ABSTRACT
Knowledge of auditory processing should be valuable in a search for correlates of phonetic categories that would be relatively invariant with respect to speaker and phonetic context (Blumstein & Stevens, 1979; Chistovich, Lublinskaya, Malinnikova, Ogorodnikova, Stoljarova & Zhukov, 1982; Fant, 1973; Searle, Jacobson & Rayment, 1979). Psychophysical models based on critical-band filters have been successfully used in speech analysis and recognition (Fant, 1973; Pols, van der Kamp & Plomp, 1969; Searle et al., 1979; Syrdal, 1982; Zwicker, Terhardt &; Paulus, 1979). Recent studies of the responses of auditory-nerve fibers to speech-like sounds (Delgutte, 1980; Delgutte & Kiang, 1984a, b; Hashimoto, Katayama, Murata & Taniguchi, 1975; Kiang & Moxon, 1972, 1974; Miller & Sachs, 1983; Sachs & Young, 1979; Sinex & Geisler, G; Young & Sachs, 1979) have shown that peripheral auditory processing is considerably more complex than that of linear filter-bank models. These studies used only a minimal number of stimuli from each phonetic category, therefore they could not investigate the variability of speech. In this paper, a functional model of the peripheral auditory system that simulates selected properties of the discharge rates of auditory-nerve fibers (Delgutte, 1982) is used to analyze French stop consonants produced by male and female speakers in several vowel contexts.
