Files
craft-libs/crates/cms/tests
Khawar Zamman WaniandDylan M. Taylor 872e001552 cms: Coated FOGRA39 separates like the reference application (#2028)
* cms: Coated FOGRA39 separates like the reference application

Coated FOGRA39 gave a saturated blue about 97 % magenta instead of about
75 %, a mid grey about 30 % black instead of about 20 %, and Saturation
reused the Perceptual table.

- GamutMapping::Oklab (opt-in, CmykParams::mapping): out-of-gamut colours
  take the closest colour the printer reaches in Oklab, lightness and hue
  weighted above chroma and more so the further outside the colour is;
  black is generated from the clipped colour. Perceptual and saturation
  map the v4 perceptual reference black onto the device black for
  neutrals; perceptual compresses Oklab chroma, saturation gets its own
  table.
- CmykParams::gcr_linear: a linear part of the black generation curve.
- Legacy (the default) keeps the Photocraft profile's tables: its
  generator output is byte-identical to main's.
- The FOGRA39 built-in is regenerated with the new mapping and curve.

For storytold/photocraft#14267.

* cms: Oklab round-trip test tolerates the published matrices' precision

* cms: FOGRA39 keeps in-gamut round trips within the existing limits

The first version broke in_gamut_lab_round_trips (max 7.7 ΔE00): dark
colours got too little black to be reached, and clipped nodes next to
light tints and dark colours differed from their in-gamut neighbours, so
interpolation spoiled colours between them.

- Black generation is a curve of points (CmykParams::gcr_curve) instead
  of a linear term: red, mid grey and dark colours each get the black
  the reference and the round trip need.
- The clip keeps the plain separation when that is closer, tries the
  separation's black as well, and blends to the separation near paper
  white (L* 85 to 95).

* prepress: follow the Coated FOGRA39 round trip the new profile gives

---------

Co-authored-by: Dylan M. Taylor <dylan@dylanmtaylor.com>
2026-11-17 14:14:36 -05:00
..
…
…
2026-11-12 10:38:37 -05:00
…
…