Binaural and Monaural Localization of Sound in Two-Dimensional Space

Two experiments were conducted. In experiment 1, part 1, binaural and monaural localization of sounds originating in the left hemifield was investigated. 104 loudspeakers were arranged in a 13 x 8 matrix with 15 degrees separating adjacent loudspeakers in each column and in each row. In the horizontal plane (HP), the loudspeakers extended from 0 degrees to 180 degrees; in the vertical plane (VP), they extended from -45 degrees to 60 degrees with respect to the interaural axis. Findings of special interest were: (i) binaural listeners identified the VP coordinate of the sound source more accurately than did monaural listeners, and (ii) monaural listeners identified the VP coordinate of the sound source more accurately than its HP coordinate. In part 2, it was found that foreknowledge of the HP coordinate of the sound source aided monaural listeners in identifying its VP coordinate, but the converse did not hold. In experiment 2, part 1, localization performances were evaluated when the sound originated from consecutive 45 degrees segments of the HP, with the VP segments extending from -22.5 degrees to 22.5 degrees. Part 2 consisted of measuring, on the same subjects, head-related transfer functions by means of a miniature microphone placed at the entrance of their external ear canal. From these data, the 'covert' peaks (defined and illustrated in text) of the sound spectrum were extracted. This spectral cue was advanced to explain why monaural listeners in this study as well as in other studies performed better when locating VP-positioned sounds than when locating HP-positioned sounds. It is not claimed that there is inherent advantage for localizing sound in the VP; rather, monaural localization proficiency, whether in the VP or HP, depends on the availability of covert peaks which, in turn, rests on the spatial arrangement of the sound sources.