How to Analyze a Reference Photo’s Shadow and Highlight Colors for Color Matching

Start by deciding what you want to match: the reference’s overall appearance, the color of a particular object, or its relationship between shaded and illuminated surfaces. For this workflow, the goal is the overall appearance, with particular attention to how color changes from shadows to highlights.

Keep the reference unchanged while analyzing it. Neutralizing its whites immediately could remove the tint you want to reproduce. The darktable manual distinguishes between correcting unwanted casts and adding the final creative appearance; treating those as separate decisions helps clarify what you are trying to copy. darktable: color balance rgb

Open the reference and your target in the same editor, against the same neutral gray background. Keep display brightness and room lighting consistent during comparison. darktable’s workflow guidance recommends a calibrated monitor, white environmental lighting, and a background approximating middle gray for color assessment. darktable: process

Check the documents’ color profiles before comparing RGB numbers. Profiles determine how those numbers translate into visible colors. If you need a shared RGB space for numerical comparison, convert working copies to that space. Assigning a different profile keeps the numbers and changes their interpretation; conversion changes the numbers to preserve appearance as closely as the conversion permits. Adobe: Change a document’s color profile

Begin with a material that appears in both light and shade: a matte ceramic pot, a painted panel, or a sheet of paper. Comparing two areas of the same material reduces one obvious ambiguity: whether a color difference comes from comparing different objects.

Do not assume shadows must be blue or highlights must be yellow. The darktable manual explains that directly illuminated areas and areas receiving reflected light can have different color temperatures, and a single global white balance may not reconcile them. Your task is to observe the relationship present in this reference. darktable: color balance

Use the following sample roles as a practical selection guide. These are proposed analysis steps, not universal rules about where every photo’s colors should fall.

Sample role	Choose	What you want to learn
Main shadow	A shaded patch with visible texture	The usable shadow tint
Midtone	A moderately lit patch of the same material	The transition between dark and light
Main highlight	A bright patch retaining texture and color	The illuminated surface’s tint
Cross-check	Another suitable material in comparable lighting	Whether the pattern extends beyond one object

Keep tiny glossy reflections separate from the main highlight sample. Likewise, avoid making the deepest crevice your only shadow reference. Your main samples should describe substantial, usable areas of the photograph; you can record small extremes separately if they matter to the look.

A sample reading pure white cannot tell you the original surface’s hue. In 8-bit RGB, equal maximum channel values produce white, while equal zero values produce black. Move to a nearby patch that retains variation when the chosen endpoint provides no useful color distinction. Adobe: Color modes in Photoshop

Use an averaged sample rather than relying on a single pixel. Adobe documents that Point Sample reads an individual pixel, while options such as 3 by 3 Average and 5 by 5 Average smooth readings by averaging surrounding pixels. This reduces the influence of noise and minor variations. Adobe: View pixel color values in an image

In Photoshop, use this sequence:

1. Open Window > Info.
2. Select the Eyedropper and choose 5 by 5 Average as a starting point.
3. Shift-click to place persistent Color Samplers on your chosen patches.
4. Set the sampler readouts to Lab for lightness and color analysis; record RGB too if useful.
5. Label each recorded value with its material and lighting role. Adobe: View pixel color values in an image

Make sure the averaging area stays inside the patch. A sample straddling a yellow object and a blue background answers a different question from a sample contained within either surface. Enlarge the sample only when the surface is broad enough to accommodate it.

Check two or three nearby positions within each chosen region. If they suggest the same color direction, retain one representative reading and note the variation. If they disagree strongly, inspect the area for changing illumination, texture, or mixed materials before treating it as one color.

Lab readouts are useful because they separate lightness from two color axes. L* describes lightness on a 0–100 scale; a* describes the green–red axis, and b* the blue–yellow axis. Positive a* points toward red and negative a* toward green; positive b* points toward yellow and negative b* toward blue. Adobe: Color modes in Photoshop

Read each sample in two passes. First, compare L* to understand how dark or bright the patch is. Then inspect a* and b* to describe its color direction. This avoids confusing a brightness mismatch with a tint mismatch.

