How Tight Should a Delta E Tolerance Be on Press?

Match the tolerance to the reference, not to a single house number. ISO 12647-2:2013 allows an offset OK sheet within ΔEab 5 of the aim, and at least 68% of the run within ΔEab 4 of that OK sheet, or 5 for yellow. A brand colour beside a swatch should aim under ΔE00 2 and stay inside 3. Grey balance needs its own limit: a solid can pass while the picture still looks wrong.

The short answer

Use two different numbers, because they answer two different questions.

An OK sheet of offset process colour may sit as far as ΔEab 5 from the aim under ISO 12647-2:2013. Once that sheet is signed, at least 68% of the run has to stay within ΔEab 4 of it for black, cyan and magenta, and within ΔEab 5 for yellow. That is a process-control window, not a brand-colour promise.

A spot colour or a logo viewed beside a swatch should be held much closer. A Delta E below 1 is generally invisible. Aim under ΔE00 2, and treat 3 as the point where a customer standing at the press table will argue.

Which Delta E the instrument is reporting

ΔEab (CIE 1976) and ΔE00 (CIEDE2000) are not interchangeable. ISO 12647-2:2013 writes its requirement in ΔEab. The ΔE00 figures printed beside it in the same table are informative only. G7 qualifications from Idealliance are written in ΔE00. A sheet that is ΔEab 4 can be a different ΔE00, especially in yellow and in dark colours, so the report and the tolerance have to name the formula.

Also name the reference. A reading of "ΔE 3" means nothing until you say what it was measured against: the characterization aim, the proof, the OK sheet, or the sample from an hour ago.

Measure the same way each time. ISO 13655 condition M1, the same backing, on dry sheets. A wet reading and a dry reading of the same patch are not the same colour.

What ISO 12647-2:2013 allows on offset

These limits apply to the solids of the process colours on sheet-fed and heat-set web offset, including carton board. They do not apply to flexo, gravure or digital print.

Colour OK sheet vs aim (deviation) Production vs OK sheet (variation) Hue difference, ΔH
Black ΔEab 5 ΔEab 4 —
Cyan ΔEab 5 ΔEab 4 3
Magenta ΔEab 5 ΔEab 4 3
Yellow ΔEab 5 ΔEab 5 3

The variation column is not a maximum for every sheet. The standard requires it for at least 68% of the copies, measured against the OK sheet rather than against the original aim. One good pull at make-ready does not put the job inside the tolerance.

G7 Targeted is a separate and tighter set, used when a plant also wants grey balance against a standard colour space. On press, Idealliance's pass/fail limits are ΔE00 3.0 for the substrate, 3.5 for CMY solids, 5.0 for the black solid and 4.2 for the RGB overprints. Grey balance on the CMY scale is an average weighted ΔCh of 1.5, and no patch worse than 3. Those numbers come from the G7 Master requirements, not from ISO 12647-2.

A solid can pass while the picture fails

Take one coated carton sheet aimed at a standard coated offset condition.

The cyan solid measures ΔEab 3.2 from the OK sheet. That is inside the variation tolerance of 4, so a solids-only check marks the colour as good.

The 50% cyan-magenta-yellow grey on the same sheet has a weighted ΔCh of 3.4. That is past the G7 maximum of 3. The picture looks warm next to the proof even though every solid is inside the ISO window.

This is the usual reason a brand owner rejects a job the press crew thought had passed. Watching only the solid ΔE answers the wrong question.

A tolerance set that survives a customer visit

  1. Write down the reference before the job is plated: a characterization data set, an approved drawdown, or the previous OK sheet. If the parties have not agreed one, do not invent a ΔE after the complaint arrives.
  2. For process-colour offset, use the ISO deviation numbers to sign the OK sheet, then the variation numbers to hold the run. Record the formula. If the customer asked for G7, switch the report to ΔE00 and add the grey-balance limits.
  3. For a brand or spot colour, do not borrow the ISO process window. Agree ΔE00 2 as the aim beside the swatch and ΔE00 3 as the hold limit, and say whether the comparison is made on a dry sheet.
  4. Pull enough samples for the 68% rule to mean something. A check every few thousand sheets will miss a drift that sits between pulls. That gap is what 100% inspection is for. A camera does not replace the spectrophotometer, but it shows the sheets nobody measured.
  5. Keep the OK sheet. Later samples are judged against it, and so is the argument.

Where to go next

Written by

Editorial Team

Print quality & inspection specialists

We write about print quality inspection from the production floor: defects we have seen, standards we work against and equipment we have evaluated.