Voice quality: Speaker identification across age, gender, and ethnicity
Sonja A. Trent-Brown · The Journal of the Acoustical Society of America · 2018
Listeners can perceptually identify speaker gender and ethnicity at better than chance guessing for adult speakers (Lass et al., 1978; Thomas & Reaser, 2004; Trent-Brown et al., 2012). These findings are supported across varying levels of phonetic complexity and across temporal manipulations that impact perceptual processing. There is less evidence regarding the extent to which these patterns hold for child speakers and whether listener gender/ethnic background and experience influence accuracy of identification. African American and European American male and female adult speakers were recorded producing passage-, sentence-, and word-length (/hVd/) stimuli including 11 General American English vowels. African American and European American male and female child speakers, ages 8–12, were recorded producing sentence and word stimuli. A temporal manipulation introduced to distort perceptual expectations presented stimuli to listeners in both forward and reversed conditions. Listeners were male and female African American and European American undergraduate students. Significant outcomes were observed for accuracy across speaker gender, speaker ethnicity, phonetic complexity, and temporal condition. In addition to accuracy of identification, listener confidence ratings, identification reaction times, and confidence reaction times also showed differential outcomes with respect to speaker age, gender, and ethnicity.