Eyeglasses and Eye Tracking: Lens Types, Coatings, and Tints

Vue d'ensemble

Eye trackers read the eye using invisible near-infrared (NIR) light, so what matters about a lens is how it treats NIR, which can differ from how it treats visible light. Tracking can be affected by reflections off the lens surface, by coatings and tints that block or reflect NIR, by coatings whose NIR behavior is unknown, and by lens designs that redirect or interrupt the light, such as prism lenses and multifocals with visible dividing lines.

• Least likely to cause problems: clear lenses with no coating, light dye tints (including light yellow blue-light-blocking tints), and blue-light-filtering lenses.
• Unpredictable: standard anti-reflective (AR) coatings, high-index lenses, standard sunglass tints, photochromic lenses, and multifocal lenses.
• Most likely to cause problems: mirror coatings, lenses that block or reflect infrared, AR-style coatings built to reflect NIR, prism lenses, and high-power prescriptions.

How a lens looks, including how dark it is, is not a reliable guide. The sections below explain each rating.

This article draws on lens-coating patents, Tobii Dynavox and Tobii support guidance, and two published laboratory studies. Those studies tested a small number of lenses on other trackers, so the ratings are informed estimates, not guarantees.

How glasses and eye trackers interact

The tracker lights the eye with NIR and uses a camera to find two things: the pupil, and a small bright reflection on the cornea called the glint. The glint is the reflection the tracker needs. Reflections off the surface of a lens in front of the eye are a separate thing, and they can compete with it. The light passes through any lens in front of the eye twice, once on the way in and once on the way back. A lens can interfere in several ways:

• Reflections off the lens surface. Each uncoated lens surface reflects roughly 4% to 7% of the light that hits it, depending on the lens material. This matters mainly when the lens faces the tracker nearly head-on, much like glare on a shop window that appears only from certain angles. In an effort to avoid the eye tracker from facing the lens head-on, you might notice that the eye trackers are designed at an angle below the display. Nevertheless, posture, device height, and frame tilt therefore decide whether it becomes a problem.
• Blocked or redirected NIR. Some coatings, tints, and lens materials absorb or reflect NIR deliberately. Some lens designs also redirect or interrupt it: prism lenses bend the light path, and the dividing lines in some multifocal lenses can interfere with the glint.
• Lens power. Stronger prescriptions change the apparent size and position of the eye. In one laboratory test on a Tobii tracker, lenses of ±6 diopters increased gaze error, and the −6 diopter lens also caused more lost tracking data.
• Dirt and scratches. Scratches and dirt on a lens can scatter or block light and reduce tracking accuracy.

Trace iron in glass is not a factor. Iron in ordinary glass absorbs a little NIR, but eyeglass lenses contain only trace amounts, and most lenses today are plastic with essentially none. Only heat-absorbing glass, which has iron added deliberately, blocks meaningful amounts of NIR.

Anti-reflective (AR) coatings

AR coatings reduce glare by cutting the light reflected from the lens surface, but standard AR coatings are designed for visible light only. Their effect on eye tracking therefore varies: depending on the coating, they can cut distracting reflections off the lens surface (good outcome) or add NIR reflections of their own (poor outcome).

• Designed for visible light. Lens-coating patents state that traditional AR is optimized to reduce reflection from the lens surface across the visible range, about 380 to 780 nm. Reflection from the lens surface in the NIR range is not optimized, which means it is not controlled. It does not mean NIR is blocked. Neutral to not good for eye tracking. The coating may leave NIR largely alone, or it may reflect more NIR than an uncoated lens would. Which one happens cannot be predicted from the spec sheet.
• NIR reflection off the lens can be high. A patent for a gaze-tracking supplemental lens states that typical ophthalmic lenses can reflect much more NIR than visible light. Not good for eye tracking. A lens that reflects more NIR sends more of the tracker's own light back toward the camera, where it can be mistaken for the glint.
• Some AR-style coatings are built to reflect NIR. Patents describe coatings that look like ordinary anti-glare coatings in visible light but reflect at least 70% of light at 850 nm or 940 nm. A wearer cannot identify these by eye. In one laboratory test, an AR lens that also blocks part of the infrared caused the largest loss of tracking data of any lens tested: valid tracking frames fell from about 98% with no lens to about 73%. Not good for eye tracking. These coatings reflect the tracker's NIR instead of letting it through, which can weaken the signal from the eye and add glare.
• Basic AR coatings degraded tracking in one laboratory test. In a test of single-vision lenses on a display-mounted Tobii tracker, a basic AR coating caused more lost tracking data than an uncoated lens, while two other AR coatings had no significant effect. The authors attribute the difference to basic coatings being optimized for visible light only. Not good for eye tracking (basic AR); neutral (the other two AR coatings tested).
• Some users report improvement. Some users report better tracking with AR-coated glasses. A plausible explanation is that the coating cuts reflections off the lens surface, including reflections of nearby light sources, so the tracker has less competing glare to separate from the glint. A laboratory study of a different remote tracker found better accuracy for single-vision spectacle wearers than for people without lenses, and its authors suggested that lens material and AR coatings might play a role. Neither explanation has been tested. Good for eye tracking, if the explanation holds. Less competing glare makes the glint easier for the tracker to find.

