Simulating Olfactory Cocktail Party Effect in VR: A Multi-odor Display Approach Based on Attention

PubDate: April 2022

Teams: The University of Tokyo

Writers: Shangyin Zou; Xianyin Hu; Yuki Ban; Shin’ichi Warisawa

PDF: Simulating Olfactory Cocktail Party Effect in VR: A Multi-odor Display Approach Based on Attention

Abstract

Odor display has been a popular approach in virtual reality (VR) to enhance users’ multi-sensory experience. The existing multi-odor presentation methods in VR are mostly based on spatiality of scent sources to produce mixed scents, which will possibly compromise users’ olfactory experience because humans normally have poor ability to analyze distinct odorant components from a mixture. To tackle this problem, we present a VR multi-odor display approach that dynamically changes the intensity combinations of different scent sources in the virtual environment according to the user’s attention, hence simulating a virtual cocktail party effect of smell. We acquire the user’s gaze information as attention from the eye-tracking sensors embedded in the head mounted display (HMD), and increase the display intensity of the scent that the user is focusing on to simulate the cocktail party effect of smell, enabling the user to distinguish their desired scent source. We conducted a user study to validate the perception and experience of 2 ways of intensity settings in response to the user’s attention shift, which were a strong level of focused scent mixed with weak levels of non-focused scents and strong focused scent only. The results showed that both of the two intensity settings were able to improve olfactory experience in VR compared to the non-dynamic odor display method. Meanwhile, only the method of presenting strengthened focused scent while maintaining the weaker mixture of background scents showed significant improvement on simulating the olfactory cocktail party effect by giving the users an enhanced sense of their own olfactory sensitivity.