Rate effects in speech and nonspeech

Abstract

A well-documented effect in speech perception is the rate normalization effect. Temporally defined cues are held to be perceived in relation to the prevailing rate of speech. In CV syllables information about rate of speech may be provided by the duration of the vowel. Recently this account has been challenged. The effect of vowel duration on consonant identification disappears if invariant cues are built into stimuli [P. C. Shinn, S. E. Blumstein., and A. Jongman, Percept. Psychophys. 38, 397–407 (1985)]. A rate effect also occurs with nonspeech stimuli, suggesting that it is not entirely phonetic [R. L. Diehl and M. A. Walsh, J. Acoust. Soc. Am. 85, 2154–64 (1989)]. The present experiments show that the rate effect returns when the stimuli of Shinn et al. (1985) are placed into a sentence. Information for rate of speech may be more reliable in sentences than in CV syllables. Another series of experiments addresses the question of whether rate effects occur for nonspeech analogs to voiced-voiceless continua.