Accuracy is not the same as vivid
A vivid screen is not necessarily an accurate one. Color accuracy means that the colors on the screen match what they are supposed to be. If a file says a shirt is a particular blue, an accurate screen shows that blue, and a less accurate one shows something a little greener, darker or more saturated.
Many people prefer a punchy image, and that is a reasonable taste for entertainment. Accuracy matters most when color is a decision: choosing a paint, matching a brand color, editing a photo, or comparing images made on different devices.
Three separate pieces
Color accuracy is built from several qualities that can each be good or poor.
Gamut. The color gamut is the range of colors a screen can show. A screen with a small gamut cannot reach very saturated reds, greens or blues. A wide gamut can show more, but only helps if the content and software are set up to use it. The article on color gamut and color spaces explains the standard boxes that content is made for. A wide gamut shown with content made for a smaller one looks oversaturated unless the system converts it.
White point. The white point is the color of the screen’s white. It can lean warm (yellowish) or cool (bluish). Because white is the reference for everything else, a shifted white tints the whole image. A separate article covers the white point and color temperature.
Tone response. This is how signal levels turn into brightness, often described by a gamma curve. A screen can have the right primary colors and white and still show wrong mid-tones if the response is off. Gamma gets its own article too.
When all three are set to the same target as the content, colors tend to look as intended. If one is wrong, colors can be off in different ways.
The basic idea of testing
Testing color accuracy is a comparison. You show the screen a known color, measure what it actually produced, and compare that with what it should have produced. The difference is the error.
A typical general procedure has these steps:
- Warm up the screen and set it to the mode and settings you intend to use. Switch off features that alter the image automatically.
- Show a series of known colors, often a set of patches that include grays, primaries and everyday colors such as skin tones.
- Measure each patch with a color-measuring instrument that is suitable for the screen type.
- Compare each result with the target and express the difference as a number.
- Summarize the errors, often as an average and a worst case.
Color error is commonly expressed with a metric that combines differences in lightness, hue and saturation into one number. Smaller means closer to the target. A common rule of thumb is that very small differences are hard to see, but exact thresholds depend on the person, the colors and the conditions, so this article does not set targets. Different purposes accept different amounts of error.
Why the eye is not enough
The eye is good at comparing two things side by side and poor at judging absolute color from memory. Surroundings affect what we see: a gray patch on a blue background looks a little warmer than the same patch on a yellow one. Our vision also adapts to a screen’s white within minutes, so a tinted screen can look normal until it is next to another.
This is why people who care about color measure. Without a reference, it is hard to know if a screen is accurate or merely pleasing. Even so, the eye can reveal a lot with a few quick checks, such as comparing gray ramps for tints or comparing the same photo on two screens.
Viewing and conditions
Several conditions change the result.
- Room light. Color and brightness are judged against the surroundings. Strongly colored walls or bright lamps shift perception. Many professional setups use neutral surroundings and controlled lighting.
- Viewing angle. On some panel types colors shift when you look from the side, which is discussed under viewing angles. Testing is usually done straight on.
- Position on the screen. Variation across the panel can make one area differ from another, a topic for screen uniformity testing.
- Warm-up and age. Colors can drift as a screen warms up and as it ages, as covered in the article on panel aging.
- Picture mode. Modes with names like vivid or cinema alter color on purpose. The article on picture modes explains why a test should use the intended mode.
Hardware, software and the pipeline
A calibration or test result describes the screen as connected to a particular computer and settings. Color passes through an application, an operating system, a graphics output and the display’s own processing. Any step may convert colors. Software that is color-aware tags images with a color space and converts as needed; software that is not may show colors incorrectly even on a good screen.
Depth of the signal also matters. Smooth gradients depend on bit depth, and a coarse path can cause visible steps even when the colors are right.
What calibration can and cannot do
Calibration adjusts the screen toward a target by changing its settings or by building a correction that the computer applies. It can usually improve white point, tone response and, within limits, color error. It cannot add colors the panel cannot physically produce, and it cannot fully fix uneven areas. Some displays store corrections in the screen itself; others rely on the computer. The practice is described in the article on calibration in professional settings.
Panel technology influences how close a screen can get. Technologies such as quantum dots are used partly to widen the range of colors, and IPS panels are often chosen for color stability from different angles, though no technology is guaranteed to be accurate without checking.
A modest approach
For ordinary use, a sensible approach is to choose a sensible mode, such as a standard or sRGB-like mode when available, avoid exaggerated modes, and judge by side-by-side comparison. For work where color decisions count, using a measuring instrument and a documented process is more dependable than judging by eye. Either way, results are only as good as the conditions, and no screen is perfect.