Why Lab mode instead of RGB for this
Lab separates an image into three channels that don't overlap in what they represent: L is lightness only, no color at all, while a (green to magenta) and b (blue to yellow) carry color only, with no brightness information in either. In RGB, every channel mixes tone and color together, so correcting a color cast by nudging one RGB channel almost always shifts perceived brightness too, forcing you to compensate across multiple channels at once by eye. In Lab, you can move color without touching tone at all, which is exactly what a precise match needs.
The tool that makes this precise: Color Sampler plus Curves on a and b
Eyeballing color in the a/b channels is still just eyeballing, more precise than RGB but not exact. The Color Sampler tool (in the Eyedropper's tool group) is what turns this into a real numeric match.
1. Place two Color Samplers
Click a Color Sampler point on your target reference color, the one you're matching to, and a second one on the color you want to shift. Both now show live before/after readings in the Info panel as you make adjustments.
2. Add a Curves layer and switch to the a channel
With the Curves dialog open and the channel dropdown set to a, click directly on your source sampler point on the canvas. That places a point on the a-channel curve at that pixel's exact current value. Watch the Info panel's "after" reading for that sampler and drag the point up or down until it matches your target sampler's a value exactly.
3. Repeat on the b channel
Switch the Curves channel dropdown to b and do the same thing: click the source point on the canvas, drag until the after value matches the target's b reading.
4. Leave L alone
Don't touch the Lightness curve at all. That's what keeps the whole correction tonally neutral, only the color shifted, brightness and contrast in that area stayed exactly where they were.
Why this beats eyeballing it in RGB
An RGB curve on any single channel mixes tone and color by definition, so fixing a cast by pulling down Blue, for instance, also darkens that area's perceived brightness, which then needs its own separate compensation. In Lab, a and b are single-purpose axes with no tonal side effect, so you converge on an exact numeric target in two clean moves instead of iterating across three interdependent ones.
When to actually use this
This earns its extra steps for jobs where precision matters more than speed: matching a skin tone patch shot under two different lighting setups within the same composite, matching a product's exact color across several photos for consistency, or neutralizing a color cast in one specific area against a known-neutral reference elsewhere in the frame.
Quick answers to a few related questions
Do I need to stay in Lab mode for the rest of the edit?
No. Do the correction, then convert back to RGB (Image → Mode → RGB Color) for everything else. Flatten any pending Smart Object edits first, since mode conversion applies to the document as it currently renders.
Can I skip the Color Sampler and just eyeball the a/b curves?
You can, but you lose the exact numeric target that's the whole point of this method, you're back to judging the match visually, just on cleaner axes than RGB.
Does this work for matching more than one color at a time?
Photoshop allows up to four Color Sampler points at once, so you can track several reference pairs in the Info panel simultaneously. Each precise a/b match described here still targets one source color to one target color at a time.
EN
IT