Dynamic binaural sound source localization with ITD cues: Human listeners

Abstract

In real life, human listeners rarely experience cone of confusion errors in localization of sound sources due to the limitation of interaural time difference (ITD) as a spatial hearing cue. Previous work on robotics suggested that the confusion is disambiguated as successive observation of ITD is made over time, when self-motion of the microphone system was allowed. In this behavioral study, we investigated whether horizontal/vertical planes localization with time difference is possible when self-motion is allowed for human listeners. In particular, the case of a static sound source playing low frequency pure tone signal was studied. Human listeners were seated in a rotating chair in the middle of a loudspeaker array, and a low frequency audio signal was played at an elevation. The task was to identify the elevation spatial angle under conditions when vision was allowed and denied. The hypothesis was, with vision, a much more robust observation of self-motion is available, and the report of elevation would be more accurate. The results largely agreed with the hypothesis. A similar study for robotic hearing generated similar results.