Gray Uniformity Test for OLED & LCD Monitors

How to run the test

01Warm the display for 20 to 30 minutes and use the brightness you normally watch or work at.
02Enter full screen and begin at 5% gray in a dim, reflection-free room.
03Compare 1%, 2%, 10%, 20%, 50%, and 100% without changing picture settings between levels.
04Note whether bands or patches stay fixed in the same panel location and remain visible in normal content.

How to read the result

Fixed vertical stripes near black can indicate OLED banding or panel non-uniformity.
Cloudy bright areas on dark gray can come from LCD backlight or pressure variation.
A color tint that changes with viewing position may be a viewing-angle characteristic rather than a fixed uniformity defect.

Interpretation guide

Read gray fields as a map of panel consistency

Why several gray levels are necessary

Uniformity defects rarely have the same visibility at every signal level. Very dark gray can expose backlight leakage and local-dimming behavior, middle gray often makes vertical bands, dirty-screen effect, and tint variation easiest to see, while light gray can reveal edge shading or broad brightness gradients. Cycle through several levels instead of choosing the most dramatic frame. A finding that persists through a useful range of tones is more likely to affect real documents, skies, sports fields, and camera pans than an effect visible at only one extreme.

Brightness variation and tint variation are different

A region may be lighter or darker without changing hue, or it may lean green, magenta, blue, or warm while retaining similar brightness. Describe these separately. Luminance non-uniformity can arise from backlight distribution, optical films, panel pressure, or local-dimming decisions. Color non-uniformity can have a different spatial pattern and is especially relevant to editing neutral images. Do not use a grayscale photograph alone to evaluate tint, and do not assume that reducing global brightness will remove a spatial color shift.

Panel technology and viewing angle affect the map

On LCDs, the apparent center and edges can change as your head moves because light passes through the liquid-crystal layer at different angles. Large screens viewed from close range naturally present greater angles at the corners. OLED and other emissive displays avoid a conventional backlight, but may still show near-black banding, tint differences, compensation-cycle effects, or unit variation. Recheck from the intended seating position and then move slightly. If the patch follows your viewing angle rather than staying fixed, geometry and angle are contributing to the observation.

Distinguish stable structure from temporary state

Temperature, panel warm-up, automatic brightness limiting, compensation routines, and local dimming can alter a uniform field over time. Keep the pattern on screen only as long as needed, especially on emissive panels, and do not use a prolonged static test as a stress test. Repeat after normal mixed content. If a broad band is consistently visible in ordinary panning shots or neutral application backgrounds, document it at the same settings and distance. That evidence is more representative than an aggressively exposed phone photo made in total darkness.

Reliable visual testing

Build a repeatable test before judging the display

A browser pattern is a controlled visual stimulus, not a measuring instrument. It is excellent for finding visible non-uniformity, clipped tones, scaling errors, halos, color shifts, and motion artifacts under the conditions in which you actually use the screen. It cannot, by itself, report luminance in candelas per square metre, a color difference value, a gamut percentage, or a response-time number. Those claims require a meter, defined test windows, timing equipment, and a documented procedure. Use this page first as a practical screening and comparison tool: it can tell you where to look, whether a change is repeatable, and whether one setting or one display looks better than another.

Treat the entire signal path as part of the experiment. The image passes through the page, browser, operating-system color management, graphics driver, cable or wireless link, monitor input mode, picture preset, and panel before it reaches your eyes. Browser zoom, system scaling, non-native resolution, night-light software, HDR conversion, an ICC profile, dynamic contrast, local dimming, sharpening, noise reduction, and variable refresh can each alter the result. A suspicious pattern therefore does not automatically prove a panel defect. Record the active resolution, refresh rate, browser, picture mode, brightness, contrast, color temperature, HDR state, and any adaptive features so that you can reproduce the observation.

Control the room before comparing results. Allow the display to reach a stable operating state, clean the screen with the method recommended by its manufacturer, and prevent direct reflections from crossing the test area. Use moderate, stable ambient light for general checks. Use a dim room only for near-black, black-uniformity, and blooming observations, because those artifacts are easily hidden by reflections and eye adaptation. Do not judge a dark screen immediately after looking at a bright white page; give your vision a short period to adapt. Keep your normal viewing distance and eye height unless the test specifically asks you to move off-axis.

Run the first pass at the panel's native resolution with browser zoom at 100 percent. Full screen reduces distractions, but it does not cancel operating-system scaling or color processing. Inspect once from your normal position, then move closer only to identify the structure of an artifact. Change one control at a time and repeat the same pattern after every change. If several settings move together, you lose the ability to say which one helped. A useful comparison is A/B/A: observe the baseline, apply one change, then return to the baseline. If the effect disappears and returns with the setting, the relationship is more credible than a single impression.

Look for repeatability across content, positions, and inputs. Rotate or move the pattern when the tool allows it. Test a second browser, input, cable, refresh rate, or computer if the finding matters. A mark that stays in the same physical place on the panel is more likely to be display-related; a mark that follows a browser window or screenshot may originate earlier in the signal chain. Photographing the screen can help document location and severity, but a camera introduces exposure, focus, lens shading, moiré, rolling shutter, tone mapping, and white-balance errors. A photograph is evidence of what the camera captured, not a substitute for a calibrated reading.

A compact pre-test checklist

01Stabilize the display

Let brightness and temperature settle, disable screen savers, and avoid testing while the device is rapidly changing power or thermal state.

02Confirm the signal

Use native resolution, verify the intended refresh rate and RGB range, set browser zoom to 100 percent, and note whether SDR or HDR is active.

03Neutralize temporary processing

Pause night light, automatic brightness, dynamic contrast, content-adaptive brightness, and vendor enhancement modes unless one of them is the feature being tested.

04Control the environment

Remove glare, keep ambient light stable, sit square to the screen, and use a darker room only when the test calls for near-black observation.

05Make one change

Adjust only one setting between runs, return to the original value for confirmation, and write down values instead of relying on memory.

06Repeat before concluding

Reopen the pattern, try representative real content, and check another browser or input when the consequence of a wrong conclusion would be costly.

How to turn an impression into useful evidence

Describe what you see without jumping directly to a diagnosis. Record the pattern, color or tone, screen region, viewing distance, room lighting, time after power-on, and whether the artifact changes with angle or settings. Terms such as “a pale vertical band in the left third on 30–50 percent gray” are more useful than “the panel is bad.” Severity should be tied to normal work: state whether the effect is visible only in a dark test room, during ordinary desktop use, or in real video and games. This makes comparisons between presets, firmware versions, and replacement units much more meaningful.

Escalate the method when the decision requires numbers. Photographers and designers may need a colorimeter or spectrophotometer for white point, gamma, tone response, color accuracy, and uniformity. HDR evaluation needs defined test windows and an instrument that can measure bright highlights and dark levels. Motion measurements need synchronized capture or a validated pursuit-camera method. Warranty decisions should follow the manufacturer's pixel, uniformity, and return policies rather than an unofficial universal threshold. The visual test remains valuable: it identifies the question that a formal measurement needs to answer.

What a visible symptom may mean

Fixed vertical or horizontal band

Repeat at several gray levels and with real panning content. A stable band can indicate panel or backlight non-uniformity; scaling artifacts usually move with the rendered image.

Bright corners on dark gray

Check head position, room reflections, brightness, and local dimming. LCD glow changes with angle, while backlight leakage tends to remain tied to the chassis.

Green or magenta patch

Disable night-light and dynamic color features, then compare another neutral preset. A persistent, position-specific hue change is color uniformity rather than simple brightness variation.

Patch visible only in a phone photo

Lower camera exposure and compare with the naked-eye result. Automatic processing often exaggerates gradients and vignetting, so the photograph should not define severity by itself.

What this browser test cannot measure

This is a visual screen. It does not calculate luminance uniformity percentages or compensate for browser color management, room reflections, panel compensation cycles, or camera artifacts.

Gray uniformity questions

Which gray level is best for OLED banding?+

Start near 5%, then compare 1%, 2%, and 10%. Very low levels are sensitive to room light and display processing, so confirm the pattern with dark real-world scenes.

Is some gray variation normal?+

Yes. Most consumer panels show some variation, especially near black. The practical concern is whether it is visible and distracting at a normal distance during ordinary use.

Methodology and further reading

The interpretation above follows established visual uniformity practice: EIZO asks viewers to inspect multiple gray levels for even brightness and unwanted color, while RTINGS documents a repeatable photographed and calculated black-uniformity procedure and explicitly notes unit-to-unit variation. Their laboratory numbers are not reproduced by this browser page.

EIZO Monitor Test — UniformityEIZO

Our Monitor Black Uniformity TestsRTINGS

GRAY UNIFORMITY TEST

Uniform gray makes spatial variation easier to see