Spatial Separation Benefit for Speech Detection in Multi-Talker Babble-Noise with Different Egocentric Distances

Experiments performed in free field showed that a task of speech stimuli detection is solved more successfully when a speech signal and a speech-like noise source are spatially separated in distance. The spatial separation of the sources of the target speech signal and the speech-like masker improved detection of the target speech signal in noise at a low signal-to-noise ratio of -14 dB. This indicated the release from masking when the sources of the target signal and the masker were separated by a distance of 3 m, the value of which amounted to 2-3 dB. Release from masking at equal ratios of the intensity level of the target signal and the masker did not depend on which source was closer to the listener. This result provides a possibility of adjusting the auditory system to a specified azimuth angle, at which the sound wave arrives at the ears of the person.