Foveated rendering
Also known as: foveation, eye-tracked foveated rendering, dynamic foveated rendering
Foveated rendering is a graphics technique that renders the small area you're looking at in full detail and the rest of the image at lower quality.
It works because human eyes only see sharp detail in a tiny central region. While the technique can save a large share of graphics processing, the best version needs fast, accurate eye tracking, which adds cost to headsets.
Foveated rendering, explained at three levels
At home
Hold your thumb at arm's length and look at the nail. Only that small spot is sharp; everything around it is surprisingly fuzzy, though your brain hides it. Foveated rendering uses that trick. A VR headset draws the area you're looking at in full detail and the rest at lower quality. You don't notice the difference, but the graphics chip does far less work.
At work
Foveated rendering is a technique that renders the centre of a viewer's gaze at full resolution and peripheral areas at reduced quality. It matches how human vision works, so users don't perceive the difference. For businesses deploying VR and mixed reality, it means sharper images on the same hardware, cheaper or cooler-running devices, and longer battery life on standalone headsets.
At home
There are two kinds of foveated rendering. Fixed foveated rendering lowers detail around the edges of the lenses, where optics already blur things and you rarely look. Eye-tracked, or dynamic, foveated rendering uses cameras inside the headset to follow your pupils and moves the sharp zone wherever you look, so the blurry area can be larger. Headsets such as PlayStation VR2 and Vision Pro use the eye-tracked kind.
At work
Foveated rendering is implemented with GPU features such as variable rate shading, which shades blocks of pixels once instead of each pixel, or with multi-resolution rendering into separate regions. On standalone headsets with tile-based GPUs, edge tiles can simply be rendered at lower resolution. Eye-tracked versions need gaze data from infrared cameras at high frame rates, plus prediction, so the sharp region lands before the eye settles. Savings depend on resolution and content.
At home
Your eyes don't glide; they jump between points in rapid movements called saccades, several times a second. During a jump, your brain briefly suppresses vision, which gives the headset a tiny window to move the sharp zone. If tracking or rendering lags, you land on a blurry patch. In my opinion, that timing challenge is why eye-tracked foveation took so long to reach consumer products.
At work
Eye-tracked foveation adds a sensitive data stream. Gaze can reveal attention, interest, fatigue, and some health indicators, so platforms such as visionOS process it on-device and withhold raw gaze from apps. Enterprise deployments should document how gaze data is handled. For streamed VR, foveated encoding sends peripheral regions at lower bitrate to save bandwidth. Gains vary widely by content and resolution, so benchmark on your actual application before relying on them.
Real-world examples
At home
Lena's new headset has eye tracking. When she plays a demanding VR game, the area she's looking at stays crisp, and the game runs smoothly on hardware that would otherwise struggle.
At work
Marcus is a VR developer at a Halifax medical training company. Enabling foveated rendering let his team keep detailed surgical simulations running smoothly on standalone headsets without a PC.
How it compares
| feature | Fixed foveated rendering | Eye-tracked foveated rendering |
|---|---|---|
| Sharp area | Always centre of view | Wherever you look |
| Hardware needed | None extra | Eye-tracking cameras |
| Performance savings | Moderate | Higher |
| Privacy concerns | Minimal | Gaze data is sensitive |
Eye tracking makes foveation smarter, at the cost of extra hardware and sensitive data.
For analysts
Foveated rendering can extend the life of existing hardware for VR training programs. I'd check whether your content and devices support it before upgrading headsets for performance reasons.
History & origin
Researchers studied 'gaze-contingent' displays from the 1970s onward, matching image detail to where people look. As VR headsets reached consumers in 2016, high-resolution displays strained graphics hardware, and foveated rendering became a practical goal. Meta's standalone Quest headsets used 'fixed foveated rendering,' lowering detail at the edges of the lenses regardless of gaze. Sony's PlayStation VR2, released in February 2023, included eye-tracked foveated rendering, and Apple's Vision Pro in 2024 used eye tracking to focus detail where users look. In my view, foveated rendering is a clever case of designing around human biology instead of brute-forcing more pixels.