Short answer
It should contain everything needed for two firms to measure the same batch and reach the same verdict: the reference (a physical standard, its recorded spectral values, or both), the formula with its parameters, the instrument geometry and colorimetric condition, how samples are prepared, the acceptance rule, and what happens in a dispute. The published standards deliberately leave the limit itself to the two parties. A digital standard strengthens the agreement by fixing the target as spectral data, but it carries neither the tolerance nor the surface.
A reader looking for the permitted colour difference in a standard usually finds a method instead. ASTM's practice for establishing colour and gloss tolerances says in terms that it does not indicate their extent and that purchaser and seller should agree them. Its practice for calculating differences requires the two to agree both the tolerance and the calculation. The textile standard for CMC describes a single maximum set by the end use, without naming one. The automotive recommended practice is written as a common footing for buyer and seller across paint, trim and textiles. This is not evasion. The acceptable difference depends on the product, its neighbours and its price, so the standards fix the language and leave the bargain to the contract.
Most colour disputes are disagreements about something the agreement never said. A complete agreement names the reference and who holds the master; the formula and every parameter; illuminant, observer, geometry, specular mode and aperture; the sample form, backing, conditioning and number of readings; the rule for acceptance, including whether the limit applies to one reading or an average and whether an action margin exists; a second illuminant and a metamerism limit where the product is seen under shop lighting; a gloss tolerance; and a referee procedure naming the instrument or laboratory whose reading settles a disagreement. Where an instrument cannot capture the property, as with effect finishes, physical limit samples showing the lightest, darkest and furthest-off acceptable pieces still do the job.
A digital standard replaces 'match this swatch' with 'match this reflectance curve'. The Colour Exchange Format, CxF, is an XML format that stores spectral data together with the metadata needed to interpret it: device, light source and measurement condition. It has been standardised as ISO 17972, whose fourth part covers the characterisation of spot-colour inks. Because the target is a spectrum and not a single CIELAB triplet, the recipient can recompute it for any illuminant, observer or difference formula, check for metamerism before making anything, and avoid the copying errors that creep into re-measured swatches. Textile supply chains use the simpler QTX text format for the same purpose, and a separate appearance format carries gloss, texture and effect data that a spectrum alone cannot.
Three limits matter in practice. A spectral target was produced by one instrument geometry, and a recipient with a different geometry is not measuring the same quantity; the agreement has to fix the geometry or accept a correlation exercise. The ISO spot-colour format is explicitly limited to paper-like substrates and to the ink and substrate combination that was characterised, so a curve measured on coated board says little about film. And a spectrum is silent about surface: two pieces with identical curves and different gloss will not look alike. A digital standard also is not a tolerance. The file says where the centre is; how far a batch may sit from it is still a clause in the agreement.
A tolerance tighter than the disagreement between the two firms' instruments cannot be enforced, however reasonable it looks. Makers publish agreement between instruments of the same model as an average over a set of ceramic tiles, around a tenth of a CIELAB unit in the data sheets Colourwise read, and say nothing about agreement with other makers' models. The contract-proofing standard in print addresses the gap directly: it treats limits below 2.5 ΔE00 as impractical between different models and recommends halving its tolerances when one instrument measures both reference and sample. The practical step before signing is a correlation study: both parties measure the same stable tiles and the same real samples, and the limit is set above what that study shows.
Why: They measure with different geometries or specular modes, so each compares its own reading with a target made another way.
Fix: State the geometry in the agreement, and run a correlation on shared tiles and real samples.
Why: Gloss or texture differs; the target carried colour only.
Fix: Add a gloss tolerance and a physical standard for surface, alongside the spectral one.
Why: The spectral characterisation is valid only for the ink and substrate it was measured on.
Fix: Characterise the ink on the production substrate, or agree a substrate-specific target.
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) describes a procedure for establishing colour and gloss tolerances for non-fluorescent opaque specimens, states that it does not indicate the extent of tolerances, and leaves them to agreement between purchaser and seller.
Source: ASTM D3134-15(2024) Standard Practice for Establishing Color and Gloss Tolerances
StandardStrong evidence
SAE J1545 applies to opaque or nearly opaque vehicle parts and materials intended to match a colour standard, covering topcoat paint, interior soft trim, interior and exterior hard trim and exterior film, and excludes transparent materials.
Caveat: From the public scope of the edition reaffirmed in December 2021.
StandardStrong evidence
ISO 17972-4:2018 defines an exchange format for spectral measurement data of spot-colour inks, limited to isotropic, paper-like substrates and to applications where the characterised ink and paper combination is the one used in printing.
FactModerate evidence
CxF is an XML-based format that stores spectral colour data with metadata such as device, light source and measurement conditions; QTX is a text format for spectral reflectance used in textile quality control; AxF is a binary format for appearance properties including gloss, texture and effect.
Caveat: Described by the company that originated CxF and AxF.
Source: CxF, AxF, and QTX: Essential File Formats Powering Digital Textile Color Management
StandardStrong evidence
ISO 12647-7:2016 treats ΔE00 tolerances below 2.5 as impractical between instruments of different models and recommends halving its tolerances when the same instrument measures both sets of values.
Caveat: Written for contract proofs in the graphic arts; the principle, not the figure, carries to other industries.
Source: ISO 12647-7:2016 … — Part 7: Proofing processes working directly from digital data
StandardStrong evidence
ASTM E2214-23 provides standard terms and procedures for specifying and verifying the performance of colour-measuring instruments; it gives the format in which a specification is determined, communicated and verified and does not itself set specifications.
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.