Speech perception in gated noise: the effects of spectral/temporal resolution and auditory integration

Abstract

The current investigation examined differences between hearing impaired (HI) listeners and normal-hearing listeners (NH) in the amount of masking release (MR) for sentence recognition and syllable identification tasks when they listened in modulated noise. HI and NH listeners’ performance was compared when the performance levels of the two groups were equal in steady noise and in quiet. The relationships between the amount of MR to hearing threshold and suprathreshold abilities of forward masking, auditory filter bandwidth, and auditory stream segregation/integration were also investigated. To compensate for reduced hearing sensitivity for HI listeners, the spectrum levels of both speech and noise were adjusted based on the individual hearing loss. There was no significant performance difference between NH and HI groups in steady noise and in quiet. However, the amount of MR for sentences and for CV syllables was significantly reduced for HI listeners. For sentence recognition, the amount of MR seemed to be more related to hearing sensitivity for low-to-mid frequencies and the characteristics of auditory filters. Performance for gated sentence recognition was also strongly correlated with sentence recognition in gated noise. In contrast, forward masking thresholds appear to be the main contributor to the amount of MR for syllable recognition.