抄録
The COVID-19 pandemic forced many academic conferences to transition from in-person to virtual formats, in- creasing the use of VR social platforms. In virtual poster presentations, multiple users occupy the same space and talk simultaneously, often resulting in overlapping audio. Most VR social platforms maintain a uniform vol- ume for all users, regardless of position, leading to discomfort and reduced audibility. While some platforms attempt to simulate real-world audio by adjusting volume based on distance, we argue that directly replicating the real-world setting may not always provide the best user experience in the Metaverse. To address this chal- lenge, we developed two new audio models: the Area Model, which restricts audio to designated regions, and the Directional Model, which adjusts volume based on the user’s facing direction. These two models break the barrier of real-world audio to provide a special way to improve audibility for multiple simultaneous voices. The two proposed audio models are evaluated together with the two existing models by analyzing user comfort and level of understanding in the presentation content. To assess the comfort level, we used subjective question- naires and physiological data. Results indicate that our new proposed models could promote focus and audibility. These findings highlight the potential of advanced VR audio systems to improve immersion and usability in virtual environments.
https://ror.org/021a26605 
