How to Test a Monitor in 6 Steps

A six-step monitor test you can run in the browser: which field to show first, why the order changes what you find, and what to do when one fails.

Six full-screen fields, shown in a fixed order, will find every fault a monitor can show you without instruments: dead pixels, stuck pixels, sub-pixel faults, backlight bleed and uneven brightness. You can run the guided monitor test now and answer six questions, or work through the same sequence by hand below. There is no instrument in this — how to test a monitor by eye comes down to changing what is on screen, one full field at a time, and knowing what each field can and cannot reveal.

The order is the part most guides leave out, and it is the part that decides whether you find anything. A pixel with all three sub-pixels switched off is invisible on a black screen. Show black first and you will walk past it.

Before you start

Four conditions, and only the first three come from a manufacturer.

Clean the display. EIZO's testing guidance is explicit about why: "You should also clean the display prior to testing, since reflected light could cause dust particles to look like defective pixels." A speck you cannot wipe off is the only kind that counts.
Warm the monitor up. EIZO states that "your monitor should already be warmed up prior to testing (ideally for 30 minutes)". A panel checked in its first minute is not the panel you will be using.
Darken the room for the black field. EIZO recommends carrying out the test in a dark room. This matters for step 2 and almost nowhere else — edge bleed is invisible under normal lighting, and the speck hunts are fine in a lit room.
Sit at a normal viewing distance. Dell's display pixel guidelines put this at about 35 to 50 cm for a standard monitor, and about 100 to 150 cm for large-format displays. This is not a detail: manufacturers assess pixel faults at normal viewing distance, so a fault you can only find with your nose against the glass is not the fault you will be discussing with them.

Note

This guide gives you no brightness setting, deliberately. Neither EIZO's testing page nor Dell's pixel guidelines states one, and we would rather say that than invent a number. Test at the brightness you actually work at, and write it down, so a second look later is a comparison rather than a fresh guess.

Why the order matters

Each position is forced by what the field can physically reveal. White and black are mutually blind to each other's faults, the primaries are uninterpretable without both, and grey is a different kind of looking altogether.

Run the fields in the wrong order and the tests stop being independent — each one is only interpretable in the light of the ones before it. Three things follow from that.

No single field shows every pixel fault. Dell’s own definitions are the clearest statement of this. A dark pixel is one where “all three subpixels in one pixel are permanently turned off”, and it produces “a black dot that you can see on white or colored backgrounds (but not on black backgrounds)”. A bright pixel is one where “all three subpixels in one pixel are permanently turned on”, giving “a bright white dot that’s always visible, especially on dark backgrounds”. The two faults hide on each other’s field. White is the only step in this sequence where an all-off pixel appears at all, and black is the only step where an all-on pixel or edge bleed appears.

The primaries are only meaningful after white and black. A speck on the blue field alone means one blue sub-pixel is out at that point. A speck that appeared on white and on all three primaries is a single pixel with everything off — one fault, not four. Run red, green and blue first and you finish with a list of specks and no way to tell how many pixels they represent. This is why the guided test asks you to check a coloured speck back against the white field before it reports anything.

The faults you cannot fix are the ones with a deadline. Dell’s guidance for a pixel fault is a replacement display, and it says plainly: “Do not attempt DIY fixes.” A dead pixel found on day two of a return window is a different situation from the same pixel found on day thirty. So the sequence puts the checks whose only remedy is a warranty or return claim before the ones you might reasonably attempt yourself.

Grey comes last for a different reason: it is a whole-field judgement, and the other five are speck hunts. Switching between the two kinds of looking mid-test is how uniformity problems get missed.

The six-step sequence

The six fields in order: white, black, red, green, blue, 50% grey. These are the exact values the guided test paints, and the strip above is a swatch rather than a test — a real check needs the colour filling the whole screen.

Each field goes full-screen. Look across the whole panel, not just the middle, and give each one a few seconds.

