Frequency and time variations of the first formant: Properties relevant to the perception of vowel height

Abstract

Perceptual experiments using consonant–vowel–consonant (CVC) syllables were carried out to examine the perceptual relevance of the first formant frequency (F1) trajectory in the perception of high vowels versus nonhigh vowels. Results show that stimuli characterized by a higher onset frequency and F1 maximum at the beginning of the vocalic portion are perceived as lower vowels than stimuli with a lower F1 onset frequency and F1 maximum toward the end of the vocalic portion. These findings are in agreement with the hypothesis, based on the acoustic analyses of Di Benedetto (1989), that stimuli with higher F1 onset frequencies and F1 maximum at the beginning of the vocalic portion characterize lower vowels. Results are similar for native speakers of different languages, leading to a suggestion that this phenomenon may have either an articulatory or an auditory basis. Possible interpretations based on an overshoot hypothesis or a formant time average theory were investigated through an additional perceptual experiment. Results of this last experiment agree with a weighted average time formant theory.