Dept. of Linguistics, Ohio State Univ., 222 Oxley Hall, 1712 Neil Ave., Columbus, OH 43210-1298
Listeners can recognize both the voice of a familiar speaker and the words spoken. The experiment reported in this paper explored the interaction of these two perceptual abilities. In particular, it was hypothesized that vowel perception is tied to speaker identity. This hypothesis predicts that vowel perception performance will vary as a function of the listener's exposure to the voice of the speaker. Two speakers served as models for sets of synthetic vowels which matched the speakers' voice characteristics, formant bandwidths, and higher formant values. Eight groups of listeners manipulated the frequencies of the two lowest formants in a method of adjustment paradigm. The task was to make the synthetic vowel match the vowel qualities of words illustrating the monophthongs of English. For each voice, one group of listeners had no prior exposure to the target speaker, and three groups listened to one of three tape recordings of the speaker (and were instructed to ``memorize the speaker's voice and make the synthetic vowels match the speakers pronunciation''). The results showed that exposure to a speaker's voice had an influence on the formant values chosen by the listener, that listeners were able to retain acoustic/auditory information about vowel exemplars.