Monitor test

Twelve checks, in an order that finds the expensive faults first and the settings faults last. Each one looks for a single thing — a pixel that does not light, light leaking at the edges, a shadow that trails a moving object — and says whether what it found is a defect, a setting, or the panel being what it is. Runs in the browser; nothing to install and nothing uploaded.

Start the monitor testRuns entirely in your browser

Nothing checked yet

The run, in order

Screen resolution and display detailsSee your real screen resolution, pixel ratio, aspect ratio, colour depth and colour gamut. No install, nothing uploaded.
Sharpness testBlurry text has three causes and none is the panel: chroma subsampling, non-native scaling, or BGR subpixels. Four patterns show which you have. In the browser.
Dead pixel testFill your screen with solid colours and find dead, stuck or hot pixels. Runs in the browser, nothing to install.
Backlight bleed testFill the screen with black and check the edges and corners for light leaking through. Tells backlight bleed apart from normal IPS glow.
Screen uniformity testFlat grey fields at five levels show patches, bands, a dark corner or a colour tint — the dirty screen effect a black field hides and a panning shot reveals.
Contrast testTwo ladders of near-black and near-white patches show whether your monitor clips blacks or highlights. Black crush and white clipping, checked free.
RGB range testBlacks grey and washed out, or shadow detail gone? The signal and the screen disagree about 0-255 or 16-235. Three patterns say which, and how to fix it.
Colour testSolid colours, stepped grey and RGB ramps to reveal banding, tinting and uneven colour on any screen. Runs in the browser.
Refresh rate testMeasure the refresh rate your monitor is really running at, plus frame time and skipped frames. No install, results in three seconds.
Motion blur testThree patterns at four speeds show how much a moving object smears on your screen, and whether a higher refresh rate would help. Runs in the browser.
Ghosting testMoving blocks on fifteen grey-to-grey transitions show where your panel trails, and whether the fault is weak overdrive or overshoot. Runs in the browser.
Flicker testFind out whether your monitor backlight is strobing. PWM flicker causes eye strain and headaches long before anyone notices it visually.

Also for this device

Stuck pixel fixerFlash a movable square of rapidly changing colour over a stuck pixel to try to bring it back. Free, runs in the browser, nothing to install.

Why this order

The run is not arbitrary and it is not alphabetical. It goes from the fault that costs most to the fault that costs least, with one detour to get the eyes ready.

First, the readout. Resolution, pixel ratio, refresh rate as reported. This is not a test; it is a check that the display is being driven properly, because a monitor at the wrong resolution or a laptop panel at half its refresh rate produces symptoms — softness, smear, judder — that every later test would otherwise attribute to the panel. Fix the setting first, then judge the hardware.

Then sharpness, for the same reason. Blurry text has three causes — colour detail dropped by the cable, a picture resized on its way to the panel, smoothing tuned for the wrong subpixel order — and none of them is the panel. It sits with the readout because it is the last of the settings checks, and everything after it is about the hardware.

Then the dark tests, together. Dead pixels, backlight bleed, uniformity, black-level detail. All four are seen best in a dim room, and the eye takes a minute to adapt. Doing them consecutively means adapting once. Bleed is looked for on black and uniformity on grey — the same room, the other half of the question. They are also the expensive findings: a dead pixel or bad bleed on a new monitor is a return, and on a used one is a reason not to buy.

Then the range check, which answers the question contrast raises and cannot settle. Contrast counts how many steps survive at each end; the range test asks whether the loss stops at exactly 16 and 235, which is the signature of the graphics card and the screen disagreeing rather than of a badly set monitor. Different cause, different place to fix it.

Then colour, which needs a normal room and a normal eye, and is more often a setting than a defect.

Then motion. Refresh rate measured, motion blur, ghosting. These need the most attention and are the ones people are least sure how to read, which is why they come once the easy ones are done and the visitor has the habit of the sequence.

Flicker last, because it is the one test that can be unpleasant, and because its result is a decision about eye strain rather than about the panel being defective.

What a browser test cannot see

It is worth being precise about the limits, because most monitor test sites are not.

A web page draws pixels and that is all. It can show the panel any pattern and let you judge what comes back. It cannot read light off the screen. So it cannot measure brightness, contrast ratio, colour accuracy, gamma, uniformity in percentages or response time in milliseconds — every one of those needs an instrument held against the glass. Where a page here mentions such a figure, it is telling you what the maker claims or what a reviewer measured, never what it measured itself.

What it can do is find defects, and separate a defect from a setting. That is most of what anyone actually needs from a monitor test.

When to run it

A new monitor, within the return window. Everything the shop could not show you.

A used one, before paying. The used-monitor checklist puts these tests in the order that matters for a purchase and says which findings should end the deal.

A laptop, new or used — the same tests, the same faults, a different decision about what a fault is worth. The used-laptop checklist covers the rest of the machine.

When something looks wrong and you want to know whether it is the panel, the cable, the settings or the content. Half the complaints about monitors are settings, and the tests here are built to tell the two apart.

Before a warranty claim, so the claim names the exact fault, the count, and the position. A dead pixel at coordinates is a claim; “the screen has a mark” is a conversation.

Running it properly

Full screen, always — every test here goes full screen on its own, and the browser’s own toolbars would otherwise sit exactly where a fault might be.

Clean the glass first. A speck of dust on a black field is indistinguishable from a stuck pixel, and there is far more dust than pixels.

Brightness up, and the room as dark as you can make it for the black tests. Then normal lighting for colour.

Turn off anything that edits the picture: night light, blue-light filters, HDR, dynamic contrast, any “eye care” or “vivid” mode. The tests look at what the panel does, and all of those sit between the test and the panel.

Then start, and let each page hand you to the next. Skipping ahead is fine; the counter keeps track, and the report at the end collects everything found.

Common questions

What can a browser monitor test actually find?

Anything that shows on the panel: dead and stuck pixels, backlight bleed and glow, crushed blacks and clipped whites, colour banding, ghosting and overshoot, PWM flicker, and the refresh rate the display is really running at. What it cannot do is measure — there is no photodiode or colorimeter in a web page, so brightness in nits, a contrast ratio, colour accuracy in ΔE and response time in milliseconds are out of reach. A page that quotes those numbers from a browser is estimating.

How long does the full run take?

About ten minutes done properly, and the pixel and bleed checks are the ones worth doing slowly. The rest are quicker. Each test hands you on to the next, and the counter shows what is left, so it can be stopped and picked up later.

Should I test a brand-new monitor?

Yes, inside the return window and before the box goes in the recycling. A dead pixel is the one defect that is entirely visible on day one and entirely invisible in the shop, and most makers only exchange a panel above a threshold — often several dead pixels, or one in the centre. Knowing exactly what you have is what turns a vague complaint into a return that gets accepted.

Why does the run start with a resolution readout?

Because the first thing to establish is that the display is being driven correctly. A monitor running below its native resolution, or at 60 Hz when it supports 144, looks soft and smeary in ways that no panel test can separate from a real fault. Thirty seconds on that page rules out the cause that is a setting before anything is blamed on the hardware.

Why are the dark tests early?

Your eyes adapt. Backlight bleed and dark-level detail are far easier to see once you have spent a minute in a dark screen and a dim room, and far harder once you have just looked at a white field. The run puts the black tests together, early, while the eye is adapted, and the bright ones after.

My laptop screen — is it the same test?

Yes, and the same faults, with one difference in what to do about them. On a monitor, a panel defect is a return or a replacement monitor. On a laptop the screen is one part of a machine, so the question becomes whether a panel swap is worth it — cheap on many models, not on some. The tests find the same things either way; the checklist for buying a used laptop deals with that decision.

Can this test an OLED?

Everything except bleed, which an OLED cannot have because there is no backlight. The pixel test still matters — OLED pixels fail too — and the dark tests are where an OLED shows what it is for. Ghosting does not apply either; any trail on an OLED is persistence rather than response time.