Further discussion:
Question: Using an adobe profile, if we perform adjustments to the brightness histogram to counteract the adobe tone curve, is that the same as using a linear profile, or will we introduce multiple corrections which introduce banding and other artifacts ?
ChatGPT: Good question—and the short answer is: no, it’s not the same as using a true linear profile, and yes, trying to “undo” Adobe’s tone curve can introduce subtle issues (though not always catastrophic ones).
Let’s unpack why.
1) Why “undoing” the tone curve isn’t equivalent
In Adobe Camera Raw, the tone curve in a profile is applied very early in the rendering pipeline—before most of your slider adjustments.
That has two important consequences:
a) The transformation is not perfectly invertible
The Adobe tone curve:
Compresses highlights (rolloff)
Expands some midtones
Lifts shadows selectively
This is not a simple gamma curve—it’s often piecewise and perceptual
When you try to reverse it using sliders (Exposure, Tone Curve, etc.), you’re:
Applying a different curve later in the pipeline
Trying to approximate the inverse, not actually remove the original
So mathematically:
You get “curve A + inverse-ish curve B,” not “no curve.”
b) Information has already been redistributed
This is the more subtle issue:
Highlight compression may have already squeezed tonal differences together
Shadow lifting may have already spread noise apart
Even if you “reverse” the shape visually:
You don’t recover the original distribution of values
You’re working on post-shaped data, not raw-linear data
2) Does this cause banding or artifacts?
Usually: not severe—but possible
Modern raw pipelines (including Adobe’s) use:
High bit depth (often 16-bit or floating point internally)
Smooth curve math
So:
You won’t usually see obvious banding just from compensating curves
However, you can introduce:
a) Micro-contrast distortions
Slightly unnatural tonal transitions
Especially in skies or smooth gradients
b) Highlight reconstruction differences
Once highlights are rolled off, reversing the curve:
Does not restore clipped or compressed relationships perfectly
c) Noise reshaping
Shadows lifted → then pushed back down ≠ untouched shadows
Noise can look different than with a linear profile
3) The deeper issue: pipeline order
Think of it like this:
With Adobe profile:
RAW → tone curve → your adjustments
With linear profile:
RAW → your adjustments
If you try to “cancel out” the curve:
RAW → tone curve → inverse curve
That is not the same as skipping it entirely, because:
Operations are not commutative
Some steps are lossy (especially highlight compression)
4) When the difference actually matters
You’ll notice it more when:
Doing technical evaluation (like with RawDigger)
Matching renders across software (e.g., RawTherapee)
Performing precise tone mapping or scientific imaging
You’ll notice it less when:
Making normal photographic edits
Working within moderate dynamic range
Outputting to JPEG/web
5) Practical takeaway
Can you approximate linear behavior using Adobe profiles + adjustments?
→ Yes, visually.
Is it technically the same as a true linear profile?
→ No.
Will it break your images?
→ No—but it can subtly alter tonal relationships and noise.
A useful mental model
Think of Adobe’s tone curve as:
“baking contrast into the negative”
Trying to undo it later is like:
“editing a baked image to look unbaked”
You can get close—but you’re not working with the same underlying material.