LCD Colour Displays: How a Colour System Is Actually Specified

A colour requirement is the most expensive line in a display specification to get wrong, because it is the one that cannot be fixed in firmware after the parts are committed. An LCD colour display does not carry a single colour; it carries a backlight spectrum, a filter array, a liquid crystal mode and a front stack, and the colour that reaches a viewer is the product of all four. Specifying that combination is a short exercise, but it has to be done in the right order.

What does an LCD colour display actually specify?

A white point, a gamut and a tolerance.

A colour requirement is a target white point, a colour space with its coverage, and a tolerance that says how far a unit may sit from the target before it is rejected.

Those three entries replace the vague instruction to make the screen look accurate. The white point names the neutral the product is calibrated against. The colour space and coverage state which triangle the panel can reach, which is a property of the filter array and the backlight spectrum rather than of the driver. The tolerance states the acceptance limit, and it is the entry that turns a colour discussion into a testable condition. Without it, two units that differ visibly can both be described as meeting the specification, which is exactly how a colour argument starts at the sampling stage.

Where does colour sit in a display specification?

After the optical block, not before it.

Colour belongs after luminance, contrast and viewing angle, because each of those changes the colour that is measured and the order stops the requirements contradicting each other.

The reason is geometric. Luminance sets the operating point of the backlight, and the backlight spectrum is the largest single influence on the white point. Contrast and the front stack determine how much stray light reaches the measurement, and stray light shifts the coordinates toward the neutral of the ambient. Viewing angle determines where the measurement has to be taken, because an in-plane panel holds its colour across a wide cone while other modes shift, and a specification written at 0 degrees may not describe the operator's position. Writing colour last means it is stated against a defined optical state instead of against an ideal one.

How do the standard colour spaces differ?

Different triangles, different coverage expectations.

sRGB covers the desktop and industrial default, DCI-P3 extends the red and green corners for cinema and wide-gamut content, and NTSC percentages describe coverage of an older television triangle, which is why a 72% NTSC figure and a 100% sRGB figure are not interchangeable claims.

The comparison matters because a single panel can be described as covering most of sRGB and only a fraction of P3, and both statements can be true. The definitions behind those triangles come from standards bodies and industry bodies, and the measurement vocabulary follows the publications of the International Commission on Illumination with traceability conventions from NIST. Where a supplier quotes only a percentage without naming the space, the number cannot be used in a comparison.

Common misreadings of colour figures

Four readings distort a comparison.

The first is comparing a coverage percentage without the space it refers to. The second is treating a wide gamut as accuracy, when a panel can reach a large triangle and still render a colour incorrectly. The third is ignoring the measurement geometry, because a reading taken on axis says nothing about the position the operator occupies. The fourth is quoting colour before the front stack is fixed, which produces a figure that the assembled product will not reproduce.

A fifth reading is subtler and worth naming: gamut is not the same as colorfulness perceived by the user. A panel pushed for the widest possible triangle can render mid-tones in a way that looks oversaturated on content authored for a smaller space, so a wider gamut is not automatically better for an industrial interface. The useful question is which space the content was authored in and how the panel maps into it, which is a rendering-pipeline question rather than a panel-specification one.

How does the front stack change colour?

It shifts the white point and adds stray light.

A cover glass, a bonding layer and any anti-glare treatment change both the colour that reaches the viewer and the stray light the instrument collects, so a colour specification has to be verified on the assembled front.

Bonding removes two internal reflections and raises contrast, which makes the black state darker and the measured coordinates more stable. An anti-glare treatment spreads the reflection of ambient light across a wide angle, which reduces the peak reflection but raises haze, and haze mixes a small amount of white into dark content. The measurement methods for haze and for colour difference are standardised, with ASTM D1003 covering haze and ASTM D2244 covering instrumentally measured colour differences, and instrument makers publish application guidance around both, including X-Rite and Konica Minolta.

How does sourcing change with a colour requirement?

Fewer acceptable panels and tighter filtration.

A demanding colour requirement narrows the approved panel list because it depends on the backlight spectrum and the filter array, both of which are fixed at panel level rather than adjustable at module level.

That has three consequences a buyer should price in. The backlight becomes a controlled item, since a change of LED bin can shift the white point enough to fail a tight tolerance. The panel's colour variation between production lots becomes a sampling question rather than a formality. And a second source for the panel means re-qualifying colour as well as geometry, which is a larger exercise than a mechanical substitution. Where the tolerance is loose, all three costs fall, which is why the tolerance is the entry most worth negotiating rather than the coverage figure. Colour measurement service providers such as Radiant Vision Systems publish the imaging methods used for uniformity checks on large panels, and the applications that follow from the choice are collected on the site's news section.

What should you verify before design freeze?

Measure the assembled front, not the bare panel.

Before freezing, measure white point, gamut coverage and colour difference on the assembled module at the viewing position the operator will use, with the production backlight current and after a defined warm-up.

Three supporting checks prevent most sampling arguments. Record the instrument, the aperture, the geometry and the warm-up time with every measurement, because a colour difference of a few units can be produced by changing any of them. Take the measurement at the extremes of the viewing cone the product specifies, not only on axis, because a mode that holds colour well on axis can still shift at the edge of the cone. And keep a golden sample of the assembled front so a later production unit can be compared against a physical reference rather than against a memory of how the first one looked. Holding expectations for the module follow the workmanship conventions published by IPC.

Long-term availability for a colour-matched build

Colour makes every substitution larger.

When colour is tightly specified, the lifecycle risk extends beyond the panel to the backlight, the LED bin and the bonding materials, because a change in any of them can move the white point outside the accepted tolerance.

A workable plan names those items separately and states what happens when each reaches end of life: the notice period in writing, a last-time-buy quantity held against the forecast, and a documented alternative that has been evaluated for colour rather than only for fit. Where an alternative panel uses a different backlight, the emission spectrum has to be compared before the substitution is accepted, and the golden sample is the reference that makes that comparison quick. A second-source qualification carried out while the original build is in healthy supply is the control that keeps a colour-critical product buildable, and the measurement record from the original qualification is what makes the second one cheap.