White. Look for dark or coloured specks. A pixel that is permanently off shows here as a dark dot. This is the step that matters most for a monitor still inside its return window, so take it slowly. Run the dead pixel test for a fuller set of fields, or open a plain white screen if you only want this one.
Black. Look for bright dots, and separately for glow at the edges. They are two different findings: a bright dot is a stuck pixel, and glow spreading inward from an edge or corner is backlight bleed, which is a panel characteristic rather than a pixel fault. This is the step that needs the dark room. Use the backlight bleed test for the bleed question, or a plain black screen for a quick look.
Red. Look for specks that are not red. A speck here means the red sub-pixel is out at that point.
Green. Look for specks that are not green. Green is the channel the eye is most sensitive to, so this is usually the primary where a marginal fault becomes obvious.
Blue. Look for specks that are not blue.
50% grey. Stop hunting for specks and look at the field as a whole. Are some areas brighter than others? Is there a colour cast towards one edge? Patchiness here points to uneven backlighting or panel variance, not to a pixel fault. The screen uniformity test breaks the field into a grid of patches, which makes a gradual falloff much easier to see than a single flat field does.

What each result means

This is the table to keep. Find the row that matches what you saw, and the last column tells you what that result actually obliges you to do.

What you saw	Step	Most likely fault	What to do next
Dark dot on white, still dark on all three primaries	1, 3-5	Dead pixel — all three sub-pixels off	One fault, not four. Photograph it and check your warranty and return position while the window is open.
Bright white dot on black	2	Stuck pixel — all three sub-pixels on	Document it first, then read Fixable or not below before trying anything.
Small coloured dot on black	2	One sub-pixel stuck on	Same as above. It is a sub-pixel, so it will look smaller and dimmer than a full stuck pixel.
Dark speck on white and on one primary only	1, 3-5	One sub-pixel permanently off	Note which primary. That names the failed channel, which is what a support ticket wants.
Glow spreading inward from an edge or corner	2	Backlight bleed	Not a pixel fault. Position and severity decide whether it is worth raising — run the backlight bleed test in a dark room.
Patches, or a colour cast toward one side, on grey	6	Uniformity variance	The hardest fault to judge by eye. Confirm with the screen uniformity test before drawing a conclusion.
Nothing on any of the six	none	No visible fault of these kinds	It does not mean the panel is perfect. See what this sequence does not cover, below.

Six fields, seven outcomes. The last column is deliberately an action rather than a verdict — how many faults your particular manufacturer accepts is a policy question, and policies differ by brand and by model.

There is no universal number of acceptable dead pixels, and this guide does not give you one. Manufacturers publish their own thresholds and they are not interchangeable — Dell’s display pixel guidelines are one example of where to find such a policy, and yours will be on your own manufacturer’s support site. Find the one that governs your panel before you decide anything.

Fixable or not

A schematic, not a photograph: filled stripes are sub-pixels that are lit, outlined stripes are sub-pixels that are not. The pattern is what decides which field the fault shows up on, and therefore which step in the sequence finds it.

The distinction that matters is whether the sub-pixels are stuck on or stuck off, because that is what decides whether changing the image on screen can do anything at all.

A dead pixel is dark because its sub-pixels are not lighting. Nothing you display can drive a sub-pixel that is not responding, which is why manufacturers handle this as a hardware matter. A stuck pixel is the opposite case — sub-pixels that are lit when they should not be — and that is the case a colour-cycling tool is aimed at. It is an attempt, not a procedure with a known success rate, and nobody should promise you one.

Before you try anything, note Dell’s position, which is representative: its pixel guidelines direct you to contact support for a replacement display and say “Do not attempt DIY fixes.” Document the fault and establish where you stand on warranty first. A replacement panel is a better outcome than a fix that does not work, and you cannot un-spend a return window.

Caution

The stuck pixel fixer flashes colours rapidly. Rapid flashing can trigger seizures in people with photosensitive epilepsy. The tool shows a warning first and will not start until you dismiss it, it keeps the flashing area bounded rather than filling the screen, and it has no fullscreen mode. If you are photosensitive, or unsure, do not run it.

Capturing evidence

Nobody covers this, and it is the step that determines whether a support conversation goes anywhere. Photograph the fault while it is still on screen — you will not want to reproduce the whole sequence later to find it again.

Shoot the fault on the field that shows it best. A dark dot photographs on white; a bright dot or a coloured dot photographs on black. Using the wrong field produces a picture of nothing.
Turn the flash off. A flash reflects off the panel and washes out exactly the small feature you are trying to record.
Brace the camera or use a tripod. A dark-room shot means a long exposure, and handheld shake smears a single pixel into invisibility.
Give the frame something for scale. A window edge, the cursor, or the edge of the panel lets someone else locate the fault in the picture instead of taking your word for where it is.
Shoot straight on. A photograph taken from an angle does not show what someone sitting in front of the monitor sees, and edge glow in particular changes with viewing angle.
Write down the date, the model and the brightness. The date is the one that matters: return and warranty windows run from a purchase date, so a dated photograph establishes when you found it.

What this sequence does not cover

Six colour fields tell you about pixels, bleed and uniformity. They tell you nothing about the following, and no browser test will.

Response time and motion blur. Not testable with static fields.
Refresh rate. Whether a panel is genuinely running at the rate it claims is a separate question with its own method.
Colour accuracy. Judging colour against a reference needs a colorimeter — a hardware probe held against the screen. What you are doing here is looking, not measuring, and it is not calibration.
HDR behaviour. Peak brightness and tone mapping need instrumented measurement.
Burn-in. Samsung's monitor support describes burn-in as "when a static image becomes a permanent fixture on your screen, regardless of the content being displayed", and notes that its OLED monitors ship panel maintenance features — Pixel Shift on by default, and a Screen Optimization pass that runs automatically after several cumulative hours of use. If your panel is OLED, run the OLED burn-in test, which is built for that fault rather than for pixels.

Questions

How long should I let a monitor warm up before testing?

EIZO's guidance is that the monitor "should already be warmed up prior to testing (ideally for 30 minutes)". Uniformity and edge glow are the two findings most likely to read differently on a cold panel, so the warm-up matters most for steps 2 and 6.

Do I need to test in a dark room?

Only for the black field. EIZO recommends a dark room for monitor testing, and for the bleed check it is genuinely required — edge glow is invisible under normal lighting. The white, primary and grey steps are fine in a normally lit room.

What is the difference between a dead pixel and a stuck pixel?

Direction. A dead pixel has all three sub-pixels permanently off, so it is a dark dot on white and invisible on black. A stuck pixel has them permanently on, so it is a bright dot on black and hard to spot on white. That is also why the sequence shows you both fields rather than one.

How many dead pixels are acceptable?

There is no single answer, and anyone who gives you one number for every brand is guessing. Thresholds are set by each manufacturer and vary by model and product line. Find your own manufacturer's published pixel policy before you decide whether what you found qualifies.

Is backlight bleed a defect?

It is a panel characteristic rather than a pixel fault, and some amount is common on LCD panels. Severity and position are what decide whether it is worth raising. Check it on a black field in a dark room with the backlight bleed test.

Can I test a monitor from my phone?

Yes. The tools are browser pages, so anything with a browser can display the fields — which makes this practical for a second monitor, a laptop screen or a TV. You are testing whatever screen is showing the page.

Does this test refresh rate or response time?

No. Six static colour fields cannot measure either, and this guide does not claim to. It covers dead pixels, stuck pixels, sub-pixel faults, backlight bleed and uniformity, and stops there.

How this guide was made

The six-field order is the sequence our own guided monitor test runs, taken from the implementation rather than from a description of it. The reasoning for that order in "Why the order matters" is built on manufacturer definitions of what each fault looks like, cited below.

Some things were deliberately left out. We looked for a manufacturer statement on what brightness to test at and found none at either source, so this guide gives no brightness figure. We give no acceptable-dead-pixel count, because those are set per manufacturer and a single number would be wrong for most readers. And we make no claim about how often a colour-cycling tool recovers a stuck pixel, because we have no measurement to support one.

Warm-up, dark room and screen cleaning: EIZO, monitor test basics, retrieved 23 August 2026.
Bright, dark and stuck sub-pixel definitions, normal viewing distance and the manufacturer's own remedy: Dell Display Pixel Guidelines, retrieved 23 August 2026.
Burn-in definition and OLED panel maintenance features: Samsung, prevent burn-in on Samsung OLED monitor, retrieved 23 August 2026.