Foveated rendering: Difference between revisions

Fixed Foveated Rendering is the most basic implementation of the concept. It operates without any eye-tracking hardware and instead relies on the assumption that a user will predominantly look towards the center of the screen.<ref name="WikipediaFR" /><ref name="JigSpace">{{cite web |url=https://www.jig.com/spatial-computing/foveated-rendering |title=What Is Foveated Rendering? - JigSpace}}</ref> Consequently, FFR systems render a static, high-resolution region in the center of each eye's display, while the quality degrades in fixed concentric rings towards the edges.<ref name="MetaFFRvsETFR">{{cite web |url=https://developers.meta.com/horizon/blog/save-gpu-with-eye-tracked-foveated-rendering/ |title=Save GPU with Eye Tracked Foveated Rendering}}</ref>

Dynamic Foveated Rendering represents the full realization of the concept. It requires a [[head-mounted display]] with integrated eye-tracking cameras to determine the user's precise point of gaze in real-time.<ref name="VarjoWhatIs" /><ref name="WikipediaFR" /> The high-resolution foveal region is then dynamically moved to match this gaze point on a frame-by-frame basis, ensuring that the user is always looking at a fully rendered part of the scene.<ref name="TobiiDFR">{{cite web |url=https://www.tobii.com/resource-center/reports-and-papers/eye-tracking-and-dynamic-foveated-rendering |title=Eye tracking and dynamic foveated rendering - Tobii}}</ref>