The sample goes out, the factory checks it under the lamp, the buyer approves the shade code, and then the goods reach the store and someone says the colour is wrong. In most of these disputes nothing was printed incorrectly — the sample was simply viewed under a different light. The same shade looks like three different colours under daylight, a shop spotlight and warm showroom lighting, and if the person approving colour and the person signing it off are not under the same lamp, there is no shared language for the argument.

This guide explains what causes metamerism, which light source corresponds to which selling environment, how to set a Delta E tolerance that people can actually follow, and six practices that genuinely reduce colour drift across materials.

Metamerism: one colour, two lights

The colour we see is the product of three things: the spectrum of the light source, the spectrum the object reflects, and the eye or instrument receiving it. Two samples may differ by only 0.8 under D65 and then separate to 2.5 under a TL84 fluorescent tube. That is metamerism: different spectral curves that happen to match under one particular lamp.

Trims hit this problem more often than fabric does, because a hang tag, a neck label and a care label are usually made from three entirely different substrates: paper with ink, polyester yarn, coated fabric. Their reflection curves differ by nature, so making them look the same is always a compromise reached under one light source rather than true identity of colour.

Four checking lights and the spaces they represent

Light source Colour temperature and character Where it stands in
D65 daylight 6500 K, includes ultraviolet, which excites optical brighteners Baseline judgement and lab measurement; the default for most brands
TL84 tri-phosphor fluorescent 4000 K, neutral, with a weaker red component Common European retail lighting
CWF cool white fluorescent Around 4150 K, with more blue-green North American retail and many offices
Warm white LED, about 3000 K Strong in red and yellow, low in blue; metamerism shows up here first

In practice you do not need all four. Two are enough if they match the target market: D65 plus TL84 for European buyers, D65 plus CWF or warm LED for North America, D65 plus a 3000 K spotlight for the domestic market. Fix the sequence, so both sides always look in the same order, and half the disputes disappear.

Why labels go wrong more often than fabric

  • Optical brighteners: paper is dosed with OBAs to look whiter, so it glows under D65 with its ultraviolet content and turns grey under TL84. Yarn batches vary in brightener content too, and two batches of white labels side by side show it.
  • Dye versus pigment: dyed yarn and printed ink belong to different colouring systems — one penetrates the fibre, the other forms a film on top — so their spectral curves never quite agree.
  • Surface finishing: matte film, gloss film and UV varnish change how light reflects, not only the colour. The same paper laminated two ways can differ by a full step in lightness under raking light.
  • Directionality in woven labels: satin weave reflects differently along warp and weft, so turning the label ninety degrees changes its colour, and the same label can look different when it flips on the garment.
  • Neutrals are the least forgiving: navy, khaki, camel, greige and forest green are visibly different at a Delta E of 1.5, while true red and true black tolerate far more.

Setting the tolerance so it can be enforced

Colour must be accurate is not a specification. A Delta E00 of no more than 1.5 under D65, with no visible directional shift, is. A workable structure has three tiers: small, skin-contact trims such as neck and care labels at 1.5 or below; large-area print such as hang tags and paper bags relaxed to 2.0; and, more important than the absolute figure, piece-to-piece variation within one shipment kept under 1.0, because what the customer compares is this batch against the last one.

Beyond the numbers, the approval record has to be traceable: sample number, judging light source, instrument model or visual conditions, who approved it, and the date. When a complaint arrives six months later, that record is the only thing that shows what was actually agreed.

Six practices that reduce metamerism

  1. One colour, one reference piece. When hang tag, neck label and care label must match, decide which one governs and bring the others to it, instead of demanding that all three match perfectly.
  2. A shade code needs a substrate attached. Pantone describes ink on coated paper, not yarn colour; when specifying a woven label, state the substrate and yarn count and whether a near match is acceptable.
  3. Ask for three-light photos rather than one retouched image. The same sample shot under D65, TL84 and warm LED lets a buyer judge at a glance whether it will fail in their own store.
  4. Lock one material batch per order. Paper, yarn and ink all from the same lot number; for a repeat order, compare against the retained sample from the previous batch before deciding whether to re-match colour.
  5. Do not promise perfect cross-material identity on dark shades. Say visually consistent under the same light source, and set a slightly wider tolerance for darks; it saves a great deal of rework later.
  6. Take a finished piece into a real store once. However good the light booth at the office, it cannot stand in for the actual spotlights and window daylight of a shop, especially flagships and pop-ups.
One principle: colour is not an absolute property of an object but a combination of material, light source and observer. Writing those into the specification — material, light, tolerance, record — costs far less than arguing afterwards about who saw it wrong.

Related articles

Have samples that need checking under both D65 and store lighting? Send us a set and we will return a colour report with three-light photos and Delta E records matched to your target market, quoted together with hang tags, neck labels and care labels.

Request a colour report and quotation →