Demonstration of the Binaural Gain

This demonstration illustrates the effects of binaural listening in different spatial conditions on the speech intelligibility, i.e. the room size and the positions of a speech and a noise source relative to the listener. You can choose from seven different azimuth angles for the speech and the noise source, respectively, two distances (1m or 2m) between the sound sources and the listener, and either a small or a large room ("office" or "hall"). The selected setting is displayed in the left figure below. The 'S' and the 'N' symbols represent the speech and the noise source, respectively, the circle indicates the listener, looking to the top of the figure. You can listen to the sound that a person would hear at the displayed position. Click on one of the 'Play' links for playing only the speech, the noise, or the mixture of both. It is strongly recommended to listen to the sounds by headphones in order to be able to fully experience the binaural effects. All sounds have been pre-processed (offline) by the auralisation software “tinyAVE”. The recorded speech from a male and a female speaker originates from the AUDIS catalogue. You will always hear the same non-sense sentence in order to compare the different acoustical conditions.

In addition, the reverberation time (T60) and a measure for the expected speech intelligibility computed by the Binaural Speech Intelligibility model (BSIM) for the specified spatial configuration is given at the top of the figure. (The value of this measure is the estimated speech level adjustment, relative to the given level, that would be needed to achieve 50% intelligibility.) Model values are given for normal-hearing (NH) listeners and an examplary hearing-impaired (HI) listener, whose hearing loss is characterised by the audiogram shown in the right figure. (Blue and red lines indicate hearing thresholds of left and right ear, respectively.)

When varying the spatial settings, especially the angle difference between noise and speech source, considerable changes of the perceived and estimated speech intelligibility can be observed. (Please note that changes in the spatial settings will not become effective before a new "play" command.) This demonstrates the potentially large “binaural gain”, i.e. the benefit of listening with two ears in acoustical conditions with spatially distributed sound sources. This benefit can become quite apparent when we lack binaural listening, e.g. by listening with only one ear (binaural ⇒ monaural), or, in the present case, when listening to only one sound channel of this demo (dichotic ⇒ diotic) or playing it by only one speaker (stereo ⇒ mono).