Comparable Sound Source Localization Of Plausible Auralizations And Real Sound Sources Evaluated In A Naturalistic Eye-Tracking Task In Virtual Reality

Highly plausible audiovisual virtual scenes can be created using head-tracked binaural audio renderings presented via headphones combined with a visually-realistic scene presented via virtual reality (VR) glasses. Open questions are whether these plausible auralizations enhance social presence in VR and whether they allow sound source localization comparable to real sound sources. To address these questions, we implemented an eye-tracking paradigm in VR as naturalistic tool to measure spatial attention and sound source localization. In this study, 25 participants completed localization tasks and rated social presence and spatial audio quality. We compared three highly plausible auralizations to loudspeakers and to an anchor (gaming audio engine). Participants reported higher (almost 100%) externalization rates for all plausible auralizations compared to the anchor. Sound distance perception of plausible auralizations and loudspeakers do not differ. For azimuthal error, only for audio renderings based on individual HRTFs lower accuracy was found in comparison to the loudspeaker condition. Social presence was significantly higher in loudspeaker and plausible auralizations compared to the anchor condition. Furthermore, social presence and audio quality are strongly correlated. The implementation of audio renderings is therefore suggested for VR settings in which high levels of (social) presence are relevant (for example, VR exposure therapy).