Two different physical processes

Emitted light and reflected light

Your screen makes colour by emitting. Three phosphors or LEDs per pixel put out red, green and blue light and your eye adds them. All three at full is white. All three off is black, or rather, whatever black your screen manages, which on an LCD is dark grey.

Paper makes colour by subtracting. White paper reflects nearly all the light that falls on it. Cyan ink absorbs red and reflects the rest. Magenta absorbs green. Yellow absorbs blue. Lay cyan over magenta and you have absorbed red and green, leaving blue. Lay all three and you have absorbed almost everything, leaving a muddy dark brown, which is why black ink exists as a fourth.

Two consequences follow immediately.

White is not a colour you can print. It is the paper. There is no white ink in a four-colour process job; the white in your logo is an absence of ink. On coloured stock, on a dark T-shirt, on transparent film, this becomes a real problem that needs a fifth station and a real white ink.

Black is limited by the paper. The darkest black on offset litho is around 5% reflectance. On a screen with a decent contrast ratio it is far lower. A photograph with deep shadows will always look flatter on paper than on the screen you approved it on, and that is physics, not a mistake.

Why two screens disagree

Even before ink enters the argument, two screens showing the same file disagree, for three separate reasons.

• Different primaries. A wide-gamut monitor’s red is a different physical red from a laptop’s. The same RGB triplet is a different colour on each.
• Different white points. Screens ship at 6500K by default, which is bluish. Print is judged under 5000K, which is warmer. The same neutral grey looks cool on one and warm under the other.
• Different transfer curves. Gamma, brightness, and whatever the manufacturer’s “vivid” preset is doing.

Nothing in that list is fixable by being careful. It is fixable by profiling, which is two lessons away.

Metamerism, which catches everyone once

Two colours can match under one light source and differ under another. This is metamerism, and it happens because the eye reduces a whole spectrum to three numbers; two different spectra can produce the same three numbers under one illuminant and different ones under another.

The practical version: a brand colour printed as a spot ink and the same colour built from CMYK can look identical in the shop’s fluorescent light and visibly different in daylight, because their spectra are different even though their appearance matched under one light.

This is why print is judged under a standard illuminant, normally D50 (5000K), in a viewing booth or a light box with D50 tubes. If you are approving a proof under a shop’s tube light, you are approving it under an unknown illuminant, and so is the client, and you may be approving different things.

A cheap and honest fix: buy a D50 bulb for one desk lamp. Around the price of a decent lunch, and it makes every colour conversation reproducible.

What to say to a client

Two sentences, before the job starts, save an argument later:

Screen colour and printed colour are made by different physical processes, so some colours on screen have no printed equivalent. I will pick from a printed chart and show you a proof on the actual stock.

Then do it. Choosing from a printed swatch removes the screen from the decision entirely, which is the only fully reliable method available at any budget. It also changes what the client is approving: a sheet of paper in their hand rather than an image on a device neither of you has measured.

Ask for the chart on the stock the job will actually run on. The same CMYK build on gloss art card, matt art and uncoated bond gives three visibly different colours, because coated paper holds ink on the surface and uncoated absorbs it. A shop that runs all three will usually have all three charts.

The limitation

None of this is settled by precision. Two presses running the same file on the same paper on the same day produce slightly different sheets; ink density drifts through a run; paper varies by batch. Print tolerances are real and the trade works in Delta E, which is the next lesson. Anyone who promises an exact match without naming a tolerance is either being imprecise or has not run enough jobs.