Speech recognition in coherently-amplitude modulated noise

Abstract

Comodulation masking release is a phenomenon that improves the detectability of a masked pure tone or speech signal by addition of a coherently amplitude-modulated energy above and/or the signal frequency. The majority of the studies done on comodulation masking release have studied threshold detection for a pure tone, or speech identification at or near threshold levels with a favorable signal-to-noise ratio. Moreover, various studies examining comodulation masking release under conditions of reduced frequency selectivity in listeners with cochlear hearing impairment indicate absence or reduced comodulation of masking release effect (Hall et al., 1984; Hall and Grose, 1985; Moore et al., 1993). In the current study, the speech recognition task was studied for the CMR effect at supra-threshold level and unfavorable signal-to-noise ratio of −10 dB. Results indicated absent CMR effect at supra-threshold level for the word recognition task for normal-hearing individuals.