Individual differences in speech production: Voice-onset-time

Abstract

As part of an ongoing study of talker-specific speech processing, we sought to document the existence of individual talker differences in voice-onset-time (VOT). Four female and four male talkers, all native speakers of American English, were asked to produce 30 repetitions each of 18 different words beginning with /p/, /t/, or /k/. Words were displayed visually on a computer monitor one at a time in a randomized order at a constant presentation rate, and talkers were asked to produce the words at a comfortable speaking rate. The VOT and duration of each spoken word were measured. The eight talkers showed substantial individual variation in speaking rate (as measured by word duration) and, as would be expected, slower speaking rates were associated with longer VOT values. In order to determine whether there were individual differences in VOT beyond those associated with speaking rate, we conducted analyses that controlled for speaking rate on a word-by-word basis. These analyses revealed cases of consistent individual differences in VOT. We are currently examining whether the structure of listeners’ perceptual voicing categories reflects such individual talker differences. [Work supported by NIH/NIDCD.]