Articulatory and perceptual factors controlling the age- and sex-conditioned variability in formant frequencies of vowels

Abstract

Speech sounds, as heard by listeners, contain phonetic, personal, and transmission information. The differences between the formant frequencies of vowels spoken by men, women, and children show a fairly uniform tendency in several studies and languages, and they are regarded as personal quality differences. The differences between the sexes are mainly due to the descent of the larynx in males during puberty. The observed tendency in female/male formant frequency ratios is reproduced in a calculation taking into account the physiological consequences of larynx descent and assuming that the vowel specific neural commands to the articulators remain unchanged. The perception of phonetic quality is seen as a process of tonotopic gestalt recognition. The tonality (= critical-band rate) distances between the formants in phonetically identical vowels are claimed and shown to be invariant as long as they are smaller than 6 Bark. The absolute position of the formants allows personal variation. The tonality distance between the first formant and the fundamental is smaller in most vowels spoken by women than in those by men and children. As for the role of the fundamental in this connection, some alternative hypotheses are discussed.