Analysis of different pointing methods in localization experiments of sound sources

Sound localization experiments are described which aim to analyze and compare different pointing methods for sound-source localization in virtual and real environments. In the first experiment, subjects indicated the perceived direction of sound incidence by pointing directly to the source with their hands. In the second experiment, the direction was indicated using the GELP (Gods eye localization pointing) technique, i.e., by pointing at a spherical model of the auditory space. In the third experiment, a head-pointing task was applied, that is to say the direction of localization was indicated by turning the head toward the direction of the source. All the localization tests were carried out in a virtual auditory environment. Individually measured HRTFs were used for the auralization; the stimuli (pulsed white noise) were presented via headphones. The subjects were allowed to perform small head movements in order to improve the localization capability. The results of the experiments are discussed. The systematic errors that occur when using the different pointing methods are analyzed. It is shown that the localization blur averaged over all the directions is smaller using the GELP technique, while the blur for the front directions is smaller when the subjects point directly to the source.