Localization of virtual sounds in dynamic listening using sparse HRTFs

Reproduction of virtual sound sources that are perceptually indistinguishable from real-world sounds is impossible without accurate representation of the virtual sound source location. A key component in such a reproduction system is the Head-Related Transfer Function (HRTF), which is different for every individual. In this study, we introduce an experimental setup for accurate evaluation of the localization performance using a spatial sound reproduction system in dynamic listening conditions. The setup offers the possibility of comparing the evaluation results with real-world localization performance, and facilitates testing of different virtual reproduction conditions, such as different HRTFs or different representations and interpolation methods of the HRTFs. Localization experiments are conducted, comparing real-world sound sources with virtual sound sources using high-resolution individual HRTFs, sparse individual HRTFs and a generic HRTF.