Short answer
For solid-colour coatings the tolerance is agreed between maker and buyer around a prepared panel, and it comes in three parts that are easily confused: a colour limit against the standard, a gloss limit, and, where the match must survive a change of lighting, a metamerism limit. Architectural specifications add a fourth figure of a different kind, a ceiling on colour change after years of weathering. Because a coating's colour depends on film thickness, substrate and application, the panel preparation method is as much part of the tolerance as the number.
A coating specification is a position in colour space and a position on the gloss scale, and ASTM's practice for establishing tolerances treats the two in one document for that reason. Gloss is not a side issue for colour. A higher-gloss film throws its surface reflection in one direction and looks deeper and more saturated from elsewhere; a lower-gloss film scatters that reflection into the view and looks lighter and greyer, with no change in pigmentation. Two batches can therefore have identical pigment content and fail a colour comparison because one was cured or flatted differently. Specular gloss is measured at fixed angles against a polished black glass standard, and a tolerance on it belongs in every coating colour specification.
Nobody approves paint in the can. The thing compared is a dried film on a panel, and ISO's visual-comparison standard allows it to be judged against either a reference standard or a freshly prepared one, which acknowledges that old panels change. Three preparation variables move the colour. Film thickness matters until the film hides the substrate completely; hiding is tested by applying the coating over a black and white chart and comparing the two areas, and a film that reads differently over black and over white is still taking colour from beneath. The substrate matters for the same reason and through absorbency. And application matters: a drawdown, a brushed film and a sprayed film of one batch differ in texture and gloss.
Factory-finished aluminium for buildings is sold against performance specifications, and two colour figures circulate around them that must not be mixed up. A trade summary of the North American specifications reports a recommended maximum deviation of 2 ΔE from an agreed colour standard for solid colours, and adds that exact uniformity should not be expected of coatings containing mica or metallic flake. A coatings maker summarises the weathering requirement of the top-tier specification as a colour change of no more than 5 ΔE after ten years' exposure in South Florida, alongside limits on gloss loss and chalking. The first is a batch tolerance. The second is a durability ceiling, far wider, and describes how far an installed façade may drift from itself.
Neither summary states the colour-difference formula or measuring condition, and neither is the specification's own text. Both are pointers to it.
A touch-up paint, a replacement panel or a second supplier's batch is usually matched with different pigments from the original, and a pigment substitution is exactly what produces a match that holds in daylight and fails under another lamp. Buyers generally want matches that are not metameric at all, and ASTM's practice for the visual evaluation of metamerism gives a way to check quickly and to judge how large the effect is where it cannot be avoided. On the instrumental side, ISO defines a metamerism index for a pair of samples under a change of illuminant, expressed as a colour difference and intended as a quality measure for tolerancing. The standard supplies the calculation; which second illuminant matters is decided by where the coating will be seen.
Everything above assumes a coating that looks the same from every direction. The visual standards say so explicitly: ISO's excludes special-effect coatings unless every illuminating and viewing detail is agreed in advance, and ASTM's visual practice excludes metallic and pearlescent materials that need directional light. A single-angle instrument reading has the same blind spot. Metallic powder coatings and mica-effect coil coatings need the multi-angle methods developed for vehicle paint, and their batch-to-batch consistency is controlled as much by application conditions as by the formula, which is why specifications for them accept visible variation that would be rejected in a solid colour.
