Consonant identification was measured for a stationary and amplitude-modulated noise masker in four listeners with flat cochlear hearing loss, and four age-matched normal-hearing listeners. The masker modulation rate was systematically varied between 2 and 128 Hz. Masking release (MR), that is better identification performance in fluctuating, than in stationary noise, was highest in a masker fluctuating at 8-16 Hz in all normal-hearing listeners. In comparison, MR was only observed in two out of the four impaired listeners. In these listeners, MR was poorer than normal, and peaked at lower rates, that is 2 or 8 Hz. MR corresponded to increased reception of information for voicing, place, and manner between 2 and 64 Hz in all normal-hearing listeners. In impaired listeners, increased reception of information was mainly observed for manner, and mainly reduced for place, but these differences were not significant. For all phonetic features, MR was observed at lower masker fluctuation rates (< or =32 Hz) than in normal-hearing listeners. This study therefore shows that cochlear damage affects MR, both quantitatively and qualitatively.

The present study examined the effects of temporal-envelope expansion on speech perception. Sentence identification performance was measured in normal-hearing and hearing-impaired listeners for stationary and fluctuating noise. Two expansion schemes were used to increase the depth of the slow (<16 Hz) and fast (<256 Hz) amplitude fluctuations of the stimuli. In the first scheme, the envelope of the stimuli was raised to the power 2. In the second scheme, the high- and low-level segments of the envelope were compressed and expanded, respectively. When envelope processing was applied to speech before the addition of noise, the first form of expansion generally degraded identification, while the second form generally improved identification. When envelope processing was applied to speech after the addition of noise, both forms of expansion improved or did not affect identification scores in both groups of listeners when applied to the slowest fluctuations. When applied to the broadest range of fluctuations, both forms of expansion degraded identification. However, in hearing-impaired listeners, the second form of expansion yielded an increase in performance for fluctuating noise. This complex pattern of results will be discussed in light of previous studies on envelope expansion.