Intonation and Speaker Identification

Abstract

The work described investigated the ability of listeners to identify familiar speakers solely on the basis of suprasegmental laryngeal information. The results of perceptual experiments using both natural stimuli, and synthetic stimuli with manipulations in the time and frequency domains, show that mean fundamental frequency and fundamental-frequency contour shape provide important speaker identifying information for an age-, sex- and accent- matched group even in the absence of all supraglottal features. The investigation is set in the context of the normalisation level of speech perception.