts to improve speech recognition in noise via hearing aid algorithms, communication where there is background noise continues to present serious challenges to hearing aid users. The viability of using speech reading to augment amplification to improve communication in noise has not been fully investigated. One reason for this oversight is the assumption that the decline in speech recognition performance with decreases in S/N ratio is similar for communicative situations in which speech reading is available and for those in which it is not. In this study, comparison of speech recognition performance for auditory only (A) and audio-visual (AV) presentation of sentences was made at four signal-to-noise-ratios ranging from 0 to –6 dB. Twenty-six participants with normal hearing and vision, 20 to 30 years, responded to eight experimental conditions (A and AV presentation of speech at each of four signal-to-noise ratios). Speech recognition performance was evaluated using a 2 X 4 repeated measures ANOVA. The main effects of presentation mode and S/N ratio were statistically significant. More importantly, there was a significant interaction between these factors for speech recognition performance. The interaction revealed a growing benefit of speech reading, defined as the difference between audio-visual and auditory only performance (AV-A), with decreasing signal-to-noise ratios. This finding of increasing benefit with deteriorating listening conditions refutes the assumption that performance decline with increasing background noise is independent of presentation mode. This finding suggests that the need for individuals with hearing impairment to avail themselves of visual speech cues increases with a deteriorating listening environment. In view of these findings, hearing aid evaluations involving both A and AV presentation of speech at different signal-to-noise ratios are recommended. Such procedures allow for both objective and subjective assessment of speech recognition performance under combinations of presentation mode and S/N ratios.
Identifier: FSU_migr_etd-1618 (IID)
Submitted Note: A Dissertation Submitted to the Department of Communication Disorders in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy.
Degree Awarded: Fall Semester, 2003.
Date of Defense: October 31, 2003.
Keywords: Comparison of audio with audio-visual speech recognition
Bibliography Note: Includes bibliographical references.
Advisory committee: Richard J. Morris, Professor Directing Dissertation; Akihito Kamata, Outside Committee Member; Virginia Walker, Committee Member; Howard Goldstein, Committee Member; Lee Terrio, Committee Member.
Subject(s): Communication
Persistent Link to This Record: http://purl.flvc.org/fsu/fd/FSU_migr_etd-1618
Owner Institution: FSU

Choose the citation style.
Southard, S. D. (2003). Speechreading's Benefit to the Recognition of Sentences as a Function of Signal-to-Noise Ratio. Retrieved from http://purl.flvc.org/fsu/fd/FSU_migr_etd-1618