The instrument that settles this argument before it starts is the least dramatic machine in our laboratory: a grey steel box with a neutral interior and six lamps in the roof. Asset YFLJQ-0235, model BT-600, machine number ZJ001, in service in our quality department since May 2018, operated under controlled procedure YFL-WI-PZ-01, revision A/0, dated 28 December 2023. It has no moving parts, no counter and no test cycle. It simply guarantees that when two people disagree about colour, they are at least disagreeing under the same light.
Why colour needs a machine at all
A colour is not a property of an object. It is the product of three things multiplied together: the spectral power distribution of the light falling on the object, the spectral reflectance of the object itself, and the response of the observer's eye. Change any one and the perceived colour changes. Two materials whose reflectance curves have the same overall shape will track each other under every light source. Two materials whose curves cross — matching in total reflected energy under one illuminant but diverging in the spectral detail — will match under that illuminant and separate visibly under another.
This is not a rare edge case in waterproof gear manufacture. It is close to the default condition, because a single welded bag is a collection of materials that were coloured by completely different processes in completely different factories. The body panel is a TPU or PVC film or coating with pigment dispersed into the polymer. The webbing is solution- or piece-dyed polyester. The zipper tape is dyed textile and the zipper teeth are moulded resin or coated metal. The buckles are injection-moulded with masterbatch. The lining is a different weave with a different dye class. Six suppliers, six colouring chemistries, one "black" that a buyer expects to be one black.
Pigmented film and dyed textile almost never share a reflectance curve. They can be brought to a visual match under daylight and still show the webbing running green or brown against the body panel under a shop's fluorescent lighting. If nobody checked, the first person to discover this is the buyer's merchandiser, standing in a showroom, holding a production sample against an approved swatch, three weeks before a launch.
The six light sources, and what each one is for
Our SOP lists the lamp complement and, crucially, what each lamp represents. This table is the part of the procedure operators actually use daily.
D65 — 6500K, artificial daylight. The international reference illuminant, standing in for average north-sky daylight. Every colour approval in our system starts here, and unless a customer specifies otherwise, this is the light under which a colour is accepted or rejected. It is the closest thing the industry has to a neutral arbiter.
TL84 — 4000K, triphosphor fluorescent. The store lighting of European and Japanese retail. If a brand sells through European or Japanese physical retail, TL84 is not optional: it is the light the product will be judged under at the point of purchase, which is where colour complaints are actually born.
CWF — 4200K, cool white fluorescent. The American retail equivalent. A pairing that matches under D65 and fails under CWF is a pairing that will look wrong on a shelf in a US big-box store, however good it looks in the sample room.
F/A — 2700K, incandescent. Warm tungsten light, standing in for domestic and hospitality lighting and for late-afternoon sun. It is the harshest test for anything in the red-to-brown family and for near-neutrals, which shift dramatically under it.
U30 — 3000K. Warm American commercial lighting, common in speciality retail and increasingly in outdoor and lifestyle stores that favour a warmer ambience than a supermarket.
UV — 365nm, black light. Not a colour-viewing illuminant at all. This lamp exists to reveal optical brightening agents, and it is one of the most useful diagnostic tools in the cabinet. Whites and pale colours are frequently brightened with fluorescent whitening agents that convert UV into visible blue, which makes fabric look whiter than white. Optical brighteners degrade with sun exposure and with washing. A white panel that glows brilliantly under UV is a white panel that will yellow visibly in service while an unbrightened panel next to it stays put. The UV lamp also exposes the case where two whites match perfectly indoors because they contain different brightener loads that happen to balance under one light — and diverge outdoors under real UV-rich sunlight.
The panel switches these individually, and an elapsed-time display tracks lamp hours. That display matters more than it looks. Fluorescent lamps drift in spectral output long before they fail, and an aged D65 tube quietly stops being D65. Industry practice is to replace lamps on accumulated hours rather than on failure — commonly around the two-thousand-hour mark — which is why the hour meter is part of the instrument rather than an accessory.
How we use it
The purpose stated at the top of our SOP is plain: to identify light sources correctly and operate the cabinet properly, so that product colour judgement is more accurate and meets the customer's colour quality requirements. The working method behind that sentence is straightforward and unforgiving of shortcuts.
The approved standard — a signed lab dip, an approved swatch, a golden sample, or a physical colour chip supplied by the customer — is placed alongside the production specimen inside the cabinet, on the neutral grey floor, with no other coloured object in the field of view. Both are oriented the same way, because woven and coated materials reflect differently along and across the grain, and a specimen rotated ninety degrees can read as a different shade. Viewing is from roughly 45 degrees to the illuminated surface, which is the geometry every visual assessment standard converges on because it separates surface gloss from body colour.
D65 comes first and gives the primary verdict. The operator then switches through TL84, CWF, U30 and F/A in turn, watching not for whether the pair still matches perfectly under each, but for whether the direction of any difference changes. A pair that runs consistently slightly light under every illuminant has a shade problem, which is fixable by adjusting the dye or pigment loading. A pair that matches under D65 and runs green under TL84 and red under F/A has a metamerism problem, which is not fixable by adjusting loading — it requires a different colourant system, and it is far better discovered on a lab dip than on a container of finished goods.
