DeviceBenchFree online burn in test
Burn in test that tells a ghost from permanent wear
Seven flat fields take over the screen — gray at five levels plus half red and half blue — because a ghosted taskbar or a station logo appears on some of them and disappears on others, and knowing which is half the diagnosis. Flag every field where you can see a shape with edges, then run a scrolling wash or a five-color invert cycle against a clock that counts only the time the pattern was really on screen, and mark the same seven fields again. A shape that fades was image retention, a temporary charge the panel will let go of; a shape that does not move after half an hour is differential pixel wear, and no picture displayed on a panel has ever put that back.
- 100% free
- No signup
- 7 detection fields
- Wash with a clock
- OLED, plasma and LCD
Detect, wash, detect again. The difference between the two looks is the whole diagnosis.
Seven fields to check
Almost nothing is emitting, so an element that was left bright for weeks shows as a pale rectangle floating in the dark.
The field a Windows taskbar or a phone status bar usually gives itself away on first.
The single most productive field for a channel logo, a scoreboard or a game HUD that lived in one corner.
Mid gray drives all three subpixels equally, so a patch where one of them has faded reads as a tinted rectangle rather than a dim one.
A bright field reverses the polarity: here it is the elements that stayed dark for hundreds of hours that appear, as a darker patch.
One channel on its own. Red emitters last longest, so a shape visible here has had a great deal of use.
Blue is the first OLED emitter to fade, which makes this the field where wear turns up before it turns up anywhere else.
The wash and its clock
0 ms
Counted from frames actually drawn, so time in a background tab does not count toward the 10 minutes. This browser has no way to keep a screen awake, so check your display sleep timer before starting a long run.
Where this leaves you
Step through the seven fields first and flag every one where you can see a shape that should not be there. Nothing below can be decided until that list exists.
Two things a flat field here is not for. Light pooling along an edge or in a corner is the backlight escaping past the panel, which the backlight bleed test gives you a way to judge, and a single dot that stays the same color on every field is one pixel rather than a retained image — take its position to the dead pixel test. Retention is a shape with recognizable edges: a bar, a rectangle, a logo, something you can name.
How to test a screen for burn-in
Three stages and a comparison. The wash on its own proves nothing — it is the two looks either side of it that produce an answer.
Step the seven fields and flag what you can name
Dim the room, put the panel at the brightness you normally use, and walk the fields from 5% gray upward. You are looking for a shape with edges you can describe — a bar across the bottom, a rectangle in a corner, the outline of a logo — not for a general unevenness, which is a manufacturing variation every panel has. Some fields will show nothing and one will show it clearly: on most sets a taskbar first appears around 10% and a channel logo around 25%, which is exactly why there are seven of them rather than one.
Run a wash against the clock
Pick scrolling bars, the invert cycle, or both swapping every 30 seconds, choose 10, 20 or 60 minutes, and start. Every pixel then spends the run being driven hard in both directions, which is the only mechanism a picture on screen has for relaxing a retained charge. The clock counts frames that were actually delivered, so if you switch tabs it stops rather than crediting you with time the panel spent showing something else, and the screen is held awake for the duration wherever the browser supports that.
Look again, and let the difference decide
Switch the marking control to the second look and step the same seven fields once more. Fewer flags than before means the shape is moving and another run of the same length is worth doing. The same flags after half an hour or more at full brightness means you are looking at wear rather than retention, and the honest conclusion is that no further washing will help — on an OLED at that point the useful questions are about warranty terms and about keeping the next one from happening.
Technical specifications
| Detection fields | 7 — neutral at code values 13, 26, 64, 128 and 191 out of 255, plus red and blue at 128, because a shape shows on some levels and hides on others |
|---|---|
| Wash — scrolling bars | 96-px black and white bars sweeping at 700 px/s, so one black-white pair passes any given pixel every 274 ms |
| Wash — invert cycle | White, black, red, green and blue held 3 s each: one change every 3 s, or 0.33 Hz, an order of magnitude below the 3 Hz where flashing becomes a photosensitivity concern |
| Wash clock | Accumulated from delivered frames, not from the wall clock — a background tab is throttled to about 1 Hz and the count stops rather than crediting the time |
| Keeping the screen awake | Screen Wake Lock, requested when a run starts and re-taken after a tab switch; Firefox has no implementation, so there the display sleep timer is yours to set |
| Full screen on iOS | Unavailable — Safari grants it to video elements and nothing else, so the field is drawn as a fixed panel and the address bar stays above it |
| What a wash cannot do | Reverse wear: a faded OLED emitter puts out less light for the same drive permanently, and no image on the screen adds it back |
| Stored between runs | Nothing — the flagged fields and the clock reset the moment the tab closes, so a second opinion tomorrow starts from scratch |
Frequently asked questions
How do I tell image retention from burn-in?
Three things separate them, and this page is built around the third. Time is the first: retention is a ghost of something that was on screen for the last few minutes or hours, while burn-in is a ghost of something that was there for hundreds of hours. The shape is the second: a crisp outline that matches a permanent interface element — a taskbar, a scoreboard, a station logo, a HUD that never moved — points at wear, while a soft impression of whatever you were last looking at points at retention. Response is the third and the only one you can test: retention relaxes when the pixels are driven differently, so it fades over a wash or simply over a night with the set switched off, and wear does not change at all because the emitters themselves have degraded.
Can an LCD burn in?
Practically never in the way an OLED does, because nothing in an LCD wears differentially — the backlight is one uniform light source that is on whatever the picture shows, and the liquid crystal layer only shutters it. What an LCD gets instead is image persistence: a long-static image leaves a DC bias in the crystal layer, the molecules stop returning fully to their neutral orientation, and a ghost appears that looks identical to burn-in for a while. It clears. Give it a wash, or leave the panel off overnight, and an LCD ghost almost always goes completely; the rare cases that do not have usually had the same image for months on end, which is a signage problem rather than a desktop one.
