Skip to content
DeviceBench

Free backlight bleed and uniformity check

Backlight bleed test for bleed, glow and clouding

Light escaping past the edge of a panel only shows itself in a dark room on a black field, so the conditions come first here and the field second: black for the leak at the rim, then 5%, 10%, 25% and 50% gray for the evenness of the surface, which black hides completely. Nine zones sit on a pad you tap — none, slight, obvious — and the verdict reads the pattern instead of the count, because marks at the corners mean light getting past the frame while marks in the middle mean clouding in the panel itself. The distinction no spec sheet prints is the one this page is built around: IPS glow shifts and fades as you move your head, and real bleed does not move at all.

  • 100% free
  • No signup
  • Black + 4 gray steps
  • 9-zone map
  • Glow vs bleed
  1. Switch the room lights off and close the curtains. A lamp behind you reflects off the glass and reads as bleed that is not there.
  2. Leave the monitor on the brightness you actually use. Winding it to 100% for the test measures a screen you never look at.
  3. Give your eyes two minutes in the dark before you judge anything — most of the sensitivity you need arrives in the first 120 seconds.
  4. Sit square to the middle of the panel at your normal distance, roughly 60 cm for a desk monitor.
Verdict

Nothing flagged yet. Walk the four gray steps before you call a panel even — plenty of screens are flawless on black and blotchy at 10%.

Judging on Black The crystal is doing everything it can to block the backlight. Whatever you can still see got past it.

Full screen is refused in this browser, and on a phone that hardly matters: a handset backlight is bonded to a panel a tenth of the area of a monitor and edge leakage is usually invisible on it. Run this on the screen you are actually judging.

How to test a monitor for backlight bleed

The room, then your head, then the zones. Doing them in that order is what makes the answer worth anything.

  1. Set the room up first, because the room decides the result

    The four conditions above the button change the outcome more than the monitor does: the same panel looks flawless under a ceiling light and alarming in a blacked-out room at midnight. Work through them before you open the field, and resist photographing anything until you have looked with your own eyes — a phone decides its own exposure and will show you a screen nobody has ever seen.

  2. Move your head before you write anything down

    With the black field up, shift 20 cm left, then 20 cm right, then lean back half a meter. Anything that brightens, fades or changes shape as you move is IPS glow: it is light scattering in the crystal layer at an oblique angle, every IPS panel has it, and no manufacturer accepts it as a fault. What stays in the same place with the same shape from every seat is backlight bleed, and that is what the zones are for.

  3. Walk the gray steps, tap the zones, copy the map

    Black finds leakage at the rim, but unevenness across the middle of a panel only appears once there is some light in the image, which is what the 5%, 10%, 25% and 50% fields are for. Tap each zone once for slight and twice for obvious as you find it, then copy the nine-character map — it names positions and severity in a form a support desk can act on, which a photograph of a dark screen never does.

Technical specifications

FieldsBlack #000000 plus four gray steps — #0D0D0D, #1A1A1A, #404040 and #808080, which is sRGB 13, 26, 64 and 128
ZonesNine in a 3 × 3 grid — four corners, four edge midpoints and the center — each rated none, slight or obvious
Verdict ruleMarks at corners only read as leakage past the frame; anything reaching the center reads as clouding or mura in the panel
Control placementFloating in the middle of the field, the one region this test never inspects, and dimmable to 20% opacity
KeyboardUp and down arrows walk the five fields, G draws the zone grid, Esc leaves
What a browser cannot measureLuminance — no page may read a light sensor, so the output is your own zone map rather than a cd/m² figure
Photographing it honestlyAuto exposure and night mode lift a dark scene by two to four stops, which is 4× to 16× the brightness in the file; lock ISO 100 and 1/60 s
Display sleepWake lock held for the whole session, because this test is several minutes of sitting still and looking at almost nothing

Frequently asked questions

How do I tell IPS glow from backlight bleed?

Move your head: glow changes and bleed does not. IPS glow is light from the backlight scattering as it leaves the crystal layer at an angle, so it depends entirely on where your eyes are — it appears as a broad silvery or bluish wash over a corner, it swings to a different corner when you move, and it shrinks as you sit further back because the viewing angle to the edges narrows. Backlight bleed is light physically escaping between the panel and the frame holding it; it is warmer in tint, tends to have a defined edge or a streak shape, and looks identical from every seat in the room. Photograph both from dead center and only the bleed will still be there in a second photo taken from one side.

How much bleed is acceptable before I send the monitor back?

There is no published threshold you can hold a maker to, which is the honest and unhelpful answer — the pixel-fault standard covers dead and stuck pixels and says nothing about backlight uniformity, and while a laboratory method for measuring uniformity does exist, almost no consumer monitor is sold against a stated figure. That leaves two practical tests. Does it intrude on what you actually watch, at the brightness and ambient light you actually use? And does it show at the panel's normal brightness rather than only in a blacked-out room with the backlight at maximum? If both answers are yes, it is worth a claim. In the EU and the UK a distance sale also carries a 14-day right to return without giving a reason, which is a far shorter road than arguing about a tolerance nobody has published.

My phone photo looks far worse than what I can see. Which one is right?

