Evaluating Foveated Frame Rate Reduction in Virtual Reality for Head-Mounted Displays

TL;DR

This study investigates foveated frame-rate reduction in VR HMDs, introducing a five-region gaze-based scheme that updates peripheral regions at lower frame rates while preserving central detail. A within-subject user study (N=15) evaluates perceptual artifacts using 10 frame-rate configurations across 5 environments and two gaze tasks, employing a $5$-point Likert scale. The results show that up to $63.6\%$ of pixel rendering can be saved before users report noticeable artifacts or discomfort, with slightly higher sensitivity when actively tracking objects, highlighting the practical potential of temporally adaptive rendering in VR. The work also outlines future directions, including spatio-temporal reprojection and longer exposure studies to better understand perceptual limits and VR sickness risks.

Abstract

Foveated rendering methods usually reduce spatial resolution in the periphery of the users' view. However, using foveated rendering to reduce temporal resolution, i.e., rendering frame rate, seems less explored. In this work, we present the results of a user study investigating the perceptual effects of foveated temporal resolution reduction, where only the temporal resolution (frame rate) is reduced in the periphery without affecting spatial quality (pixel density). In particular, we investigated the perception of temporal resolution artifacts caused by reducing the frame rate dependent on the eccentricity of the user's gaze. Our user study with 15 participants was conducted in a virtual reality setting using a head-mounted display. Our results indicate that it was possible to reduce average rendering costs, i.e., the number of rendered pixels, to a large degree before participants consistently reported perceiving temporal artifacts.

Figures (12)

• Figure 1: In our user study, participants tracked objects in virtual reality under different frame rates and scene environments. Frame n shows the fully rendered environment, in frame n + 1, the periphery is unchanged and only the foveal region is rendered freshly. Forest Asset courtesy of Oleh Lila, used with permission.
• Figure 2: (Left) Eccentricity mask defining fovea/periphery image regions, with diameters chosen according to MOHANTO2022474. (Right) Image regions are color-coded according to frame rate reduction for a 12345 frame rate configuration.
• Figure 3: Illustration of the rendering pipeline. Green boxes implement foveated rendering, blue boxes implement temporal resolution reduction. The images depict an object moving from the left to the right. HMD illustration courtesy of user Juhele "Jan Helebrant" via Open Clip Art Library, used under CC0 license. Suzanne Asset courtesy of TurboSquid, used with permission.
• Figure 4: Temporal artifacts resulting from object movement (Left) and camera movement (Right) with our rendering method. HDRI Room/Attic (left) and Vestibule (right) Assets courtesy of user "Sergej Majboroda" via Poly Haven, used under CC0 license.
• Figure 5: Screenshots of the environments: (a, b) HDRI scans, (c, d) pre-rendered scenes converted into cube maps, and (e) cube map with a constant color. (f) Object shapes moving in front of the user. HDRI Street Asset courtesy of users "Dimitrios Savva" and "Jarod Guest" via Poly Haven, used under CC0 license. HDRI Room/Attic Asset courtesy of user "Sergej Majboroda" via Poly Haven, used under CC0 license. Apartment Asset courtesy of Alexis Jose Guevara, used with permission. Forest Asset courtesy of Oleh Lila, used with permission.