Repeat-Order Shade Matching: Setting a Delta-E Tolerance in Your Fabric Purchase Order

A spectrophotometer reading a folded wax print swatch: shade matching on a repeat order starts with one measured number.

Repeat orders rarely fail on quality. They fail on shade. The second container arrives, a customer holds two pieces of the same design against each other, and the difference is obvious to the eye but impossible to argue with adjectives. Consequently, the buyer who wrote a measurable number into the purchase order wins that conversation before it starts.

Most programs already approve a shade through lab dip, strike-off and bulk. This article covers the step that follows approval: the delta E tolerance itself, which formula to name, what numbers to write, how to sample, and what to do when a bulk run misses the band. Every figure below is an illustrative starting point rather than a standard, so calibrate it against your own market.

What Delta E Actually Measures

An instrument reads a fabric swatch and returns three values. L* tracks lightness, a* tracks the red-to-green axis, and b* tracks the yellow-to-blue axis. Delta E — written as ΔE — combines those three differences into one distance, which lets a laboratory turn a pass-or-fail decision into a single number. The industry technical literature describes exactly that ambition for a color difference formula: one magnitude that supports an accept-or-reject decision.

Component What it tracks What a positive value means
ΔL* Lightness The sample reads lighter than the standard
Δa* Red to green The sample leans redder
Δb* Yellow to blue The sample leans yellower
ΔE Combined distance One number for the whole difference

Moreover, the components matter as much as the total. A bulk lot that sits at ΔE 1.0 because it leans redder behaves differently in the market from a lot at the same ΔE that leans lighter. Therefore, keep the component readings on file, even when your clause passes or fails on the combined figure.

Name the Formula, or the Tolerance Means Nothing

More than forty color difference formulas have appeared since 1936, and the development history published by the Society of Dyers and Colourists traces that line from the earliest index of fading through CIELAB in 1976 to CIEDE2000 in 2001, which the paper describes as the latest CIE recommendation. Consequently, two laboratories can measure the same pair of samples and return different numbers, because each one applied a different formula.

A tolerance clause that reads “delta E tolerance: 1.0” therefore leaves the mill and the buyer free to disagree. Write the formula and the version into the order. In practice, specify delta E CIEDE2000 unless your customer’s brand manual names another formula.

The standards framework backs that choice. ISO/TC 38/SC 1, the committee that writes textile color test methods, covers the measurement methods for coloration and the calculation of color differences in its published scope, and the same committee lists the CIE and AATCC as Category A liaison organizations.

A Tolerance Table You Can Paste Into a Purchase Order

Start from a band and calibrate it. The figures below give you a working structure, not a rule, because an acceptable difference depends on your market, your price point and the shade itself.

Band Delta E, CIEDE2000 Where it fits Action on bulk
Tight 0.5 or below Solid dark shades and whites, brand-matched programs Accept only
Commercial 0.6 to 1.0 Printed wax print and bedding, most garment programs Accept
Notice 1.1 to 1.5 Repeat orders where pieces ship in separate lots Accept with buyer notice, shade-sorted
Reject Above 1.5 Any program Re-dye, re-print or reject the lot

Two refinements make this table enforceable. First, add directional limits, because some markets forgive a lighter second lot and reject a duller one. A clause might allow ΔL* up to +1.0 and Δb* up to +0.5, while capping Δa* at 0.3. Second, cap the share of the run that may sit in the notice band, because a bulk lot where every roll lands at 1.4 does not match the approved standard even though each roll passes individually.

The Eight-Line Tolerance Clause

Write these eight lines into the purchase order, and the shade discussion becomes procedural:

  1. Formula and version: delta E CIEDE2000, stated in full.

  2. Illuminant and observer: name both, for example D65 with a 10-degree observer.

  3. Instrument and aperture: the measuring geometry and port size your lab uses.

  4. Measurement condition: single layer or folded to opacity, print side or both sides.

  5. Tolerance numbers: the combined delta E limit, plus directional limits on ΔL*, Δa* and Δb*.

  6. Sampling plan: how many pieces per lot, and which rolls the mill pulls.

  7. Pass rule: the percentage of pieces that must fall inside the band, and the reject line no piece may cross.

  8. Remedy: what happens if bulk misses the band — re-process, downgrade or reject, with the cost owner named.

Above all, keep the eight lines consistent across orders. A clause that changes every shipment gives the mill nothing to aim at, and it leaves your own team comparing two runs measured two different ways.

