
RawTherapee's Green Equilibration addresses differences between the two green populations in Bayer sensor data. Those differences can produce maze-like interpolation artifacts when a demosaicer assumes identical green responses. It is not a way to desaturate green objects or geometrically move red and blue fringes into place.
Identify the raw-data problem
The RawPedia preprocessing documentation describes slightly different green-filter responses and green separation from optical crosstalk. A shifted technical-camera lens or an adapted wide-angle film lens can be relevant, but the equipment alone does not prove the diagnosis.
Inspect the raw at a useful detail view. Compare fine repeating structures and smoother regions. Ordinary fabric moiré, sharpening texture and green-split mazing should not automatically receive the same explanation.
The demosaicing artifact guide provides background on algorithm-dependent detail, while this article isolates a particular Bayer input imbalance.
Find the Bayer preprocessing control
RawPedia separates general preprocessing from the settings inside Sensor with Bayer Matrix in the Raw tab. Use the controls belonging to the actual sensor type; a Bayer correction is not a generic X-Trans adjustment.
Green Equilibration's threshold determines which neighboring green values are equalized. The reference advises increasing it only enough for the mazing to disappear. It does not supply a universal value for every camera or source.
For a proposed comparison, retain the unchanged processing state, adjust this control alone and inspect the same fine pattern. Stop increasing it when the diagnosed artifact is no longer visible; “maximum” is not the documented target.
Distinguish the alternative demosaicing route
The Demosaicing reference describes VNG4 as a possible response to relevant crosstalk cases, with some fine-detail cost. Treat that as an alternative comparison, not an instruction to stack multiple changes without inspecting their effect.
Hold exposure, sharpening and output size steady while comparing. If an alternative reduces a maze but erases needed lettering, it has not automatically met the photographic requirement.
Keep geometric fringing separate
Green equilibration changes a response relationship before interpolation. Lens chromatic-aberration correction concerns a different problem and should be judged on its own evidence. The raw CA versus lens TCA guide explains that distinction in another application; those controls are not substitutes for the two-green response correction here.
If the issue is heavy crosstalk with dull color rather than the isolated maze pattern, do not promise this threshold restores every contaminated channel. RawPedia describes broader limits to crosstalk correction.
Accept a proposed rendition only after checking both the artifact and the detail that matters. No Bayer calibration, corrected camera sample or measured threshold comparison was performed here.
Related tasks
- Why Do Fine Lines Turn into Colored Moiré After Demosaicing in darktable 5.6?
- Should You Use Lens TCA Correction or raw Chromatic Aberrations in darktable 5.6?
Sources and scope
- RawPedia: Preprocessing — Green Equilibration: Different responses of the two Bayer green filters can cause interpolation mazing; threshold equilibrates sufficiently close neighboring green values; raise only enough for mazing to disappear; crosstalk can also split the green channels.
- RawPedia: Demosaicing — VNG4: Near-symmetrical shifted wide lenses or adapted film lenses can cause crosstalk/green separation; VNG4 trades some fine detail, and Green Equilibration is an alternative.
Research checked on 2026-10-23 (Asia/Shanghai) against actually opened official manual/RawPedia pages. No camera capture, command execution, raw processing, calibration or display test was performed. Examples and acceptance checks are proposed, and the cover is decorative.
Prepared by Imagild Editorial, the company editorial team responsible for these guides. AI assisted research and drafting. This is documentation-based guidance with proposed examples, not a claim of hands-on testing or an independent product review. Report a correction.
Cover: 85 LENS (Ngô Hoàng Auth) / Pexels, Pexels License. This is an illustrative source photograph, not an application screenshot or a generated test result.