AR can either improve or degrade tracking, depending on how the specific coating treats NIR. If tracking is poor with AR-coated glasses, the coating is a possible cause. If tracking improved after AR was added, reduced lens-surface glare is a plausible explanation.

Lens compatibility

The ratings indicate likelihood, not certainty. Here, working means usable for eye control, so a lens that lowers accuracy is rated lower even if tracking continues. Clear lenses without AR are the most likely to track well, and anything built to reflect or block NIR is the least likely.

Lens or coating	Outlook	Why
Clear plastic lens, no AR (CR-39, polycarbonate, Trivex)	Likely to work	Only baseline surface reflection (about 4 to 5% per surface), and no coating layers acting on NIR. In one laboratory test on a Tobii tracker, an uncoated lens had no significant effect.
Clear glass lens, no AR, not heat-absorbing	Likely to work	Trace iron absorbs negligible NIR.
Clear or light dye tint, including a light yellow blue-light-blocking tint	Likely to work	Light dye is not designed to act on NIR, and blue-light-blocking dye absorbs short visible wavelengths, well away from NIR. A light yellow blue-light-blocking tint has been reported to track without issues. No NIR measurements were found.
Blue-light-filtering AR coating (clear lens with a faint colored sheen)	Likely to work	The blue-blocking function itself does not act on NIR. In one laboratory test on a Tobii tracker, a blue-light-filter lens had no significant effect on tracking. Other products have not been tested.
Standard AR coating (visible-light design)	May work	NIR reflection off the lens is not controlled and cannot be read from the spec sheet, and can be much higher than visible reflection. In one laboratory test, a basic AR lens caused more data loss than an uncoated lens, while two other AR lenses did not. Some users report better tracking, possibly because the coating cuts reflections off the lens surface.
High-index lens (1.67, 1.74)	May work	Reflects more per surface (about 6 to 7% uncoated) and is usually sold with AR, so the AR rows apply too.
Standard gray, brown, or green dye tint	May work	No NIR data found. Depends on the dye chemistry.
Gradient tint	May work	Same as dye tints, and only part of the lens is tinted.
Photochromic (Transitions and similar)	May work	No NIR data found. Test fully darkened and fully faded, since the same pair can behave differently.
Bifocal or trifocal with visible dividing lines	May work	Tobii lists bifocal and trifocal lenses among types that can affect accuracy and precision. The visible dividing lines can get in the way of the glint, and results depend on where the lines fall.
No-line multifocal (progressive) lens	May work	Tobii lists progressive lenses among types that can affect accuracy and precision. In one laboratory study of multifocal wearers (lens design not specified), accuracy was lower than for single-vision wearers but similar to people without lenses. Calibrate with the glasses on.
Mirror or flash coating	Unlikely to work	Built to reflect light, and can send NIR away from the eye.
AR-style coating that reflects NIR strongly ("IR protection")	Unlikely to work	Patents describe at least 70% reflectance at 850 or 940 nm by design. In one laboratory test, an infrared-protection AR lens caused the largest loss of tracking data of any lens tested.
IR-blocking, heat-absorbing, laser, welding, or IR safety lens	Unlikely to work	Made to absorb or reflect NIR.
Prism lens	Unlikely to work	Prism redirects light, which can displace the glint from where the tracker expects it. No published testing was found, so this rating rests on optics and user reports.
High-power lens (about ±6 diopters)	Unlikely to work	In one laboratory test on a Tobii tracker, ±6 diopter lenses increased gaze error by about 0.5° to 0.7°, and the −6 diopter lens also caused more lost tracking data. Reduced accuracy makes eye control harder even when tracking continues.

A lens's visible darkness is not a reliable indicator. The tracker operates in NIR rather than visible light, so a dark sunglass tint can pass NIR while a clear-looking lens may block it.

Tints, blue-light lenses, and photochromics

No published NIR measurements were found for photochromic lenses or ordinary dye tints, so specific brands and colors cannot be ranked.

• Tints are dye added to the lens. Dyes tuned for visible color are not normally designed to act on NIR, but some tints are.
• Blue-light-blocking lenses come in two forms. A light yellow tint is a dye that absorbs short visible wavelengths, well away from NIR, and is likely to track well based on reported use. A clear lens with a blue-light coating had no significant effect on tracking in one laboratory test; other products have not been tested.
• Photochromics (Transitions is the best-known brand) darken in UV light such as sunlight and clear up indoors. Their darkening acts mainly on visible light, but no NIR data was found, and a lens can behave differently freshly darkened versus faded.
• Labels to avoid. NIR blocking is added deliberately, so lenses labeled IR-blocking, heat-absorbing, laser, welding, or mirrored should be avoided.

What to try before buying new lenses

Test existing glasses and adjust device positioning before spending money on new lenses.

1. Check Track Status. Open the Track Status window and compare tracking with and without glasses.
2. Check for glare. Tobii Dynavox troubleshooting guidance suggests shining a small flashlight from the bottom of the device toward where the tracker points, with the glasses held off the face, to see whether they reflect. Never shine it in anyone's eyes. This reveals visible glare only, not NIR blocking.
3. Check for NIR blocking. Open Track Status so the tracker's infrared lamps are on, then aim a phone camera at the tracker from where the user's eyes would be. The lamps may show up as faint glowing dots. If they do, hold the lens in front of the phone camera. If the dots dim or disappear through the lens, the lens is blocking NIR. Many phone cameras filter IR, so first confirm the dots show without the lens. If they don't, a TV remote's IR light can be used the same way. Many remotes no longer use IR, and those that do typically run near 940 nm, which may not match the tracker's wavelength.
4. Try free or low-cost fixes first. Clean or replace scratched lenses, raise or reposition the device so the eyes line up with the middle of the screen, and ask an optician whether the frame's tilt can be adjusted. Tobii Dynavox guidance adds that larger lenses and frameless or partly frameless frames are less likely to block the eye, and that sparkly or extremely shiny frames should be avoided.
5. Calibrate with the glasses on. Tobii Dynavox guidance says calibration differs with and without glasses, so calibrate while the user is wearing them. Recalibrate if the user switches between wearing glasses and not.
6. Before buying new lenses, ask the lab for reflectance and transmission data at the tracker's NIR wavelength, and get the return policy in writing.

Limits of this guidance

The ratings above are estimates, and the following gaps could change them.

• No published NIR reflectance or transmission data was found for specific lens brands, AR product lines, tints, or photochromics.
• The ratings have not been validated on Tobii Dynavox devices. The two laboratory studies cited tested a small number of lenses and participants on other trackers (a Tobii 5L and another non-Tobii eye tracker).
• The laboratory studies used trackers built for general use. AAC eye control depends on accurate gaze, so accuracy losses that look small in those studies may matter more in practice.
• Reports of AR coatings improving tracking are anecdotal, and the explanation offered above has not been tested.
• No published testing of prism lenses with eye trackers was found, so that rating rests on optics and user reports.
• The NIR wavelengths used by individual tracker models are not confirmed here, so the 850 nm and 940 nm figures from the patents are reference points only.

Bench testing common lens types would convert most of the "May work" ratings into firm answers.

Sources

The patents and the Tobii Dynavox troubleshooting article were reviewed as search excerpts rather than in full. The other sources were reviewed in full.

• Tobii Dynavox: How can I position the device to get a better calibration?
• Tobii Dynavox: How do I troubleshoot eyetracking calibration issues?
• Tobii support: glasses, lenses and eye surgery
• The Impact of Ophthalmic Lens Power and Treatments on Eye Tracking Performance (Journal of Eye Movement Research)
• Effects of tracker location on the accuracy and precision of a remote eye tracker for spectacle wearers (Behavior Research Methods)
• Patent US 11169400: interferential coating with high reflection in the NIR
• Patent US 12399389: optical article with NIR-reflecting interferential coating
• Patent US 12487378: lens coatings for a gaze-tracking supplemental lens