This research proposal is concerned with the auditory and phonetic coding processes that constitute the early perceptual analysis of speech. The goal of the project is a model of the perception of speech that could serve as a normative base for comparison to theories of perceptual development and perceptual disorders. This research may also contribute to the development of machine-based speech perception which could serve as an aid to the deaf. Human perception, though seemingly effortless, involves a complex coding process that transforms the continuous sound stream into a series of segments that can serve as the basis for recognizing words. The studies proposed here use a diverse set of experimental tasks, computer simulation modeling of perception, and acoustic measurements of speech to provide convergent evidence about the nature of human speech perception. Experimental procedures involving identification, discrimination, monitoring, naming, and reaction time will be used in conjunction with natural speech, synthetic speech and complex nonspeech stimuli. Additional studies will use stimuli that are ambiguous between speech and nonspeech. The choice and construction of stimuli is guided by an information processing model of speech perception implemented in a series of computer simulation models of phonetic perception. The models focus on two aspects of speech perception: a computational metric thought to reflect central auditory pattern recognition processes and the nature of the segmental representation that is the basis for word recognition. Some of the experimental studies focus on careful measurement of the acoustic qualities of speech distinctions based on a computational description of human auditory pattern recognition. Other studies, based on these analyses, use edited natural speech and synthetic speech to test hypotheses about how humans use the information in speech that was identified in the acoustic analyses. Additional studies focus on how listeners use the systematic acoustic details of speech to recognize where one syllable or word begins and another ends. The issue of what type of segments are the basis of word recognition is addressed using a priming task in which the similarity of segments in sequences of syllables (prime and target) can influence the speed with which human listeners respond to the target. The experimental studies and the results of the computational modeling will be incorporated into computer simulation models of speech perception. The simulation models, in turn, can be used to generate new predictions about human perception. The simulation model can also serve as a basis for understanding perceptual disorders and possible methods for rehabilitation.
| Newman, Rochelle S; Sawusch, James R (2009) Perceptual normalization for speaking rate III: Effects of the rate of one voice on perception of another. J Phon 37:46-65 |
| Sawusch, J R; Newman, R S (2000) Perceptual normalization for speaking rate. II: Effects of signal discontinuities. Percept Psychophys 62:285-300 |
| Turk, A E; Sawusch, J R (1996) The processing of duration and intensity cues to prominence. J Acoust Soc Am 99:3782-90 |
| Newman, R S; Sawusch, J R (1996) Perceptual normalization for speaking rate: effects of temporal distance. Percept Psychophys 58:540-60 |
| Sawusch, J R; Gagnon, D A (1995) Auditory coding, cues, and coherence in phonetic perception. J Exp Psychol Hum Percept Perform 21:635-52 |