For example, describe a reference as “dark shadows leaning blue-green, with brighter highlights leaning yellow.” That is more actionable than “cinematic colors,” and more flexible than a pair of unlabeled hex codes.

Look for repetition before making a global conclusion. If several suitable shadow samples lean blue-green, that supports using the direction as an initial grading hypothesis. If only one shaded object does, keep the conclusion local. This is an inference from your sample pattern, not proof of how the reference was illuminated or edited.

The following values are invented solely to demonstrate the method. Assume both photographs contain comparable matte gray surfaces, sampled through a consistent color-managed workflow. These are not measurements from an actual reference.

Patch	Reference L*, a*, b*	Target L*, a*, b*	Initial interpretation
Shadow	25, −4, −8	34, 0, −2	Target is lighter and less blue-green
Midtone	52, 0, 2	56, 1, 3	Target is slightly lighter, with a similar color direction
Highlight	83, 3, 12	80, 1, 4	Target is darker and less yellow, with less red

The lightness differences suggest addressing tonal separation first: the target’s sampled shadow is too light, while its highlight is slightly too dark. After that adjustment, sample again before deciding how much color work remains.

For color direction, subtracting the target values from the reference gives a shadow difference of −4 on a* and −6 on b*. The highlight difference is +2 on a* and +8 on b*. In this illustration, that suggests moving shadows toward green and blue, and highlights primarily toward yellow with a smaller redward shift.

These differences are diagnostic notes, not slider settings. Entering “−6” into an unrelated grading control would not reproduce a six-unit b* change. Instead, make a small adjustment, observe the new sample readings, and inspect the surrounding image.

If the two photos contain different materials, exact sample equality becomes a weaker goal. Use the example’s reasoning to identify direction and tonal separation, then judge whether the target’s own objects still have the intended appearance.

Write a short instruction from your findings before touching color controls: “Deepen the main shadows, add a restrained blue-green shadow tint, and shift textured highlights toward yellow while keeping midtones close to their current appearance.” This becomes a clear starting hypothesis that you can revise after comparison.

Work in three passes: establish the broad lightness relationship, adjust shadow and highlight color, then refine the transition through midtones. Recheck your samples after each pass so that one change does not conceal another mismatch.

Use controls that let you inspect which tonal areas receive the adjustment. For example, darktable 5.6’s color balance rgb module separates tonal regions with masks. Its middle-gray fulcrum controls their division, fall-off controls soften or tighten transitions, and mask previews show affected areas. Those controls matter when a shadow tint spreads too far into midtones. darktable 5.6 manual: color balance rgb

If a color relationship appears on just one object, begin with a local adjustment. If it repeats across several materials and lighting regions, try a broader tonal adjustment and check each material again. Treat the scope of the adjustment as part of the matching problem.

Use the following diagnostic sequence when the edit feels close but does not yet resemble the reference. It connects each visible symptom to a specific check, rather than encouraging more saturation everywhere.

What you notice	Check next	Practical response
Tint seems close, but the image feels too flat	Shadow and highlight lightness	Refine tonal separation, then resample
One object matches while others drift	Whether samples represent different materials or lighting	Narrow the adjustment’s scope
Midtones acquire the shadow tint	Tonal mask coverage and transition	Reduce overlap or soften the adjustment
Nearby sample readings disagree	Sample boundaries and surface variation	Reposition the sample within a coherent patch
Matching RGB numbers gives a different appearance	Document profiles	Resolve the comparison setup before further grading

Finish by comparing the full images and then inspecting the sampled regions at a useful detail view. Ask whether the shadow-to-highlight color direction matches, whether the lightness relationship is close enough for your purpose, and whether the target’s materials remain convincing.

Save your labeled readings together with a one-sentence description of the relationship you matched. Those notes make later revisions easier: you can return to the intended shadow tint, highlight tint, and tonal separation without reconstructing the analysis from scratch.

• Adobe — View pixel color values in an image
• Adobe — Change a document’s color profile
• Adobe — Color modes in Photoshop
• darktable user manual — process
• darktable user manual — color balance
• darktable 5.6 user manual — color balance rgb