PERCEPTION OF FILTERED SPEECH BY HEARING-IMPAIRED LISTENERS AND BY NORMALLY-HEARING LISTENERS WITH SIMULATED HEARING LOSS.
Item
-
Title
-
PERCEPTION OF FILTERED SPEECH BY HEARING-IMPAIRED LISTENERS AND BY NORMALLY-HEARING LISTENERS WITH SIMULATED HEARING LOSS.
-
Identifier
-
AAI8212205
-
identifier
-
8212205
-
Creator
-
MILNER, PAUL.
-
Contributor
-
Harry Levitt
-
Date
-
1982
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Health Sciences, Audiology
-
Abstract
-
Six subjects participated in a study to measure the perception of filtered nonsense syllables and test the ability of Articulation Theory to predict intelligibility performance. Two of the listeners had normal hearing, three had bilateral sensorineural hearing loss and one had a sensorineural loss in one ear and normal hearing in the other. Ten conditions of low-pass, high-pass and band-pass filtered speech were presented at a minimum of five different presentation levels ranging from near threshold to near discomfort level. All listener heard the speech samples with no added background noise. In addition, the normally-hearing subjects listened to the speech materials in a background of shaped white noise designed to simulate the hearing loss of selected hearing-impaired subjects. All listening was done in a sound-proof room with headphones.;Performance-intensity functions were determined for all the conditions and Articulation Index calculations were performed to determine the accuracy of predicted intelligibility performance versus the observed performance of the listeners.;The results of this study have shown that for all subjects, intelligibility of the materials used remained relatively high whether high-pass filtered at 700 Hz or low-pass filtered at 2800 Hz, and heard with sufficient intensity. The performance-intensity functions of the normally-hearing listeners with simulated high frequency hearing loss were similar to listeners with relatively flat sensorineural hearing loss.;Rollover in performance occurred at high presentation levels that were below reported discomfort thresholds. Articulation Theory calculations generally predicted lower intelligibility scores than were observed for most conditions and most subjects. Although spread out, A.I. values for the normally-hearing listeners when tested in quiet fell closer to a curve of expected performance than for these listeners when tested in noise and for the hearing-impaired listeners.;Errors in the predictions were generally systematic, in that observed performance for the narrowest band conditions showed the greatest difference from the predicted scores. Calculated Articulation Index values did not exhibit the expected monotonicity assumed by Articulation Theory. Methods of computing the Articulation Index appear to require revision to account for the differences between the observed data and performance predicted by Articulation Theory.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
-
Program
-
Speech and Hearing Sciences
