Acoustic Analysis of the Voice in Native and Nonnative Speakers

There are significant differences in voice parameters for nonspeech tasks between culturally and linguistically diverse sample populations. Normative data of standard clinical software programs is typically comprised of North American speakers of Standard American English (SAE). There is a need for normative spectral and cepstral data across sociolinguistic groups to ensure clinical objective measurements are accurately classifying the voice quality of all individuals. The purpose of this study was to (1) compare objective measures of voice quality assessment of monolingual speakers of SAE with native speakers (L1) of French and Spanish on acoustic spectral and cepstral analyses; and (2) compare ratings on auditory-perceptual assessment with acoustic data secondary to degree of accentedness. Perceptual analyses and voice quality measures for frequency, cepstral measures, and perturbation measures were analyzed using the Analysis of Dysphonia in Speech and Voice (ADSV) and the MultiDimensional Voice Program (MDVP) in 10 L1 English, 10 L1 Spanish, 10 L1 French speakers. An informal measure of accentedness of SAE was performed. ANOVAs were included for identifying group differences in perceptual ratings and acoustic data, and the relationship between the degree of accentedness and CAPE-V scores. Results indicate relative accuracy of objective measures of voice assessment for multicultural clinical populations. In this study, two variables demonstrated significance for main effect of language, with 17 other variables of voice quality at minimum, approaching significance. Therefore, appropriate consideration of cultural competence is still warranted in therapeutic and diagnostic services.