The F1 structure influences final-consonant voicing decisions

Abstract

Previous acoustic analyses [W. V. Summers, J. Acoust. Soc. Am. Suppl. I 79, S36 (1986)] showed higher F1 steady-state frequencies and higher F1 offset frequencies for consonant-vowel-consonant (CVC) utterances ending in voiceless consonants than voiced consonants. These analyses examined the vowels /a/ and /æ/ and the final consonants /b/, /p/, /v/, and /f/. The present experiments examined whether these differences in F1 structure provide perceptual information to the listener concerning final-consonant voicing. The CVC stimuli were synthesized with format steady-state frequencies appropriate to /a/ or /æ/ and with final formant transitions appropriate to the bilabial stops, /b/ and /p/. Several series of synthetic stimuli were created differing in F1 steady-state frequency, F1 final transition slope, and F1 offset frequency. Stimuli within a series varied in steady-state vowel duration. At each steady-state duration, stimuli containing higher F1 steady-state frequencies were heard as containing voiceless final consonants more often than stimuli containing lower F1 steady states. The F1 offset frequency also influenced voicing judgments with higher offset frequencies producing more voiceless responses: The F1 final transition slope did not have a consistent influence on voicing judgments. [Work supported by NIH.]