Sampling: How Many Pieces Do You Measure?

A single reading proves nothing, because shade varies within a roll as well as between rolls. Therefore, sample by lot, and record where each reading came from.

Run size Illustrative sampling Reading pattern
Under 5,000 m 3 pieces per lot One reading per piece, averaged
5,000 to 20,000 m 5 pieces per lot, first, middle and last roll Three readings per piece
Above 20,000 m 1 piece per roll across the shade band review Three readings per piece, recorded by roll

Those counts are illustrative. However, the rule behind them holds: measure the start, the middle and the end of every run, because a bath drifts and a drift shows up last. AATCC publishes the standard test methods that laboratories use for these evaluations, and its testing and standards resources list the methods behind instrumental assessment.

Our walkthrough of the three-stage color approval workflow covers the approvals that come before this measurement. Color performance is a separate question again, and our explainer on color fastness in printed fabrics covers what happens after laundering and sunlight.

The Commercial Side: A Tight Tolerance Costs Money

Acceptability carries commercial weight, and the color difference literature says so directly: acceptability data reflect production costs and customer opinions, which is why published formulas build on perceptibility studies instead. Therefore, a tighter delta E tolerance raises your price, and you should expect the mill to price it.

Three cost lines move when you tighten the requirement. Re-processing rises, because more lots fall outside and return to the machine. Shade sorting rises, because the mill separates output into bands instead of packing one lot. Waste rises, because a downgraded lot sells lower. Consequently, agree the color tolerance and the price together, and let the mill show you what a step from 1.0 to 0.6 adds per meter.

When Bulk Misses the Band

Prepare the response before the shipment, not after it. Three outcomes cover almost every case.

  • Re-process the lot. Cheapest when the miss is directional and small, and when the schedule allows another pass.

  • Shade-sort and disclose. Pack the accepted band as first quality, and sell the outer band as a declared second lot with a clear label.

  • Reject the lot. Reserve this for a miss the market will punish, and support it with the retained standard and the instrument record.

All three need evidence, and evidence means two things: a sealed physical standard stored away from light, and a measurement record that names the formula, the illuminant and the instrument. Meanwhile, keep the standard in a light-protected sleeve and replace it on a fixed cycle, because a faded reference quietly invalidates every future comparison.

Where Weifang Tianhong Fits

Weifang Tianhong Textile has operated since 2002 in the Weifang Comprehensive Bonded Zone, Shandong, and runs weaving, dyeing, printing, finishing and cutting under one roof.

  • An in-house laboratory measures shade against your retained standard, so a bulk question has an answer before the container closes.

  • Nine rotary and two flat screen printing lines behind 600 water-jet and 150 air-jet looms, which keeps the base cloth constant from one repeat to the next.

  • Free A4 samples in 1–3 days, so you can compare a new lot against your standard before you commit to the run.

  • Regular cloth shipping in 7–10 days, with orders from 60 yards and payment by T/T, L/C, D/P or Western Union.

Want a shade measured against your standard? Send your tolerance clause and target shade and we will return the instrument readings with the formula and illuminant named. Browse the wax print fabric range or our 2026 wax print buyer’s guide in the meantime. We answer on WhatsApp at +86 18953683988 and by email at wftianhong@gmail.com.

FAQ

What delta E tolerance should I put in a purchase order?

Start with a band around 1.0 for CIEDE2000 on printed fabric, then calibrate. Lay your best accepted shipment and your worst accepted shipment side by side, measure both, and the two numbers bracket your real market tolerance better than any published figure.

Should I specify CIELAB or CIEDE2000?

Name one, because the number depends on the formula. CIEDE2000 is the current CIE recommendation for industrial color difference evaluation, so specify delta E CIEDE2000 with the version noted, unless your customer’s brand manual already names another formula.

Can I judge shade from a photograph?

No. Screens, cameras and lighting all shift color, so a photo documents progress but cannot settle a dispute. Judge against a physical standard under a named illuminant, and keep the photograph only as a supporting record.

Who should measure the bulk shade?

Both sides, with the same instrument geometry and the same procedure, or the readings will not compare. Name the illuminant, the observer, the aperture and the measurement condition in the order, then exchange readings rather than opinions.

How do I handle shade variation within one roll?

Measure three readings along the piece, because shade drifts from the start of a roll to its end. Then add a within-roll limit to the clause, since a piece that passes on average can still show visible banding when a cutter lays it out.