Vowel normalization: The role of fundamental frequency and upper formants.
Item
-
Title
-
Vowel normalization: The role of fundamental frequency and upper formants.
-
Identifier
-
AAI9908321
-
identifier
-
9908321
-
Creator
-
Halberstam, Benjamin.
-
Contributor
-
Adviser: Lawrence J. Raphael
-
Date
-
1998
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Speech Communication
-
Abstract
-
Some vowel normalization schemes attempt to account for successful vowel classification by listeners despite the interspeaker overlap between acoustic vowel categories in the F1 x F2 space, by assuming perceptual exploitation of F0 and F3 information.;This study tested an implicit prediction made by these vowel normalization schemes: availability of F0 and F3 information should improve listeners' classification of naturally produced vowels with typical interspeaker formant variability presented in a mixed-speaker condition, relative to their ability to classify the same stimuli in a blocked-speaker condition.;Eight subjects classified phonated and whispered vowels from the set /i I {dollar}\varepsilon{dollar} ae a c {dollar}\Lambda{dollar} U u/ produced by speakers varying in age and sex. Stimuli were presented with and without F3 and upper formants filtered out, in blocked-speaker and mixed-speaker conditions.;Vowels were classified with greater accuracy for blocked-speaker than for mixed-speaker condition, for phonated vowels than for whispered vowels and for unfiltered than for filtered vowels.;The interaction between presentation condition and phonatory type was just short of significance at the 0.05 level (p. = 0.068). Phonated vowels were classified with similar accuracy in blocked-speaker and mixed speaker conditions. Whispered vowels were classified with significantly lower accuracy in mixed-speaker condition than in blocked-speaker condition. This finding indicates that F0 is likely to be perceptually useful in vowel normalization.;The interaction between phonatory type and presentation condition was of low magnitude, and was not significant at the 5% level (p. = 0.189). Thus, the results did not provide perceptual support for including F3 as a parameter in vowel normalization.;The interaction between phonatory type and availability of upper formants was extremely robust (p = 0.001). The implication of this unexpected finding is that third formant information is relatively inconsequential for phonated vowel perception, but of considerable importance for whispered vowel perception.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
