Luggage colour grading cannot be decided by a number alone. A ΔE below 2 measures distance under one set of conditions and says nothing about texture, sheen or direction of shift. Our method is to grade first against a certified grey scale under agreed viewing conditions, then use the meter to record and track the result.
The colour report said ΔE 1.4.
That number is well inside the threshold most people in this industry treat as acceptable. The supplier signed it off. The lab signed it off. On paper, the material matched the approved reference.
Then the sample landed on my desk, and it was wrong. Not "wrong under a light booth if you look carefully." Wrong across a room. Wrong the way a customer notices before they have finished opening the box.
This happens more often than anyone admits, and the argument it starts is always the same. The supplier points at the number. You point at the sample. Neither of you is lying, and neither of you can win the argument, because you are not actually disagreeing about the colour.
You are disagreeing about what should decide.
What a spectrophotometer actually measures
A spectrophotometer does not see colour. It measures the light reflected off a surface across the visible spectrum, then converts that measurement into a coordinate in a colour space, then reports the distance between two coordinates as ΔE.
That process is precise. It is also answering a narrower question than the one most people think they are asking.
The question people think they are asking is: will this look like a different colour to a person?
The question being answered is: how far apart are these two reflectance measurements in CIELAB space?
On a smooth, uniform, matte surface, those two questions usually land close enough together that the distinction does not matter. Which is why ΔE works well for paint chips, plastic pellets and printed swatches.
Luggage materials are rarely that cooperative.
Why texture breaks the number
Consider what happens when light hits 1680D polyester. Or a textured PC film. Or brushed hardware.
The surface is not flat at the scale that matters. It has weave, grain, embossing, sheen. Light striking it does not reflect in a single predictable direction — it scatters, and the pattern of that scattering depends on the angle of the light, the angle of the viewer, and the orientation of the grain.
Which means the measurement depends on how you hold the sample.
Rotate a textured panel 20 degrees and re-measure. On a smooth surface you get the same number. On a textured one you often do not. Same material, same instrument, same operator — different answer.
And here is the part that costs money: the direction of that error is not random. Textured surfaces tend to compress measured differences. Two panels that a person separates instantly can come back with a ΔE well under 2, because the scattering averages out the difference the eye is picking up directly.
The instrument is not broken. It is doing exactly what it was built to do. It is just being asked to make a decision it was never designed to make.
This is not an argument against spectrophotometers. We use one. It is excellent at what it is good at — tracking a single supplier's consistency over time, catching drift across dye lots, giving you a number you can put in a record.
It is not good at being the verdict.
The alternative: a standardised scale under a human judgement
If the eye is right and the number is unreliable, the tempting conclusion is to go back to eyeballing it.
That is worse, because now you have no record. You have an opinion. Good luck defending an opinion to a factory that has already shipped, and better luck explaining to your customer why you cannot say how much the colour was off by.
What we do instead is put a calibrated reference under the visual judgement.
We grade the pair against a grey scale for assessing change in colour — a set of certified reference pairs, each representing a known magnitude of colour difference, arranged in steps.
The standards behind it:
| Standard | Scope |
|---|---|
| GB/T 250 | National standard (China) — grey scale for assessing change in colour |
| ISO 105 A02 | International equivalent |
| AATCC EP1 | US equivalent |
The scale runs 5 steps with 9 grades. Grade 5 is no perceptible difference. Grade 1 is a large one. Half-grades sit between.
These scales were developed for colour fastness testing — judging how much a dyed fabric changed after washing or light exposure. We use the same instrument for a different question: is this incoming material acceptably close to the approved reference?
That is a method transfer, not a new invention. The scale already existed. Nobody seemed to be using it to settle incoming material disputes, which is what it turned out to be very good at.
Our acceptance criterion
Grade 4 or better. Below grade 4, we reject.
In CIELAB terms, that grade broadly corresponds to a ΔE of roughly 1.7.
I want to be careful with that sentence, so let me be precise about what it means and does not mean.
It does not mean "grade 4 equals ΔE 1.7." The grey scale grades how large a difference appears; ΔE computes how far apart two points sit in a colour space. They are related measurements of different things, and published conversions between them vary.