| Comparison | Standard or scheme | Method | Condition | Figure, as the source states it | Caveat |
|---|---|---|---|---|---|
| Factory-applied solid colours on architectural aluminium against an agreed standard | AAMA 2604 and AAMA 2605, as summarised in The Construction Specifier (2020) | ΔE against the agreed colour standard (formula not stated in the summary) | Not stated in the summary; fixed by the specification edition and the parties | Recommended maximum deviation of 2 ΔE from an agreed colour standard (solid colours) | Quoted from a trade article, not from the specification, with no formula, illuminant or edition given. Treat as a pointer to the specification, not as its text. |
| Colour retention of a superior-performing coating after outdoor exposure | AAMA 2605, as summarised by a coatings manufacturer | ΔE between exposed and unexposed areas | Ten-year South Florida exposure | Colour change of at most 5 ΔE after ten years | A durability limit. It is several times wider than any batch tolerance and must never be borrowed as one. |
| Paint film against a reference standard or a freshly prepared standard | ISO 3668:2017 (3rd edition, confirmed 2022); ASTM D1729-22 for the equivalent practice | Visual comparison in a standard booth under artificial light | Diffuse artificial daylight in a booth; ASTM D1729 recommends simulated D65, or D50 where printing or photography is involved | None verified; agreed by the parties | A procedure for looking, not a limit. Acceptability is expressed by agreed limit panels or by an instrumental tolerance set separately. |
| Setting colour and gloss tolerances for opaque, non-fluorescent coatings | ASTM D3134-15(2024) | A procedure for arriving at tolerances around a colour standard and a gloss value | Non-fluorescent opaque specimens | None verified; agreed by the parties | The practice states that it does not indicate how large tolerances should be; purchaser and seller agree them. |
| A colour match that must hold under a second light source | ISO 18314-4:2024 (2nd edition) | Metamerism index: the colour difference of the pair under a test illuminant | Solid surface colours; not applicable to effect coatings without adaptation | None verified; agreed by the parties | The standard gives the calculation. Which test illuminant matters, and how large an index is tolerable, depend on where the product is seen. |
Why: The production film is thinner and does not fully hide a pale substrate, or its gloss is lower.
Fix: Measure film thickness and gloss on both, and prepare approval panels by the production method at production thickness.
Why: The new batch uses different pigments and is metameric to the weathered original.
Fix: Ask for the original pigmentation, or specify a metamerism limit under the relevant second illuminant.
Why: A change-after-exposure limit has been read as a batch tolerance, or the reverse.
Fix: Quote the clause, not the number: uniformity against the standard and retention after exposure are separate requirements.
Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.
StandardStrong evidence
ASTM D3134-15(2024) covers establishing tolerances for both colour and gloss of non-fluorescent opaque specimens, on the basis that specifications are centred on exact positions in colour space and on the gloss scale that cannot be achieved repeatedly.
Source: ASTM D3134-15(2024) Standard Practice for Establishing Color and Gloss Tolerances
StandardStrong evidence
ISO 6504-3:2019 determines the hiding power of white or light-coloured paint coats by measurement over a black and white chart or over a colourless transparent foil placed on black and white panels.
Caveat: Scoped to paints for masonry, concrete and interior use; cited for the black-and-white principle.
ConventionLimited evidence· North America
A trade-journal summary reports that the AAMA 2604 and 2605 specifications recommend a maximum deviation of 2 ΔE from an agreed colour standard for solid paint colours, and that exact colour uniformity should not be expected of coatings containing mica or metallic flake.
Caveat: Secondary source; the article gives no formula, illuminant, geometry or edition. Applies to North American architectural aluminium finishes only.
Source: Design considerations for metallic finishes on architectural aluminum
StandardModerate evidence· North America
A coatings manufacturer summarises AAMA 2605 as requiring a ten-year South Florida exposure with a colour change of no more than 5 ΔE, gloss retention of at least 50 % and a chalk rating of at least 8.
Caveat: A manufacturer's summary, not the specification's text. It is a limit on change after weathering and is not a batch tolerance.
Source: FGIA/AAMA 2605 (architectural coatings overview); AAMA 2605-22 Voluntary Specification, Performance Requirements and Test Procedures for Superior Performing Organic Coatings on Aluminum Extrusions and Panels (with Coil Coating Appendix)
StandardStrong evidence
ISO 18314-4:2024 specifies a calculation of illuminant metamerism for pairs of solid surface colours as a colour difference, intended as a quality measure for specifying tolerances under different illumination, and cannot be applied to effect coatings without adaptation.
StandardStrong evidence
ASTM D4086-18(2023) describes visual methods for detecting illuminant and observer metamerism and estimating its magnitude, and excludes gonioapparent samples.
Source: ASTM D4086-18(2023) Standard Practice for Visual Evaluation of Metamerism
Reviewed 6 October 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.