Your eyes are right and the photograph is exaggerating, usually by a factor of four or more. A camera pointed at a dark screen meters for the whole frame, opens up, and in night mode stacks several seconds of exposure — two stops of extra exposure is four times the brightness in the file, and four stops is sixteen. That is why forum photos of bleed look like torches behind the panel. To take one that means something, turn off HDR and night mode, lock the exposure at ISO 100 and 1/60 s, and shoot from the distance you sit at; the result will look much closer to what you see and much less dramatic.

Can backlight bleed appear or get worse over time?

Yes, and the mechanism is mechanical rather than electrical. The panel sits inside a frame under even pressure, and heat cycling, a knock in transit or an over-tightened bezel screw can change that pressure enough for light to find a way out along one edge. Some people fix a new streak by loosening the screws around the affected corner a quarter turn, which is a real technique, voids the warranty the moment the case is open, and is worth trying only on a monitor nobody is going to take back.

What is the difference between clouding, mura and dirty-screen effect?

They are three uneven-brightness faults that appear at different signal levels, which is why this test has gray steps and not just black. Clouding is patchy blotches visible on dark scenes, common on VA panels and usually caused by uneven pressure on the diffuser. Mura is a manufacturing unevenness of the layers themselves, faint and fine-grained, most visible in the near-black steps. Dirty-screen effect is the one that ruins panning shots: vertical or horizontal bands that only reveal themselves on a mid-gray field, which is exactly what the 50% step is for and why broadcast engineers reach for it.

My OLED shows nothing at all on the black field. Is the test working?

It is working, and finding nothing is the correct result for an OLED. There is no backlight to leak: each pixel makes its own light and a black pixel is switched off, so edge leakage and clouding are faults an OLED structurally cannot have. The gray steps are still worth walking, because near-black banding and vertical lines do occur on OLED panels, and the fault that genuinely threatens one is uneven aging rather than uneven lighting — the burn-in test is the page for that.

Should I turn the brightness up to 100% for this?

No, because a backlight at maximum makes every leak proportionally larger and you would be judging a screen you never use. Test at your normal setting first and treat that as the verdict; if you want a second data point, note the brightness level next to the zone map so the two runs can be compared. It cuts the other way too — a monitor whose panel is clean at 30% and streaked at 100% is telling you something real about the frame around it, but only the first figure describes your daily experience of it.

About where the light escapes and why it moves

Almost every LCD monitor is lit from its edge. A strip of white LEDs runs along one side — usually the bottom — and fires into a light guide plate, a sheet of acrylic engraved with a pattern that turns light traveling sideways into light traveling forwards; diffuser and prism films even it out, and the liquid-crystal panel sits on top acting as a shutter. The panel is clamped into a metal frame around its rim, and that clamp is where the trouble starts: if pressure along it is uneven, if a corner is fractionally warped, or if transit has knocked the assembly, light from the guide plate finds a path around the panel instead of through it. That is backlight bleed, and it explains the shape it usually takes — streaks and pools at the edges and corners, worst near the LED strip, always in the same place. Televisions and monitors that are lit from behind rather than from the edge behave differently: an array of zones behind the panel leaks far less at the rim and instead produces blooming, a halo around bright objects on a dark background, which is a dimming-algorithm artifact rather than a build fault.

The fault people confuse it with has nothing to do with the frame. IPS glow is the crystal layer itself scattering light when you look at it off-axis, and because the corners of a screen are the parts you view most obliquely, that is where it appears — a broad, silvery, slightly blue wash with no defined edge. Everything about it depends on geometry: it swings across the panel as you move your head, it weakens when you sit further back because the angle to the corners narrows, and two people either side of the same desk see it in different places. Bleed passes none of those tests, which makes the head-movement check the single most useful thing on this page. It also explains why VA panels are described as having less glow and more clouding, and why an OLED has neither: with no backlight behind it, there is nothing to scatter and nothing to escape, and the fault that actually threatens that panel is uneven aging, which the burn-in test is built to find.

What a browser cannot do here is measure. Luminance uniformity in a laboratory means a colorimeter placed at nine or twenty-five points across the surface, reporting each point’s deviation from the center; a panel holding everything within about 10% is considered good and the figure appears in professional reviews rather than on retail boxes. No web page has a light sensor, so what this tool provides instead is the flat field, the conditions that make judging it honest, and a structured place to put your own answer — which is more useful than it sounds, because a nine-zone map with severity per zone is something a support agent can act on and a photograph of a dark screen is not. Two neighboring faults belong elsewhere: a hard-edged dot that stays put through every color is a pixel rather than a leak, which the dead pixel checker confirms, and a tint that varies across the surface rather than a brightness that does is color uniformity, judged on the monitor color test. If you would rather cover every display fault in one sitting, the monitor test runs the full pattern sequence, and the plain black screen is what to open when you want the field without the checklist.

What leaves this page

Every number on this page is worked out by JavaScript running in the tab you are reading it in. Nothing you type, paste or open is uploaded, logged or kept, which is also why the tools carry on working after you disconnect from the network.

Nine numbers, zero to two, are the entire state this test holds, and they exist only in the tab until you clear them or reload. No photograph is taken, no camera is opened, and the zone map reaches your clipboard only when you press Copy — after which where it goes is entirely your decision.