Enhancing spatial release from masking through individual head-related transfer function augmentation

Binaural hearing underpins effective communication in complex acoustic environments by increasing listeners' abilities to segregate concurrent sound sources. In certain conditions, augmenting binaural cues has been shown to improve speech-in-noise performance, yet existing methods largely rely on processing short temporal windows, often causing unwanted artifacts. Moreover, the perceptual implications of augmenting an individual Head-Related Transfer Function (HRTF) are not well understood. We revisit a previous augmentation approach applied directly to individual HRTFs and introduce a novel augmentation technique designed to overcome earlier limitations. Both strategies are evaluated using auditory models and behavioral tests employing a speech-on-speech spatial release from the masking paradigm, with a target at the listener’s midline and symmetrical maskers at varied azimuths. Initial findings indicate that augmentations enhance spatial release from masking relative to non-augmented individual HRTFs, though auditory models tend to overestimate this improvement. Conventional metrics do not seem to be able to account for differences arising from these augmentations, suggesting that listener familiarity with individual cues influences observed intelligibility gains. Our proposed method also mitigates the spatial spreading associated with previous approaches.