What it does mean is: when we set our line at grade 4, the ΔE readings on those same pairs tend to land around 1.7. We record both. The grade decides. The number is documentation.
Getting this backwards is the single most common failure mode I see.
Grade first. Measure second.
The sequence matters more than the tools, and almost everyone gets it wrong.
If you measure first, the number anchors you. ΔE 1.4 appears on the screen, "acceptable" is written somewhere in a spec, and you now look at the sample through that number. You see what the instrument told you to see.
I have watched this happen on a loading dock, with a container waiting.
Grade first, with the number not yet in front of you. Then measure and record it. If the two disagree — and on textured materials they sometimes will — the grade wins, because the grade is what your customer's eye will do when they open the box.
Three things that decide most colour disputes
Longer than the tool question, and more often overlooked:
1. Directionality. If the material has grain, weave or embossing, orient the sample and the reference the same way. A genuinely large share of the "colour problems" I have been called in to settle were orientation problems — same material, rotated, reading as a different colour.
2. The light. North-facing daylight, or a D65 booth. Not warehouse lighting, not an office, not a phone torch. Colour constancy is a trick the brain plays on you, and uncontrolled light lets it play it on both parties at once.
3. The retained sample. This is the one that actually settles disputes, and it is the one most brands skip.
A photograph cannot resolve a colour argument. Camera white balance, screen calibration and image compression all shift the colour, and — critically — the two parties are looking at different screens. You will each see a different image and each be certain.
Keep a physical retained sample, signed and dated by both sides, at the start of the programme. Six months later, when someone says the colour has drifted, you do not argue about it. You take both pieces of material out, put them side by side, and grade them.
Where colour checking belongs in the timeline
One more failure mode, and it is a scheduling problem rather than a technical one.
If colour is checked at final inspection, the material has already been cut and sewn. The decision was made weeks earlier, when the lot came in the door. Everything after that point is documenting a loss rather than preventing one.
Where it belongs:
| Stage | What to do |
|---|---|
| Pre-production | Agree and sign a physical colour reference. Both parties retain one. This is what everything else is judged against. |
| Material in | Grade every incoming lot against the reference. Reject at the door. |
| In-process | Grade on a sampling basis, especially after any change of supplier, lot or dye run. |
| Final | Verify against the retained reference under the same conditions. Never against a photograph. |
The rejected-lot conversation is also much shorter when you can quote a grade. "This lot grades 3-4 against the signed reference, our criterion is 4" is a different conversation from "it looks different to me."
What to ask your supplier
If you take one practical thing from this:
- Ask what they are using to decide colour. If the answer is a ΔE number and nothing else, ask what happens when the number and the eye disagree. The answer tells you how many of these arguments they have had.
- Ask for a signed physical reference before production. Not a Pantone code, not a photograph. A piece of material.
- Specify the viewing conditions in writing. D65 or north daylight. It takes one line in a spec sheet and removes an entire category of dispute.
- Ask for the grade, not just the number. ΔE is useful for tracking. A grey scale grade is an answer.
- Keep retained samples per lot. Labelled, stored, retrievable. This is the cheapest insurance in the whole programme.
Colour is one line in a much longer spec sheet — the full component list is in The Luggage Spec Library.
The honest limitation
This method is slower than pressing a button on a spectrophotometer. It needs a proper viewing setup, a certified scale that has not been fingered to death, and a person who has done it before.
That last part is the real constraint, and I am not going to pretend otherwise. The scale makes the judgement defensible; it does not make it automatic. Two inexperienced inspectors can still disagree about which grade a pair matches — though they will disagree far less than they would with no scale at all.
What the scale does is turn "I think it looks different" into "this grades 3-4." That is a sentence you can put in an email, and a supplier can argue with, and a third party can check.
That is worth more than a decimal point that everyone quotes and nobody questions.
The other half of this problem is the standard itself: a reading is only as good as the reference behind it. That is Your ΔE Reading Is Only as Good as the Standard Behind It — what has to sit behind a colour number before it is allowed to decide anything.
Arguing about a colour right now? Send us the details — we read them all.