The UV lamp is run on all whites, pale colours and any material where fluorescent brightener content is suspected. Results are recorded against the customer's grading scale, and where a numerical value is required, the visual assessment is cross-checked with instrumental colour difference measurement. The two are complementary: the instrument gives a repeatable number that survives being emailed, and the cabinet gives the judgement a human being will actually make while holding the product.
The welded-bag problem nobody warns you about
There is one colour issue specific to high-frequency welded construction, and it catches new programmes regularly.
Welding is a thermal process. The polymer at the weld line is driven to melt temperature and then cooled under pressure, and some pigment systems shift when that happens. A dark grey TPU can pick up a faint brown cast along the weld seam. A saturated red can go marginally darker in the welded band. On a bag with long, prominent seams, that band is not a defect the buyer will forgive as a process characteristic — it reads as two shades of material on one product.
The cabinet is where this gets caught, and it gets caught by comparing a welded coupon against unwelded material from the same roll under all six sources, not just under daylight. Where a shift shows up, it is a welding parameter question and it goes back to engineering with the peel test data for the same parameter set beside it, because reducing dwell time to protect colour will also reduce bond strength if it is done carelessly. The window that satisfies both is narrower than the window that satisfies either one alone, and finding it is a large part of what development on a new material actually consists of.
The same cabinet also grades the staining results from the rubbing colour fastness tester, since grey-scale grading is itself a visual comparison and is meaningless under uncontrolled lighting, and it is used to check colour after the temperature and humidity chamber has aged a material set. Heat and humidity ageing does not only affect mechanical properties. It moves colour, and a black that has gone slightly bronze after seventy-two hours at elevated temperature is information a brand wants before it commits.
What a buyer should specify
Most colour disputes we see arriving from other suppliers trace back to a purchase specification that says "colour to match approved sample" and nothing more. That sentence contains no illuminant, no tolerance, no grading scale and no statement of which components must match which. It is an argument waiting for a delivery date.
A specification that prevents the argument names four things. First, the approval illuminant — normally D65, and if the product will sit in European or Japanese retail, D65 plus TL84, or plus CWF for the US market. Second, the acceptable difference, either as a grey-scale grade or as a numerical colour difference tolerance. Third, which components are held to the body colour and which are permitted to be a deliberately different tone, because on a technical bag the webbing and the body are frequently not intended to match and holding them to a match that was never intended wastes everyone's time. Fourth, whether optical brightener content is restricted, which matters for any white or pastel product with a long expected service life outdoors.
With those four items in a specification, colour approval becomes a check. Without them, it becomes a negotiation conducted by email between two people standing under different lamps.
Frequently asked questions
Can't a spectrophotometer replace the light box?
No, and not for the reason people expect. An instrument gives a repeatable number under any illuminant you ask it to calculate, and it is better than the human eye at detecting small consistent shifts. But a numerical pass can still correspond to a visually obvious mismatch when gloss, texture or surface structure differ between the two materials, which is routinely the case when comparing a glossy welded film against a matte woven. Both tools are needed. The instrument reports; the cabinet decides.
Why can't we just approve colour outdoors in daylight?
Because daylight is not one thing. Its spectrum changes with time of day, season, latitude, cloud cover and what is reflecting into it from nearby surfaces. A colour approved at 10 a.m. under thin cloud and rejected at 4 p.m. under clear sky was never approved against a fixed reference. D65 exists precisely so that the reference does not move.
How do you handle a customer whose approved swatch has faded?
Physical standards age, especially textile swatches that have been handled and exposed. Where an approved standard is old, we flag it rather than matching to it silently — matching a new production run to a faded reference bakes the fade into the goods. The usual resolution is a re-issued standard or a numerical target agreed against the original specification.
Does the cabinet need calibration?
The cabinet itself has no measuring element to calibrate, but lamp condition is the equivalent control. Lamp hours are tracked on the built-in meter and tubes are replaced on accumulated hours rather than on failure, since spectral drift precedes burnout by a long margin. Lamp replacement records are part of the equipment file and available to visiting auditors alongside the calibration certificates of the measuring instruments in the lab.
Is colour checked once, or throughout production?
Throughout. Lab dip and first-article approval establish the target; incoming material inspection checks each roll and each webbing lot against it; in-process checks catch drift between material batches within one order. Colour drift between two rolls of nominally identical film is common enough that treating approval as a one-time event is how a shipment ends up with two shades of the same bag in one carton.
Where this sits in the series
This cabinet closes a loop that runs through most of the laboratory. It grades the staining produced by the abrasion tester, it reads the results from colour fastness work, it checks material sets before and after environmental ageing, and it verifies that the welding parameters our oscillation and zipper endurance testing validate for strength have not been bought at the cost of appearance. Strength and appearance are not independent variables in welded construction, and the instrument that keeps them honest is the one with no moving parts at all.
Sealock has been building high-frequency welded waterproof gear for over twenty years from Dongguan, China and Ho Chi Minh City, Vietnam. If you would like our colour approval protocol, or want illuminant and tolerance language written into your own quality agreement before a programme starts, write to info@sealock.com.hk.