Monitor Test
Fifteen fullscreen test patterns for uniformity, black level, gamma, banding, sharpness, colour and geometry — drawn locally in your browser, nothing uploaded.
Loading pattern…
Est. device pixels is an estimate, not a measurement: page zoom and OS display scaling are indistinguishable from each other in devicePixelRatio, so it only equals your real resolution at 100% browser zoom (Ctrl+0). Reported colour depth is worth even less: current browsers return 24 from screen.colorDepth in practice regardless of the hardware — the spec recommends it, for compatibility and to deny fingerprinters a signal — so it tells you nothing about whether the panel or the link is 6-bit, 8-bit or 10-bit. The grey ramp screen is the honest test for that. Every pattern is drawn locally on a canvas — no camera, microphone or screen capture is involved and nothing leaves your browser.
How to run the suite
- Darken the room and wait a minute. Your eyes need roughly 60 seconds to adapt before black level, backlight bleed and shadow detail mean anything. Colour and geometry are the only screens you can judge in a bright room.
- Go fullscreen — press F or the button. Inline in the page the browser may resample the canvas, which quietly breaks the single-pixel patterns.
- Set OS display scaling to 100% and browser zoom to 100% (Ctrl+0) before you trust the gamma or sharpness screens. Anything else resamples single pixels and you end up testing the scaler.
- Step through with ← →, click anywhere to advance, or pick a pattern from the strip. Number keys 1–9 jump straight to a pattern. F works anywhere on the page; the arrows and number keys act on the viewer once it is focused (click it once, go fullscreen, or tab to a thumbnail) so they do not steal your scrolling while you read this. In fullscreen the caption fades out after a few seconds so its dark gradient stops contaminating the field — move the mouse to bring it straight back.
- Change one OSD setting at a time and come back to the same pattern. Contrast lives on the white-level screen, Brightness (which sets black level on a monitor, not the backlight) on the black-level screen, Gamma on the gamma screen, RGB gains on the 50% grey screen.
- Press Esc to leave fullscreen at any time — that works even if the on-screen controls have faded out.
On an OLED, do not leave the colour bars, line grids or geometry screens up for more than a minute or two. High-contrast static patterns are the exact worst case for short-term image retention.
What these patterns actually measure
Every one of these screens tests a chain, not a monitor. The image travels: application → GPU output (bit depth, RGB vs YCbCr, full vs limited range) → cable and link mode → the display's scaler → the display's picture mode and internal LUT → the panel glass. A bad result tells you the chain is broken; which pattern fails tells you where to look. That is why order matters — fix resolution and signal range before you touch gamma, and fix gamma before you touch colour.
1–3. Uniformity (white, 50% grey, black)
Three flat fields, nothing else. On white you are looking for even brightness corner to corner. A smeary, blotchy, dirty-window look that seems to move with your eye is dirty screen effect — it shows up in slow camera pans and on sports broadcasts, and no OSD setting touches it. One dim edge or corner is an uneven backlight diffuser.
The 50% grey field is where a colour cast is easiest to catch. If the grey reads pink, green or blue, your RGB gains are off — drop the strongest channel's gain in the OSD (Colour → User/Custom) rather than raising the other two, which just clips the top of the range. Grey is also the best field for vignetting: mid-tones reveal it far better than 100% white does.
On black, in a dark room, three different faults look similar and have completely different fixes. Backlight bleed is light leaking past the edge seal: bright patches at the corners and edges that stay in exactly the same place no matter where you sit. IPS glow is a silvery haze over one corner that visibly changes shape and intensity as you move your head or lean sideways — it is inherent to the panel type and nothing removes it. Clouding is soft blotches across the middle, caused by uneven pressure on the panel. Lowering the backlight reduces all three; only bleed is arguably a defect worth an RMA, and only if it is severe with the backlight at a normal level.
4. Grey ramp — banding
The top strip is a continuous 0→255 ramp; the bottom strip is 32 discrete steps so you have something to compare against. The top strip should fade smoothly with no visible edges. Vertical contour bands mean the signal is being quantised somewhere: a 6-bit + FRC panel showing its dithering, an aggressive dynamic-contrast or "black equalizer" mode remapping the curve, or a broken ICC profile. Turn off dynamic contrast, local-dimming demo modes and shadow boost first, then re-check.
If the bands are coloured — stepping green or magenta rather than just changing brightness — that is a range or profile mismatch, not the panel. On Windows, open the GPU control panel and set the output dynamic range to Full for a PC monitor (Limited/16–235 is for TVs), and confirm the OS colour profile is the display's, not a stale one from another monitor.
5. Black level — 0 to 16
Nine patches at code values 0, 1, 2, 3, 4, 6, 8, 12 and 16 on pure black. Patch 0 is a control — it is identical to the background and must be invisible. Count how many of the rest you can separate from the surround in a dark room. Eight or nine is a healthy panel with black level set correctly. Only four or five and your shadow detail is being crushed: raise the OSD Brightness one notch at a time until 1 and 2 are just barely there. If all nine are obvious and blacks look grey and washed, brightness is too high — come back down. HDR modes, gamma presets and "black equalizer" all shift this, so set it in the mode you actually use.
The labels under the patches are deliberately dim. Anything brighter would raise your eyes' adaptation level and hide the very patches you are trying to find.
6. White level — 240 to 255
Patches at 255, 254, 252, 250, 247, 244 and 240 on a white field. As on the black-level screen, 255 is a control — it is the same value as the background and must be invisible; if you can see its edge, something in the chain is not passing the signal through untouched. If 254, 252 and 250 all look identical to the background, highlights are clipping — the top of the range is being crushed and every bright detail (clouds, snow, a muzzle flash) collapses into a flat white blob. Turn Contrast down until 252 separates cleanly. Other causes: a dynamic-contrast or pseudo-HDR mode, or limited-range video output being expanded on a full-range panel. Aim to see 250 and 247 distinctly; 254 is a bonus.
7. Gamma — 1.8 to 2.6
Each cell is a checkerboard of single device pixels, alternating pure black and pure white. Your eye averages it to exactly 50% luminance. Across the middle of each cell sits a solid bar whose code value would produce 50% luminance at the labelled gamma — 174 for 1.8, 180 for 2.0, 186 for 2.2, 191 for 2.4, 195 for 2.6. Step back two or three metres, or squint: the bar that disappears into its own checkerboard is your display's actual gamma.
2.2 is the target for a PC in a normally lit room, because that is what Windows, the web and sRGB assume. 2.4 (BT.1886) is the video target for a dark room. If 1.8 matches you are washed out with lifted shadows; if 2.6 matches, crushed and over-contrasty. Fix it with the OSD Gamma preset, not with brightness.
This is an estimate and you are the instrument: realistically you can resolve it to about ±0.1, and only when the checkerboard lands on exactly one physical pixel per drawn pixel. It also measures the whole chain including the OS colour profile, not the monitor's internal LUT in isolation. It is not a colorimeter and this page will never pretend it is.
8–9. Sharpness — 1px checkerboard and line grids
A full screen of alternating single pixels should look like a flat, calm grey field. What you must not see: shimmering or a moire pattern (something is scaling the image — you are not at the panel's native resolution, or the OS/GPU is resampling), a coloured green/magenta shimmer (subpixel-level scaling, or a non-RGB-stripe subpixel layout such as BGR or an OLED triangle arrangement), or sparkle and dark waves that shift when you move your head (over-sharpening).
Fix in this order: set the native resolution → set display scaling to 100% or an integer factor → set the OSD Sharpness back to its neutral midpoint (usually 50 on a 0–100 scale or 5 on 0–10) → disable "super resolution", "clarity", "detail enhancement" and any HDR effect processing.
The four-quadrant grid separates the axes. 1px verticals and 1px horizontals should look equally flat. If verticals shimmer but horizontals do not (or the reverse), the signal is being scaled on only one axis — the classic signature of a non-native resolution with aspect stretching. The 2px checkerboard in the fourth quadrant is a reference: it should always look clean, and if even that moires the whole image is being resampled.
10–12. Colour — bars, saturation sweeps, skin tones
The 100% bars (white, yellow, cyan, green, magenta, red, blue, black) should be flat, even blocks with hard edges and no gradient across each bar. Bleeding or ringing at the vertical edges is over-sharpening or a marginal cable/link mode. If the red and magenta edges look noisy, blocky or fringed while the white edges look fine, the source is running 4:2:0 chroma subsampling — set the GPU output to RGB or YCbCr 4:4:4, and check the cable is rated for your resolution and refresh rate. The pure red, green and blue fields underneath also make individual dead or mis-lit subpixels obvious: a dead green subpixel is a black dot that only appears in the green field.
The sweeps run white → primary (saturation) on the top three rows and black → primary (brightness) on the bottom three. Each row should be perfectly smooth. Steps in a single channel point at that channel's gain or LUT clipping — check your RGB gains and any custom ICC profile. If a row reaches full saturation a third of the way along and then sits flat, you are in an over-saturated picture mode (Vivid/Game, or a wide-gamut panel stretching sRGB content across P3) — switch to sRGB or Standard.
Skin tones are the most unforgiving reference there is, because everyone knows what skin looks like. These are eight representative sRGB swatches chosen to span light to deep — they are a familiarity check, not a standard, and not a calibration target. The row should read warm and believable end to end. A green or magenta cast, or a sunburnt-orange row, points at white balance: set colour temperature to 6500K/Warm and then trim RGB gains. If the two darkest swatches merge into one flat blob, shadow detail is crushed — go back to the black-level screen. Ignore this test entirely under coloured room lighting; your eyes adapt to the room, not the screen.
13. Geometry — grid and circles
A grid of exact device-pixel squares, circles that must be circular, and a 1px red border at the very edge of the screen. Three checks. The red border must be fully visible on all four edges — if it is cut off you are being overscanned, which is a TV behaviour: turn on "Just Scan", "1:1", "Full Pixel", "Screen Fit" or whatever your set calls it. The circles must be round. Eggs mean a non-native resolution being stretched, or a 16:9 signal on a 21:9 panel — set the OSD aspect to Native/1:1 and pick the native resolution in the OS. The grid lines must all be the same weight from edge to edge; thick-thin-thick variation across the screen is the scaler adding and dropping lines.
14. Viewing angle
Three uniform bands at code values 64, 128 and 191, plus a gradient. Look from dead centre first and remember what you see, then move your head about 30° left, right, up and down. TN panels invert dramatically when viewed from below — the darkest band goes silver and negative — and shift colour badly on the vertical axis. VA panels lose brightness and gamma off-axis, so the mid band washes out and a dark "crush" region follows your eye around the screen. IPS keeps the mid greys honest but picks up a silvery glow in the blacks. None of this is a fault; it is the panel technology you bought. The practical takeaway: calibrate from where you actually sit, and if a big screen looks uneven from centre, some of that is the viewing angle changing across its own width.
15. Text clarity
This screen is deliberately real browser-rendered text, not canvas: canvas text is drawn with greyscale antialiasing on most platforms, which would hide the exact thing you want to see. Look at the letter edges. Red/blue fringes are subpixel antialiasing (ClearType) assuming a horizontal RGB stripe — correct and normal on a standard LCD, wrong on a panel that is BGR, rotated to portrait, or an OLED with a non-stripe subpixel layout. On those, disable subpixel AA and use greyscale AA instead.
Fuzzy, smeared edges with no colour mean the image is being scaled (non-native resolution) or the OSD sharpness is off-centre — go back to the checkerboard screen. Light-on-dark text looking thinner and harder to read than dark-on-light is normal halation and affects everyone; it is much stronger on OLED, and there is no display setting that fixes it.
Before you trust any of this
- Browser colour management. Browsers treat canvas as sRGB and convert to your display profile. On a wide-gamut panel without a correct profile, "full red" here is not necessarily the panel's full red — the colour screens are useful for spotting faults, not for grading.
- Zoom and scaling ruin the pixel-level tests. Gamma and sharpness are only meaningful at 100% browser zoom and 100% (or integer) OS display scaling, in fullscreen, at the panel's native resolution.
- Room light and eye adaptation dominate the black-level, bleed and uniformity results. The same monitor scores wildly differently at noon and at midnight.
- Judge one screen, one setting, one change at a time. Contrast changes the white level test and the gamma test; if you chase them together you will loop forever.
- On a TV, turn the processing off first — dynamic contrast, noise reduction, motion smoothing, "AI picture". Otherwise you are testing the processor, not the panel.
FAQ
How do I test my monitor?
Run the patterns in order. Start with the three uniformity screens in a dark room to find backlight bleed and clouding, then use the black level and white level patches to set the OSD brightness and contrast, then the grey ramp for banding, the 1-pixel checkerboard for scaling and sharpness problems, and the colour and geometry screens last. Press F for fullscreen and use the arrow keys to move between patterns.
What is the difference between backlight bleed and IPS glow?
Backlight bleed is light physically leaking past the edge seal of the panel. It shows up as bright patches at the corners and edges of a black screen and it stays in exactly the same place no matter where you sit. IPS glow is a silvery haze over one corner that changes shape and brightness as you move your head or lean to one side. Bleed can be a defect worth a return; IPS glow is inherent to the panel type and no setting removes it. Lowering the backlight reduces both.
What gamma should my monitor be set to?
2.2 is the right target for a PC in a normally lit room, because that is what Windows, the web and sRGB content assume. 2.4 is the video target for a dark room and looks more contrasty. Use the gamma pattern here: the bar that blends into its surrounding checkerboard when you step back tells you what your display is actually doing, which is often not what the OSD gamma preset claims.
Why does the 1-pixel checkerboard shimmer instead of looking like flat grey?
The checkerboard is drawn one device pixel at a time, so it only averages into flat grey when each drawn pixel lands on exactly one physical pixel. Shimmer or a moire pattern means something is resampling the image: you are not running the panel at its native resolution, the OS display scaling is a fractional value such as 125 or 150 percent, the browser page zoom is not at 100 percent, or the monitor sharpness control is set above neutral. Check those in that order.
Can this find dead pixels?
The solid white, black, red, green and blue screens will reveal most stuck and dead pixels, but the dedicated dead pixel test cycles through the full set of solid colours in fullscreen and is easier to scan carefully. Use this suite for calibration, banding and scaling problems, and the dead pixel test when you are inspecting a brand new panel.
Does this work on a TV, a laptop or a phone?
Yes, but read the results with the panel in mind. On a TV turn off every picture enhancement first, including dynamic contrast, noise reduction and motion smoothing, and set the picture mode to Filmmaker, Cinema or Game, otherwise you are testing the processing and not the panel. On a phone or tablet the pixel level patterns are unreliable because the browser scales the page, so stick to the uniformity, black level and colour screens. Everything runs locally in your browser and nothing is uploaded.