Speaker Identification from Turbulent Portions of Fricatives

Abstract

Previous reports have not indicated whether inter-speaker variation in the production of fricative consonants is great enough so that turbulence can serve as a cue in speaker identification. In this experiment 12 listeners were exposed to isolated turbulence portions of fricatives, as produced by 8 male speakers, and asked to judge speaker identity. Results indicate that speakers can be identified from such stimuli, that the addition of a laryngeal source results in higher identification levels, and that performance tends to improve as place of articulation moves posteriorly.

Sex Identification of Preschool Children’s Voices

“…For example, in adult sex identification studies, respondents have demonstrated accuracy in identifying maleness and femaleness from presentations o f voiceless fricatives, isolated spectral regions o f frication noise, whispered vowels, and speech samples employing an electrolarynx sound source (Schwartz. 1968;Ingemann, 1968;Coleman, 1971;La Riviere, 1974). These results suggest that vocal tract resonance characteristics make the greatest contribution to accurate perception o f speaker sex in the absence of fundamental frequency (F") in formation.…”

An Approach to Speaker Identification

“…One area of study here is on speech suprasegmentals, included are as follows: SFF, or, speaking fundamental frequency , voice quality (often determined by long‐term spectra , temporal speech features , and so on. SI research also has been carried out on speech segmentals; that is, on vowels and their formants as well as on diphthongs , nasality , fricatives , whispers , and related segmentals . These relationships have often been revisited over the years.…”