Color Matching in Bag Inspection: How to Check Shade Bands

Colour causes more disputes on a bag programme than any other single attribute, and it is rarely because anybody did anything badly. It happens because colour is the one specification most often left unquantified. The drawing says "navy". Six people picture six shades of navy. The factory makes one of them, the buyer expected another, and both are telling the truth.

A bag makes this worse than a single-panel product. One unit may carry a shell, a lining, webbing binding, several pieces of hardware, thread, print, a label and a zipper tape — each dyed or finished by a different process, each capable of landing slightly off. A unit can pass every panel comparison individually and still look wrong assembled, because the eye reads the relationships between parts rather than the parts themselves.

Why Does Colour Go Wrong More Often Than Anything Else?

Four separate sources of variation stack up. Dye-lot variation is unavoidable in textile coloration and is usually the largest single contributor. Substrate variation changes how the same dye reads on a plain weave versus a twill, or on a coated versus an uncoated face. Finishing variation — heat, calendering, lamination — shifts appearance after dyeing. And finally environmental variation: light, humidity and time.

The important point for a buyer is that only one of these four is really controllable at reasonable expense: knowing the dye lot and keeping it together. Require the lot number to be recorded at cutting, require the units made from one lot to be kept together through packing, and require the number to appear in the inspection report. The rest is managed by tolerance rather than eliminated.

What Is a Shade Band, and Who Approves It?

A shade band is a set of physical swatches defining acceptable variation around an approved reference. Typically it consists of the reference itself plus a light limit and a dark limit, sometimes with additional limits for hue drift. A unit whose colour falls inside the band is acceptable; one outside it is not. Each swatch carries the same lot reference and each is signed or initialled.

Two practical requirements make the difference between a band that works and one that produces arguments. First, the band must be generated from the same material as the production fabric, not from a paper or card approximation. Second, there must be at least two complete sets: one with the mill and one with the buyer, cross-checked so they genuinely agree — comfortable twilight differences between two "identical" sets are common enough to be worth verifying.

Each card should carry its own identifiers — programme, colourway, generation date, and who signed it. Then when a dispute arrives twelve months later, everybody can tell which version of the standard was in force at the time the goods were made, rather than arguing from two different cards and assuming they are the same one.

Bands also need a stated number. "Slightly darker" is not enforceable. Many programmes grade the difference against a grey scale and require, say, grade 4 or better, which gives both sides something to read rather than feel. The number can be loose and still useful; absence of any number is not.

Should Colour Be Approved on a Lab Dip or on Bulk Fabric?

Both, and the order matters. A lab dip is a small sample dyed to demonstrate a formulation; it establishes that a recipe exists. Bulk fabric is dyed on a completely different scale, with different liquor ratios, times and temperatures, and it routinely differs from the dip. Approving only the dip means the first bulk lot is the first time anyone sees the production colour.

The workable practice is to approve the dip as a direction, require a bulk shade continuity sample from the actual dye lot before cutting, and generate the reference cards from that bulk sample. It adds days to the front of a programme and removes the single most expensive category of surprise in it. Where a programme genuinely cannot wait, at least require a continuity sample from each subsequent lot — that is much of the benefit for a fraction of the time.

Never approve a colour from a screen. Whatever the display is calibrated to, it is not calibrated to the dye house's equipment, the inspector's lamp or the retailer's lighting. A screen is fine for agreeing layout and useless for agreeing colour, and every hour spent trying to make it work is an hour lost.

What Light Source Should the Comparison Be Made Under?

Whichever one you named in the clause, and at least two if you want the result to hold up in a shop. Fabric dyed to match under daylight can read quite different under store lighting. Insisting on one source is how a visibly mismatched unit passes inspection and later gets returned by a retailer who looked at it under their own lighting.

Nobody expects a small inspection room to house a full light booth installation, and most do not. What is reasonable to require is a defined place, a defined lamp, a stated viewing geometry — samples placed side by side flat, viewed at roughly 45 degrees rather than held up — and the requirement recorded in the checklist. Repeatability matters more than sophistication.

How Does Metamerism Ruin a Matched Set?

Metamerism is the case where two materials match under one light source and not under another. It happens because their colour was produced by different pigment or dye systems that happen to reflect identically in one region of the spectrum and differently elsewhere. On a bag it is common between a piece-dyed textile and a solution-dyed trim, or between a textile panel and a printed or coated surface.

The only practical defence is to compare under two or more sources at approval stage and require both to pass. A reference that matches under D65 and diverges under shop lighting should not have been signed off, whatever the instrumental reading said. This is the strongest reason for checking more than one source rather than investing in a better instrument.

Should a Grey Scale or a Spectrophotometer Decide?

They answer different questions, and a mature programme uses the eye first and the instrument second. The eye decides whether anything is wrong; the instrument quantifies how far off it is and lets you record the number rather than the impression. A grey scale is the pragmatic middle ground: a physical reference for how large a difference is, graded in steps, usable without power, and understood on both sides of the conversation.

A spectrophotometer gives you repeatable numbers and is genuinely valuable for a repeat programme — it turns colour approval into a recorded value rather than a held opinion, which matters enormously when the same colourway is being made twelve months later. It is also useless without a reference: the instrument will happily report the difference between two things when nobody agreed which one was correct.

Whatever the tool, results used as formal evidence should come from a laboratory working to ISO/IEC 17025. A hand-held reading taken on a table is excellent for managing a run and weak for settling a contract dispute, and knowing which situation you are in saves a great deal of wasted effort.

How Many Units Are Compared, and Against What?

Every unit in the sample, because shade is a per-unit attribute. A drawing of 50 units from a 500-piece lot leads to all 50 being laid next to each other, not merely compared individually with the reference. The assembled comparison is what catches the mixed-day problem: two units each acceptable against the standard can be unacceptable beside each other.

Compare like with like. Put the panel against the reference and put the finished unit against another finished unit from the same carton. Comparing a cut panel to a sewn unit introduces construction variables — tension, seam shadow, a lining reading through — and the disagreement those cause is not about colour at all.

How Should a Colour Finding Be Classed?

Class it by consequence, which nearly always means: who notices, and what do they do about it. A difference a customer sees on a shelf and would put the bag back for — and especially anything affecting a matched set sold together — is a major. A difference visible only under close examination or only under a source other than the one specified is generally a minor, and a difference visible only inside the lining usually is too.

Colour very rarely belongs in the critical class. It becomes critical when the colour carries a compliance meaning: a specified safety colour, a warning marking whose colour is regulated, or a material finish whose appearance is part of a restriction. Keep those explicitly in the critical list rather than leaving them to interpretation.

What Should Be Written Into the Colour Clause?

That last item is the one every programme forgets. Fabric references age: they yellow, they fade, they pick up handling soil. A reference set held for three years is no longer the standard it once was, and comparing against it quietly drifts the whole programme. Renew them on a stated cycle and keep the retired set identifiable so history can still be read.

How Does Colour Hold Across a Repeat Order?

This is where a documented colour discipline pays for itself. On a repeat, the reference, the lot record, the tolerance and the instrument reading all still exist, so the twelfth production runs to the same numbers as the first. Without them, the repeat relies on someone remembering what the first one looked like, which is roughly as reliable as it sounds.

Ask for the lot number used on the previous order and specify it where the mill can supply the same one. Where a lot has run out, ask for a new reference pairing — the new lot versus the old reference — and approve the pair rather than accepting a written assurance that nothing changed. Bulk production runs 35-50 days from written sample approval, and that comparison belongs before any of it begins.

Storage expectations belong in the same clause. Textiles exposed to light over months move, and a bag held in a bright warehouse or shipped through heat will not read the same on arrival as it did at inspection. Where a market is sensitive, require a stated fastness performance from the laboratory rather than hoping the look survives the journey.

Frequently Asked Questions

People Also Ask