Contribution of binaural masking release to improved speech intelligibility for different masker types

摘要

In situations with competing talkers or in the presence of masking noise, speech intelligibility can be improved by spatially separating the target speaker from the interferers. This advantage is generally referred to as spatial release from masking, and different mechanisms have been suggested to explain it. One proposed mechanism to benefit from spatial cues is the binaural masking release, which is purely stimulus-driven. According to this mechanism, the spatial benefit results from differences in the binaural cues of target and masker, which need to appear simultaneously in time and frequency to improve the signal detection. In an alternative proposed mechanism, the differences in the interaural cues improve the segregation of auditory streams, a process, which involves top-down processing rather than being purely stimulus-driven. Other than the cues that produce binaural masking release, the interaural cue differences between target and interferer required to improve stream segregation do not have to appear simultaneously in time and frequency. This study is concerned with the contribution of binaural masking release to spatial release from masking for three masker types that differ with respect to the amount of energetic masking they exert. Speech intelligibility was measured, employing a stimulus manipulation that inhibits binaural masking release and analysed with a metric to account for the number of better-ear glimpses. Results indicate that the contribution of the stimulus-driven binaural masking release plays a minor role while binaural stream segregation and the availability of glimpses in the better ear had a stronger influence on improving the speech intelligibility.