CMYK and the Four-Color Printing Process

A CMYK recipe is not finished when the four percentages add up. On press, ink limit, gray component replacement, paper holdout, dot gain, trapping, screen ruling, and registration error all decide whether the color prints cleanly. This lab treats CMYK like a production system instead of a simple conversion formula.

A screen makes color by adding light. Start from black, switch on red, green, and blue emitters, and the more you add the brighter it gets - all three at full give white. Paper has no emitters. It starts white, already reflecting the whole spectrum, and ink works by removing parts of that reflected light. Cyan ink absorbs red, magenta absorbs green, yellow absorbs blue. Lay all three down and you have subtracted nearly everything - the result heads toward black.

This is why the two worlds use opposite primaries. The secondaries of one are the primaries of the other: mixing red and green light makes yellow, while mixing magenta and yellow ink makes red. Get this one idea and most of print color stops being mysterious.

CMYK names the four standard process inks: Cyan, Magenta, Yellow, and Key (black). Each is laid down as a percentage from 0 to 100 - the fraction of the area its tiny dots cover. A rich blue might be 100% cyan, 80% magenta, 0% yellow, 10% black, written C100 M80 Y0 K10. Four numbers, four printing plates.

If cyan, magenta, and yellow already make black, why add a fourth ink? Because real inks are not the perfect filters the theory assumes. Three of them overprinted produce a muddy, slightly brown dark - not a convincing black - while costing three times the ink and tripling the registration problem. The black plate (called key because the other plates were registered to it) solves all of that at once.

The amount of CMY swapped for K is a tunable choice called gray component replacement (GCR) and under color removal (UCR) - we return to it in section 08. For now the takeaway is simple: K is not a color you mix, it is a practical fourth ink that carries neutrals and detail.

A printing press can only put ink down or leave paper bare - there is no "40% ink" at a single point. To reproduce a continuous gradient, print uses halftoning: a grid of dots whose size varies. Big dots that nearly touch read as a dark area; tiny dots with lots of white between them read as light. Your eye blends them at normal viewing distance into smooth tone.

Each ink gets its own grid, rotated to a different screen angle so the four grids do not collide into ugly patterns (moiré). The classic angles - roughly 15°, 45°, 75°, and 0°/90° - interleave the dots into a tight little flower called a rosette. The grid frequency is the screen ruling, measured in lines per inch (lpi): 150 lpi for general work, 200+ for fine art.

Ink dots do not land at exactly the size the file specifies. As ink hits absorbent paper it spreads, and the optical scattering of light at each dot's edge makes it look bigger still. The combined effect is dot gain: a 50% dot in the file might behave like a 65% dot on press. Midtones fill in and the whole image prints darker and lower-contrast than the file suggests.

This is not a defect to eliminate but a known, measured behavior to compensate for. Print standards quote expected dot gain for each paper class - more on uncoated and newsprint, less on coated stock - and the press profile bakes a correction curve into the separation so the printed midtones land where they should. It is the single biggest reason a file must be prepared for its specific output condition, not just "for print" in the abstract.

The naive textbook conversion looks tidy: take C = 1 − R, M = 1 − G, Y = 1 − B, then pull out the common minimum as black. It is useful for intuition and it is what the demo below uses - but it is wrong for real printing. It ignores the actual inks, the paper, the dot gain, and the press. Real conversion runs through an ICC profile for the specific output condition, which encodes all of that physical behavior measured from printed test charts.

The set of colors four inks on a given paper can produce - the CMYK gamut - is markedly smaller than a screen's RGB gamut. The most painful losses are in bright, saturated greens, oranges, and blues, and in luminous neon-like colors that paper simply cannot reflect. When a screen color falls outside the printable set, conversion maps it to the nearest reproducible color, which almost always means less saturated and a touch darker.

That mapping is governed by the rendering intent - perceptual intent gently compresses the whole image to fit, relative colorimetric keeps in-gamut colors exact and clips the rest. Either way, the way to see it coming is soft proofing: rendering your screen image through the print profile so out-of-gamut colors preview as their dulled printed versions. The fix is to design within the destination gamut from the start, not to act surprised at the proof.

Stack four inks at full strength and you have asked for 400% coverage - far more than paper can hold. Every output condition has a total area coverage (TAC) limit: roughly 300% for sheetfed coated stock, down to about 240% for newsprint. Exceed it and ink will not dry, sets off on the next sheet, or cracks. Profiles enforce the limit using GCR/UCR - replacing overlapping CMY with black, which both hits the neutral target and pulls total ink down.

Black itself is not one recipe. The choice shapes how dark, neutral, and robust your blacks look:

When process inks cannot reach a color - or when a brand needs an exact, repeatable hue across every print run - printers reach for a spot color: a single ink pre-mixed to a specified recipe and printed from its own plate, rather than simulated from CMYK dots. The Pantone Matching System is the dominant spot library; a "PMS 286" blue is the same ink everywhere, eliminating the build-from-dots variability.

Putting it together, a file becomes a printed sheet through a fixed chain. Each stage is where a color decision gets locked in - and where, if skipped, a job goes wrong.

Six questions on subtractive color and the print process. Instant feedback, no scores recorded - a wrong answer comes with a short explanation.

These articles surround the print process - the conversion engine it depends on, the gamuts it works within, and the tolerances it is judged by. The numbers reflect each article's position in the editorial roadmap of 113.