The present study investigates the ability of an objective speech intelligibility measure to predict the effect of spatial release from masking (SRM) in binaural spatial sound reproduction. Binaural room impulse responses with varying angle of source position measured in real rooms using a head and torso simulator were convolved with dry speech and babble noise to simulate noisy speech recorded in the rooms. Their intelligibility was objectively evaluated using the modified binaural short-time objective intelligibility (MBSTOI) by varying the angular separation between the target speech and noise. Linear mixed effect models found a significant interaction between angular separation and acoustic environment. Under anechoic condition, the difference in intelligibility scores between the co-located (0˚) and lateral angular separations (45˚, 90˚, and 135˚) was substantially larger than in the reverberant condition, indicating that SRM was more pronounced in the anechoic condition. No significant differences were observed among the lateral angular separations in both conditions. Separation at 180˚ also showed a significant increase over co-located, albeit to a lesser extent. These findings generally align with previous studies reporting SRM effects based on subjective evaluations in anechoic and reverberant settings, suggesting that MBSTOI would be a valid metric that could reproduce subjective SRM trends in real acoustic environments.
Hayashi et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: