Harshitha Koppisetty^1,2, Laurie M. Wilcox^1,2, Robert S. Allison^1,2; ¹York University, ²Center for Vision Research

Material perception requires integration of information from multiple senses. In virtual reality (VR) simulations, good agreement between sensory inputs leads to better accuracy, however, the effect of conflicting inputs is less understood. Here, we evaluated the effects of cue conflicts between auditory and visual material information in a virtual environment. To create the auditory stimuli, impact sounds were recorded in a controlled environment with a mechanized rod hitting a panel made of glass, wood, plastic, or metal. We rendered panels made of these same materials mounted on a stand, presented in a virtual room. During testing we combined the visual material textures with each of the recorded impact sounds, to create sixteen different conditions that were interleaved randomly and viewed using a VIVE Pro VR headset. On each trial the target object was presented and struck with a rod to produce an impact sound. The participants then classified the target material, and we recorded their responses and response time. To study the effect of agency, on half the trials, the participant observed an agent striking the target (agent-interaction trials), and in the remaining trials the participant struck the target themselves (self-interaction trials). Our results show that most participants classify materials based on their auditory properties. Further, there was no difference in the classification response between the agent-interaction trials and the self-interaction trials. Interestingly, in one of the sixteen conditions, we observed a potential audiovisual illusion - when observing a wooden target paired with a plastic impact sound, participants predominantly responded ‘metallic’. In sum, attention needs to be paid to incorporating auditory cues in VR, as discordant signals can distort perceived material properties.

Acknowledgements: We acknowledge the support of the Natural Sciences and Engineering Research Council of Canada (NSERC) for grant #ALLRP 570802 – 21. Nous remercions le Conseil de recherches en sciences naturelles et en génie du Canada (CRSNG) de son soutien.