Will the wash actually clear it?
It clears retention and it cannot touch wear, and it is worth understanding why the two answers are so different. Retention is a charge that has settled somewhere it should not be, so driving those pixels hard through both extremes gives it a route out; ten minutes often does it, an hour is the point beyond which more time stops helping. Wear is the emitter itself producing less light for the same drive after enough total charge has passed through it. Nothing displayed on the panel adds material back, which is the difference between a wash and a repair, and any page that promises to fix burn-in is describing the first case and selling it as the second.
Is it bad for my OLED to run this?
Used once against a symptom, no; used as routine maintenance, yes. A wash is an hour of every pixel being driven near full output, which is precisely the exposure that ages the emitters in the first place, so running one weekly as a precaution costs you more panel life than it saves. The rule worth following is to run it when you can actually see a ghost and not otherwise. Also do it at the brightness you normally use rather than at maximum: a wash at 100% brightness for an hour is a meaningful amount of an OLED's rated life, and the extra output buys you very little in how quickly retention relaxes.
My television has a pixel refresher of its own. Should I use that instead?
Use it first, because it does something a web page cannot. LG's OLED sets, for example, run a short compensation cycle automatically after roughly every four hours of cumulative viewing, once the set goes to standby, and offer a manual version lasting about an hour that the manual reserves for around every 2,000 hours of use. What that cycle does is not a repair: it measures how far each pixel has degraded and adjusts the drive voltages so the panel outputs evenly again, which in practice means bringing the healthier pixels down to meet the worn ones. It hides wear rather than reversing it, and it is still the best tool available for wear, while a wash from a browser is the better tool for retention.
The ghost is a colored smear that moves when I move my head.
Then it is not a retained image, because retention and wear are fixed to the pixels and do not care where your eyes are. A shape that shifts with your viewing angle is the panel's own angular behavior — VA panels in particular shift in gamma and tint off-axis, and IPS panels get a silvery sheen sometimes called IPS glow that grows toward the corners. A rainbow or oily-looking pattern that swims across the surface is usually the anti-glare coating or, on a laptop, a polarizer that has been stressed by pressure on the lid. None of those respond to a wash, and none of them are getting worse.
How many hours does it take to burn in an OLED?
There is no single number, but the recipe is well established: a bright, static, high-contrast element in one place, at high brightness, for hundreds of hours of cumulative use. The long-running public torture tests that ran static news channels for twenty hours a day found visible logo wear on early panels within months, and modern sets survive far longer because they fight back — pixel shifting that moves the whole image a few pixels, logo luminance reduction that dims a detected static element, and the compensation cycles above. In ordinary mixed viewing an OLED is unlikely to reach that point at all, which is why the actual risk cases are so specific: a news channel left on all day, a game with a permanent HUD played for hundreds of hours, or a desktop with a fixed taskbar used as a television.
About image retention, pixel wear and why only one of them clears
An OLED pixel makes its own light from an organic emitter, and every hour of that costs it a little output: pass enough charge through the material and it produces less light for the same drive, permanently. The decay is not even across the three colors — the blue emitter has the shortest life by a wide margin, which is why a worn area on an OLED does not simply go dim but drifts toward yellow-green as blue falls away first, and why the blue field on this page is the one where wear becomes visible earliest. Plasma had the same problem through a different chemistry, phosphor that dimmed with use, and plasma sets shipped with a scrolling white bar and a pixel orbiter built into the menu for exactly this reason. Retention is a completely different animal: a charge trapped where it should not be in an OLED stack, or in an LCD a DC bias the liquid crystal has held on to, and both of those are states rather than damage. States relax. Wear does not.
The shape usually tells you which one you have before any wash does. Wear needs hundreds of hours of the same bright element in the same place, so it arrives with crisp edges that match something permanent — the bar along the bottom of a desktop, a station logo in a corner, a health bar in a game played obsessively, the subtitle strip. Retention is softer, follows whatever was on screen recently, and often covers a much larger area. Manufacturers spend real effort keeping the first kind away: modern sets shift the entire image by a few pixels every few minutes, dim any static bright element they detect, and run compensation cycles that re-measure the panel and rebalance the drive. It is worth knowing what that last one really does, because it is widely described as a repair and is not one — it evens the panel out by driving the healthier pixels less hard, so the worn region stops standing out. That improves what you see and costs a little brightness everywhere, and it is still the best available answer to wear.
The wash on this page has one job and it is worth stating its limit plainly: it drives every pixel through both extremes for as long as you let it, which gives a trapped charge somewhere to go and does nothing whatsoever for a faded emitter. Ten minutes is enough for most retention; an hour is the point past which more time stops changing anything. Prevention is the part that actually pays — auto-hide the taskbar, let the screen saver run, drop brightness from the eye-catching factory setting, and vary what is on screen if a set spends its day on one channel. If the shape you are chasing turns out to be unevenness rather than a ghost, the screen test walks the whole pattern set and will tell you what kind of unevenness it is, while a tint that covers the entire panel rather than one region belongs to the monitor color test and is usually a setting rather than a fault. And if what you actually want is one clean field to stare at for a long time, without controls or a clock on it, use black screen or white screen.
What the flags and the clock are written to
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.
The fields you flag and the minutes the wash has run are held in this page for as long as it is open and are written nowhere else — not to storage, not to a cookie, and not to any server that could build a picture of what your television has been showing. Copy hands them to your clipboard, and closing the tab is the end of them.