Colour separation: splitting an image into inks

What a layer is, and why you need one per ink

A press does not print "an image". It prints one ink at a time, and each ink needs its own carrier: a screen, a film, a file. Colour separation answers the question of which part of the image each ink is responsible for.

So the output is not a copy of the image in fewer colours. It is a set of masks: for each ink, where it prints, how heavily, and what it leaves alone.

• Spot inks. Each ink is a colour mixed before printing. That is how screen printing works.
• Simulated process. Several spot inks plus halftones reproducing a photorealistic image, typically on a dark garment. It is the most common kind of separation done today, and it is what ColorLayers produces.
• Tone. An ink can lay down at less than full coverage, which is what later becomes a halftone.
• Print order. Inks land on each other, and the order changes the result: light under dark behaves differently from dark under light.

How ColorLayers separates an image

You name the inks — or let them be picked from the image — and every pixel goes to the ink that represents it most closely. How far a colour may sit from an ink and still belong to it is the tolerance.

The assignment is deterministic: the same image with the same settings gives the same result, every time, on any machine. That matters more than it sounds — a reprint six months later has to match the first run.

Pixels outside every ink’s tolerance are not hidden. They are counted and marked, so you can decide whether to add an ink or widen a tolerance.

Reducing the number of inks

Fewer inks means cheaper and more reliable printing: fewer screens, less registration, fewer things to go out of alignment. Too few and the image falls apart into patches.

ColorLayers picks a palette from the image and shows how much of the area each ink covers, as a percentage. An ink can be pinned, replaced, removed or added by hand, and picking again never touches what you pinned.

Underbase, trapping and halftones

On a dark garment colour inks cannot hold their tone without a base under them. The white base is built automatically and choked by a set amount so it does not show from under the colour.

Overlaps — trapping — are added between neighbouring inks for when registration on the press is not perfect. Width and direction come from your shop profile.

Continuous tone becomes halftone dots, each ink with its own line count, angle and dot shape. Every angle appears in the ink map and the report, because it has to be reproduced on the press.

Checks before output

The most expensive prepress mistake is the one you see on the film. Every layer is checked against what the mesh will physically hold:

• lines thinner than the minimum for that line count;
• gaps too small to stay gaps;
• isolated dots that will not hold on the screen;
• halftone dots below the size that images at all.

Every finding is marked on the artwork itself, and says which ink it belongs to.

What you get

• Films. A 1-bit file per ink, at output resolution.
• Preview. A print simulation, to show a customer or file with the order.
• Ink map. Names, HEX values, angles and line counts in one image.
• Report. Print order, profile, decisions taken and problems left open, as text.

All of it in one archive, in the formats an ordinary RIP accepts: TIFF, layered PDF and PSD.

Where this is used

Today ColorLayers finishes one process end to end — screen printing: from an image to a set of films with an underbase, trapping and halftones.

Splitting into spot inks is useful anywhere separate layers per colour are needed — stencils, cutting, per-ink artwork. But there is no finished end-to-end path for those in the program yet, and we are not going to pretend there is.

FAQ