Abstract:

Evidence in support of several generalizations about individual differences in speech recognition will be reviewed. First, the range of these abilities among normal-hearing adults appears to be reliable and large enough to be of some functional significance. Second, auditory spectral and temporal acuity, as measured with nonspeech sounds, fail to account for individual differences in speech recognition, although they may be weakly related to them. Third, a portion of the speech recognition variance among listeners is measurable with both speechreading (vision-only) and with auditory-only tasks, suggesting a central or cognitive origin. Fourth, the variety of types of distortion or transformation that fail to render speech unintelligible implies that very little necessary information is encoded in the fine structure of the speech waveform. Last, low correlations with general intelligence and with the ability to recognize environmental sounds suggest that speech recognition may be a relatively independent cognitive skill. Relevance of these conclusions to speech recognition by impaired listeners also will be considered. [Research supported by NIDCD and AFOSR.]