ASA 127th Meeting M.I.T. 1994 June 6-10

4pSP12. Spectral restoration of speech by noise.

Keri R. Riener

Richard M. Warren

James A. Bashford, Jr.

Psychol. Dept., Univ. of Wisconsin---Milwaukee, Milwaukee, WI 53201

Two types of perceptual synthesis (or auditory induction) have been reported that can minimize or cancel the effects of masking upon speech: Phonemic restoration can enhance intelligibility of speech when segments are replaced or masked by noise, and contralateral induction can prevent mislateralization of speech masked at one ear [see R. M. Warren, Psychol. Bull. 96, 371--383 (1984)]. A third type of auditory induction, ``spectral restoration,'' is reported, which can enhance intelligibility when deleted regions of the speech spectrum are replaced by noise. Separate groups of listeners heard lists of sentences consisting of two widely separated bands having low intelligibility. When an appropriate noise was added to the spectral gap separating the bands, a significant improvement in intelligibility was observed. It is suggested that spectral restoration represents a mechanism capable of enhancing comprehension of speech when portions of its spectrum are masked.