Abstract
Controlled phonations of the 12 commonly distinguished vowels of the general American dialect were tape recorded from 4 trained speakers, and varied experimentally with respect to duration, phonetic context, and inflection. The randomized vowels were presented to 18 trained listeners at a level 60 decibels above threshold in a sound isolated room. Vowels may differ markedly in the relative identification ease. Vowels from phonation to phonation may be considerably more unstable than has heretofore been supposed. Pitch modulation may be a contributing cue to recognition. Phonetic context may provide significant cues to identification. Vowel recognition varies as a function of duration in ways which depend upon the natural duration of vowels. Complete specification of spoken vowels must provide information about its energy frequency distr. and the way in which it varies in a contextual speech pattern.
Keywords

This publication has 1 reference indexed in Scilit: