Tints and filters: what the evidence says

Thirty tints sold with a health or performance claim, grouped by condition. Each one shows what the maker states beside what the studies found.

I am a

Full study detail, grades, claim labels, FDA status and prices.

Evidence summaries, not a prescription. No screen or printer can show what a filter does to light. The spectrum bars draw the maker's own cut-off number on a wavelength scale and say how the maker defines it; the simulated view is computed from that curve. Grades are computed from the listed studies by one published scale and never from a maker's statement. Fit filters by trial in the patient's own lighting.

Choose a group

Or start from a condition

5 tints in this group

Tick up to three tints to see them side by side.

Light sensitivity, migraine and neurological

Migraine, blepharospasm, brain injury, photosensitive epilepsy, drug photosensitivity

Compare

FL-41 (rose)

A dye tint made at many densities; no standard curve. Blackburn's lens: 22.3% eye-weighted transmission (gray control 18.9%). PMC

Blocked band, as the maker defines it: 480 to 520 nm (TheraSpecs FL-Pro). The dye is made at many densities, so the depth of the cut varies.

Blue fill: how much light passes at each wavelength (full height = all of it). The color strip shows the color of light at each wavelength, not the lens.

Simulated viewInferred

The lens held against white paper. It lets through about 25.8% of light, computed from the curve.

An eye exam room with a blue wall, exam chair, slit lamp, plant and two screens, no lens

No lensThrough the lens

Drag the slider to compare. Every pixel on the lens side is scaled by the lens's red, green and blue transmission, computed from the curve and the CIE color functions, with brightness rescaled because your eyes adapt to a dark lens. It says nothing about comfort, glare or sharpness, and a screen cannot show the real lens.

Inferred: Lens color: shape of NoIR DOW125, a measured rose tint (25% transmission), because FL-41 is a dye made at many densities and TheraSpecs gives only the 480 to 520 nm band. NoIR data

Claim (who says what)

Claim partly supported

TheraSpecs: FL-Pro filters "between 480 and 520nm"; its glasses are "general wellness products" TheraSpecs

Evidence by condition

MixedBlepharospasm: Blackburn 2009, randomized crossover, n=30, beat gray on reading (2.19 vs 1.32, P = .032) but unmasked; a second study, n=26 + 26, found fewer blinks than a masked rose sham. PMC gray or placebo control

Adams 2006, 87 subjects: FL-41 = gray on discomfort thresholds. Wu 2019, n=21: at matched density neither FL-41 nor gray cut blink rate. PubMed PMC gray or placebo control

MixedMigraine: Good 1991, randomized, n=20 children, rose against a density-matched blue tint, 6.2 to 1.6 headaches a month with rose only; gray matched FL-41 in Adams 2006's migraine group. PubMed gray or placebo control

✕ No gainOver gray in TBI: Aguilar 2026, 46 TBI + 46 controls; FL-41 and gray raised thresholds above plano by the same amount (1.9 vs 1.9 log lux). PMC gray or placebo control

CChronic ocular pain and dry-eye glare: Reyes 2024, n=25, no control. Tien 2023, n=35: 480 nm and dual notch beat FL-41 on glare acuity. PubMed PubMed no gray control

Trade-offs and safety

At 22% transmission, category 2, so below the 75% night-driving floor. College A309

Regulator or society says

College of Optometrists A309: lenses under 75% transmission are unsuitable for night driving.

Status and cost

No FDA authorization of its own.

TheraSpecs $99 to $159; Zenni lens from $45.95 (2023 post). (seen 2026-10-13)

How often it is used

No survey found. 45% of 11 blepharospasm patients at NINDS wore FL-41.

Compare

Avulux (480 + 590 nm notch)

FDA order: blocks light at 480 nm and 590 nm. No transmission curve published. FDA order

Blocked band, as the maker defines it: Notches at 480 nm and 590 nm (FDA order). Their width is not published, so the marks are drawn narrow.

Blue fill: how much light passes at each wavelength (full height = all of it). The color strip shows the color of light at each wavelength, not the lens.

Simulated viewFrom maker numbers

The lens held against white paper. It lets through about 71.3% of light, computed from the curve.

No lensThrough the lens

From maker numbers: From the FDA order: notches at 480 and 590 nm, drawn 20 nm wide at 3% (Avulux states "up to 97%" absorbed). Width and the 85% elsewhere are assumed. FDA order

Claim (who says what)

Claim contradicted

Avulux: the lens "absorbs up to 97%" of blue, amber and red light; its trial was run by "an independent third party" and reached "clinical and statistical significance"; other migraine glasses are "non-evidence-based" Avulux

Evidence by condition

✕ No gainMigraine, prespecified outcome: Posternack 2023, randomized, masked, n=78, sponsor Avulux; 2-hour pain change -1.42 vs -1.14 with a near-clear sham, P = 0.65; gains only post hoc. registry PubMed no gray controlSponsor: Avulux

Hoggan 2016, 37 completers: 480 nm = 620 nm sham. A gray-controlled migraine trial (Utah, with Axon Optics) stopped at n=10: headache days 12 with gray vs 15 with the 480 nm film, no statistics. PMC registry gray or placebo controlPart-funded by Axon Optics

BBlepharospasm: Wu 2019, n=21, 480 and 590 nm films cut blink rate where matched gray and FL-41 did not. PMC gray or placebo control

Trade-offs and safety

FDA lists delayed response to color signals as a risk. FDA order

Regulator or society says

FDA-authorized through De Novo (class I, over the counter) to decrease the impact of light sensitivity in people diagnosed with episodic migraine, age 12 and older, alongside their usual care. The type is 510(k)-exempt, so a lens of the same type can be sold for this use without FDA premarket review.

Status and cost

FDA De Novo DEN250026, class I, over the counter, 510(k)-exempt; episodic migraine, age 12+, adjunctive.

$334.99 to $614.99. (seen 2026-10-13)

How often it is used

No independent figure.

Compare

Cobalt blue (Zeiss Z1 type)

Deep blue: passes blue, blocks red. TheraSpecs Z-Blue (a lighter relative) filters "580 to 640nm". TheraSpecs

Blocked band, as the maker defines it: 580 to 640 nm, from the lighter Z-Blue relative. The Z1 itself has no published curve.

Blue fill: how much light passes at each wavelength (full height = all of it). The color strip shows the color of light at each wavelength, not the lens.

Simulated viewInferred

The lens held against white paper. It lets through about 15.9% of light, computed from the curve.

No lensThrough the lens

Inferred: Shape of NoIR 26, a measured deep blue filter (16% transmission); the Z1 itself publishes no curve. NoIR data

Claim (who says what)

Claim overstated

Epilepsy Action (UK charity): Z1 lenses "can help prevent photosensitive seizures" Epilepsy Action

Evidence by condition

CPhotosensitive epilepsy: every study measures the EEG photoparoxysmal response in the lab, not seizures. Capovilla 2006, n=610, no control lens, response abolished in 75.9%. PubMed no gray control

In the only gray comparison (Wilkins 1999, 8 tested), gray reduced the EEG response about as much as colored lenses. Checa-Ros 2021, n=19: Z1 against four lighter blue lenses, no neutral control. PubMed PubMed gray or placebo control

Trade-offs and safety

Red and amber signals hard to see; Epilepsy Action says not suitable for driving. Epilepsy Action

Status and cost

No FDA authorization found.

Price not found. (seen 2026-10-14)

How often it is used

No count found.

Compare

Green-tinted lenses

Broadband green absorptive tint; not the narrow green band used in green-light studies. MigraLens

Inferred Shape of NoIR ML1, a measured green tint (45% transmission); MigraLens publishes no curve anywhere we found. Color name is the only link. NoIR data

Blue fill: how much light passes at each wavelength (full height = all of it). The color strip shows the color of light at each wavelength, not the lens.

Simulated viewInferred

The lens held against white paper. It lets through about 36.9% of light, computed from the curve.

No lensThrough the lens

Inferred: Shape of NoIR ML1, a measured green tint (45% transmission); MigraLens publishes no curve anywhere we found. Color name is the only link. NoIR data

Claim (who says what)

Claim unsupported

MigraLens (UK seller): relief "superior" to other filters "in more than 9/10 cases", citing an unpublished 2005 study. A competitor, TheraSpecs, says green-light therapy "is very different from glasses" MigraLens

Evidence by condition

NoneMigraine: no trial of green lenses. The green evidence is about dim green light in a dark room, e.g. Martin 2021, n=29, one-way crossover with a green LED strip; Noseda 2016 lab study. PubMed PubMed

✕ No gainFibromyalgia: Nelli 2023, pilot RCT, n=45, green vs blue vs clear glasses, no significant difference. Bain 2026 systematic review of green light: certainty "very low" for all outcomes. PubMed PubMed no gray control

Trade-offs and safety

Not studied.

Status and cost

No FDA authorization found.

Price not checked. (seen 2026-10-14)

How often it is used

No count found.

Compare

Plum after verteporfin PDT

NoIR 80 Plum, 3% luminous transmission (maker). NoIR catalog

Blocked band, as the maker defines it: Very dark (3% transmission). NoIR's own measured curve is drawn in blue.

Blue fill: how much light passes at each wavelength (full height = all of it). The color strip shows the color of light at each wavelength, not the lens.

Simulated viewMeasured curve

The lens held against white paper. It lets through about 2.5% of light, computed from the curve.

No lensThrough the lens

Measured curve: NoIR's own measured curve for 80 Plum (3% transmission). NoIR data

Claim (who says what)

Claim unsupported

NoIR: 80 Plum is an "Approved filter for PDT"; no approving body named NoIR catalog

Evidence by condition

NoneNo clinical study of any filter after PDT. The Visudyne label asks patients to avoid direct sunlight and bright indoor light for 5 days and to wear "dark sunglasses" outdoors; it names no filter. Visudyne label

Trade-offs and safety

Any dark lens meets the label's wording.

Regulator or society says

The Visudyne (verteporfin) label: 5 days of light avoidance and "dark sunglasses" outdoors. No filter is named, so "approved filter" has no approving record behind it.

Status and cost

No FDA authorization for the filter itself.

NoIR price not re-checked. (seen 2026-10-14)

How often it is used

No count found.

How grades work

ATwo or more randomized trials, masked, against a neutral (gray) or sham control tint, pointing the same way. No tint reaches A for any condition.

BOne controlled trial against a neutral or sham tint, or randomized trials without a control tint.

COpen-label trials, case series, or preference studies only.

MixedControlled studies disagree.

✕ No gainTested, and no better than the control tint or no filter on the prespecified outcome.

NoneNo study of the claimed outcome found.

Masking is close to impossible for colored lenses, because the wearer sees the color. That is why a density-matched gray control matters. Without one, a tint effect cannot be told apart from simple dimming.

Claim status:Claim supportedClaim partly supportedClaim overstatedClaim unsupportedClaim contradictedRegulatory text

"Contradicted" means a linked registry, regulator or study says the opposite of the claim. "Regulatory text" marks a row whose claim quotes a regulator, not a seller.

Compare

FL-41 (rose)

A dye tint made at many densities; no standard curve. Blackburn's lens: 22.3% eye-weighted transmission (gray control 18.9%). PMC

Blocked band, as the maker defines it: 480 to 520 nm (TheraSpecs FL-Pro). The dye is made at many densities, so the depth of the cut varies.

Blue fill: how much light passes at each wavelength (full height = all of it). The color strip shows the color of light at each wavelength, not the lens.

Simulated viewInferred

The lens held against white paper. It lets through about 25.8% of light, computed from the curve.

No lensThrough the lens

Inferred: Lens color: shape of NoIR DOW125, a measured rose tint (25% transmission), because FL-41 is a dye made at many densities and TheraSpecs gives only the 480 to 520 nm band. NoIR data

Claim (who says what)

Claim partly supported

TheraSpecs: FL-Pro filters "between 480 and 520nm"; its glasses are "general wellness products" TheraSpecs

Evidence by condition

Adams 2006, 87 subjects: FL-41 = gray on discomfort thresholds. Wu 2019, n=21: at matched density neither FL-41 nor gray cut blink rate. PubMed PMC gray or placebo control

✕ No gainOver gray in TBI: Aguilar 2026, 46 TBI + 46 controls; FL-41 and gray raised thresholds above plano by the same amount (1.9 vs 1.9 log lux). PMC gray or placebo control

CChronic ocular pain and dry-eye glare: Reyes 2024, n=25, no control. Tien 2023, n=35: 480 nm and dual notch beat FL-41 on glare acuity. PubMed PubMed no gray control

Trade-offs and safety

At 22% transmission, category 2, so below the 75% night-driving floor. College A309

Regulator or society says

College of Optometrists A309: lenses under 75% transmission are unsuitable for night driving.

Status and cost

No FDA authorization of its own.

TheraSpecs $99 to $159; Zenni lens from $45.95 (2023 post). (seen 2026-10-13)

How often it is used

No survey found. 45% of 11 blepharospasm patients at NINDS wore FL-41.

How grades work

ATwo or more randomized trials, masked, against a neutral (gray) or sham control tint, pointing the same way. No tint reaches A for any condition.

BOne controlled trial against a neutral or sham tint, or randomized trials without a control tint.

COpen-label trials, case series, or preference studies only.

MixedControlled studies disagree.

✕ No gainTested, and no better than the control tint or no filter on the prespecified outcome.

NoneNo study of the claimed outcome found.

Claim status:Claim supportedClaim partly supportedClaim overstatedClaim unsupportedClaim contradictedRegulatory text

"Contradicted" means a linked registry, regulator or study says the opposite of the claim. "Regulatory text" marks a row whose claim quotes a regulator, not a seller.