Voice-gender difference and spatial separation between talkers are important cues for speech segregation in multi-talker listening environments. The goal of this study was to investigate the interactions of these two cues to explore how they influence masking release in normal hearing listeners. Speech recognition thresholds in competing speech were measured, and masking release benefits by either voice-gender difference or spatial separation cues were calculated. Results revealed that the masking releases by those two cues are inversely related as a function of spatial separation, with a gender-specific difference of transition between the two types of masking release.

In multi-talker listening situations, spatial and voice-gender separation of the target from the masking speech are two major acoustic cues leading to substantial release from masking in normal-hearing listeners. A recent study demonstrated that the masking release by those two cues elicits an unequal perceptual weighting and a single point of intersection where the magnitude of masking release was the same for the two cue types, with a gender-specific difference of transition between two types of masking release [Oh et al., JASA-EL (2021)]. The goal of this study was to investigate how the interactions of these two cues are affected by various levels of the target (30, 40, and 50 dB SL). Speech recognition thresholds in the competing speech were measured, and masking release benefits by either voice-gender difference or spatial separation cues were calculated. Results revealed that the weighting of masking release from two cues is non-linearly related to the target levels. In addition, gender-related masking release was maximized when the target was presented at 40 dB SL. These results imply that a target level can be one major factor associated with maximal speech perception in noise in normal-hearing listeners. [Work supported by ASHA New Century Scholars Research Grant.]