DevicedialDead pixel or stuck pixel?
On the Devicedial dead pixel test, a dot that stays black on a full white screen is a dead pixel, and a dot that stays lit โ red, green, blue or white โ on a full black screen is stuck. Clean the screen before deciding anything, because grease and dust look exactly like a defect until they smear.
Then run the sequence full screen: white first, then black, then red, green and blue, noting where the same dot appears each time. Dead means the subpixels receive no drive and never light again, so nothing brings them back. Stuck means one or more subpixels are jammed on, and cycling colours quickly over the spot sometimes frees it โ anywhere from a few minutes to an hour, with no guarantee and no harm in trying. A dot that moves as you scroll is neither: that is dirt on the glass, or an artefact in the picture you were looking at.
What is the difference, physically?
Each pixel on an LCD is three subpixels โ red, green and blue โ with a transistor behind each one deciding how much light gets through. Two things can go wrong with that arrangement, and they look opposite.
A dead pixel receives no drive at all. Its subpixels stay shut whatever the picture asks for, so against a white field it is a black dot, and against every colour it stays black. It is a hardware failure with no software remedy.
A stuck pixel has the opposite problem: one or more subpixels are held on. Against a black field it glows, and the colour tells you which subpixel is jammed โ red, green or blue for one, cyan, magenta or yellow for two, white for all three. Because the transistor is still connected, a stuck pixel is occasionally recoverable.
That is the whole diagnosis, and it needs exactly two screens to make: one white, one black. The dead pixel test cycles both plus the primaries full screen, which is why it is worth running before you conclude anything.
Why clean the screen first?
Because dust wins this argument more often than defects do. A speck on the glass is dark, small, sharp-edged and completely convincing on a white background โ and it disappears with a dry microfibre cloth. Grease from a fingertip does the same in reverse, scattering light on a black field the way a stuck pixel does.
Two rules make cleaning safe: use a dry or barely damp microfibre cloth, and never spray liquid onto the panel. Then check again. Anything that survives a clean and appears in the same place on every colour is in the panel, not on it.
Can a stuck pixel be brought back?
Sometimes, and only a stuck one. The mechanism people rely on is simple brute force: drive the region through rapid colour changes so the affected subpixel is switched thousands of times, and it may settle back into responding. The success rate is genuinely a coin flip, and there is no way to predict it.
Give it ten to sixty minutes, full screen, and check afterwards. If nothing has changed after an hour, more hours are unlikely to help. What is worth avoiding is the pressure method โ pressing or massaging the panel over the spot โ because on an LCD that can push the liquid crystal layer out of shape permanently and turn one bad pixel into a bruise-shaped patch. Colour cycling cannot make the panel worse. Pressure can.
What if it is not a single dot at all?
Three patterns come up often enough to name, and none of them is a pixel fault.
A dot that moves when you scroll is part of the image: a compression artefact, a stray glyph or a mark in a photograph. Pixel defects never move.
A vertical or horizontal line is a row or column driver, not a pixel. It is a panel-level fault and always a warranty matter.
A dark or bright cloud that changes with viewing angle is backlight bleed or clouding, common on edge-lit panels and judged by a different standard entirely.
If what bothers you is flicker or tearing rather than a fixed dot, the fault is more likely on the timing side than the panel side โ start with the refresh rate test and the guide to a monitor stuck at 60 Hz instead.
Does the warranty cover it?
Usually only above a threshold, and the threshold is published. Every major display manufacturer puts out a pixel policy stating how many bright dots, dark dots and adjacent-pair defects a panel may have before it qualifies for replacement, referenced to the defect classes in the ISO 9241 display standards. Bright dots are generally treated more strictly than dark ones because they are more noticeable on a dark screen.
So the sequence that gets a result is: clean, test full screen, count and photograph what remains, then read the pixel policy for your exact model before you contact anyone. Going in with “there is a mark on my screen” and going in with “three bright dots, here is the photo and here is your class threshold” are very different conversations.
While the screen is in front of you, the screen resolution readout is worth a glance for a related reason: it reports CSS pixels alongside the device pixel ratio, so you can see how many physical pixels each dot on the page actually covers before you go hunting for one of them.
Frequently asked questions
Is a black dot always a dead pixel?
Not always. Clean the glass first, because a fleck of dust reads as a perfect black dot on a white background and vanishes under a microfibre cloth. If it survives cleaning and stays black through white, red, green and blue, it is a dead pixel. If it is black on white but shows colour on one of the primaries, only some of its subpixels have failed โ a partial defect, which behaves like a dead pixel in practice and is what most warranty policies count as a dark dot.
Can a stuck pixel be fixed?
Sometimes. Rapidly cycling colours over the area can shake a jammed subpixel loose, and the honest expectation is a coin flip after ten to sixty minutes of it. There is no way to know in advance, and it works on stuck pixels only โ a dead one has no drive to recover. Leave the cycle running full screen, check afterwards, and stop if nothing has changed after an hour rather than leaving it on overnight.
Does pressing on the screen fix a stuck pixel?
It is the advice that circulates most and the one worth avoiding. Pressure on an LCD can spread the liquid crystal layer permanently, turning a single defective pixel into a visible bright or dark patch, and on an OLED panel it risks the emissive layer instead. The colour-cycling method costs nothing but time and cannot make the panel worse, which is the whole argument for trying it first and stopping there.
Will one dead pixel spread?
A single failed pixel normally stays exactly as it is for the life of the panel. What does spread is a different fault: a dark cloud that grows, a line down the display, or a patch that changes shape under pressure points to the panel or its connections rather than to one pixel. Photograph the screen when you first notice a defect โ a dated picture is the only evidence you will have if the fault grows later and a warranty conversation follows.
How many dead pixels does a warranty allow?
Enough to surprise people, and the exact number depends on the manufacturer, the product line and the panel’s defect class under the ISO 9241 display standards. Bright dots on a dark background are usually treated more strictly than dark dots, because they are more visible. Look up the pixel policy for your specific model before contacting support: the makers publish these documents openly, and they tell you precisely how many defects of each type it takes to qualify for a replacement.
How the test works
Sources: