Fundamental frequency and vowel perception

Publisher Website
Google Scholar
Abstract
The set of 9 nondiphthong vowels of American English were synthesized using the male formant values from Peterson and Barney. These vowels were produced in 3 conditons of the fundamental frequency: average, 135 Hz; low, 100 Hz; and high, 250 Hz. In forced-choice testing, subjects identified vowels in the average and low condition of the F0 [fundamental frequency of phonation] with greater accuracy than the high F0 vowels. A 2nd experiment was conducted using the female formant frequency values from Peterson and Barney and the same conditions of F0. Subjects still performed better on the low and average conditon of the F0 for these vowels. The human formant frequency extractor is apparently aided by the denser spectral sampling of the transfer function by a lower fundamental. The pattern of vowel errors was the sense neither across F0 conditions nor across male or female formant values. The point vowels [i] and [u] were the most consistently identified across all conditions.

This publication has 1 reference indexed in Scilit: