Free display tests that run in your browser. Nothing is uploaded.

Colour and tone

Contrast test

Checkerboards and step patches for the two things contrast actually means: how black stays next to white, and how much tone survives at each end.

Pattern

Black and white squares across the panel, the classic simultaneous-contrast field.

How to run this test

  1. Dim the room. Not dark, just low enough that reflections are not doing the judging for you.
  2. Press Go full screen and start with the checkerboard. Look at how black the black squares stay with white squares beside them.
  3. Switch to the shadow steps and count how many patches you can tell apart from the black background.
  4. Switch to the highlight steps and count how many you can tell apart from white.
  5. Note both counts. Six or seven at each end is a display separating tone well. One or two means it is crushing that end.

What a problem looks like

Black squares turn grey next to white ones
Light from the bright squares is spreading through the panel, which is a backlight and diffuser property on an LCD. The more it happens, the lower the real simultaneous contrast. An emissive panel keeps its black regardless of what is beside it.
The first few shadow steps are invisible
Shadow detail is being crushed. The usual causes are brightness set too low, a gamma above the standard, or an RGB range mismatch between source and display. Check the range first: it is free to fix and by far the most common cause.
The top highlight steps all look white
Clipping at the top of the range. Turn contrast down until 251 and 253 separate, then stop. A contrast control set past the point where highlights separate is throwing away detail for no gain in impact.
A glow around the white areas
Blooming, and it is expected on a backlit panel with local dimming. The size of the halo tells you roughly how coarse the dimming zones are. On an OLED there should be no halo at all.
The result changes when you move your head
Viewing angle. Contrast falls off with angle on every LCD and falls off hardest on VA panels, which is where their high straight-on contrast comes from. Judge from the seat you actually use.

What to do next

Set contrast by the highlight steps and brightness by the shadow steps, in that order, then leave both alone. If shadow detail is missing at every setting, check the RGB range on the source and the display before blaming the panel.

Questions people actually ask

Why is there no contrast ratio here?

Because a contrast ratio is two luminance measurements divided by each other, and luminance is light leaving your panel. A browser has no access to that, and no API exposes it. A ratio produced by a web page would be invented, so this page gives you patterns and asks you to count instead.

What is a good number of steps to see?

There is no standard to quote, so treat it as relative. A display separating six or seven steps at each end is doing well, and one that shows only one or two is crushing. The most useful comparison is the same test on the same content before and after you change a setting.

What is the difference between this and the black level test?

The black level test uses a PLUGE ladder to find where your display stops showing shadow detail, which is a calibration question. This adds the simultaneous case: how black stays when bright content is on screen at the same time, which is where panel technology shows itself.

Does a higher contrast ratio in the specification mean a better picture?

Not reliably. Static contrast figures from manufacturers are measured in ways that vary, and dynamic contrast figures are close to meaningless because they involve turning the backlight down on dark scenes. What you see in a checkerboard is far closer to what you will notice in use.

My contrast control seems to do nothing. Is it broken?

Probably not. Many displays apply the contrast control only above a certain level, and some picture modes lock it. Try the standard or sRGB preset, then adjust while watching the highlight steps: the right setting is the highest one where 251 and 253 are still distinct.