How do Photoshop, Lightroom, and similar editors handle color spaces for tools like curves, blending, and eyedroppers?
Asked 1/5/2017
5 views
2 answers
0
I’m trying to understand how common editing tools operate with respect to color space, gamma, and white point. For example: curves/levels, blend modes, channel mixer, HSL/HSV controls, and eyedropper/color readouts in apps like Photoshop, Lightroom, Premiere Pro, or GIMP.
These operations all act on image values, so I’d like to know whether they work in the document’s working RGB space, in linear light, or in some other representation. In particular, when I’m trying to do physically meaningful edits such as additive compositing or haze removal, it matters whether the software is operating on gamma-encoded values or linear values.
Why don’t editors make this more explicit, and is the general rule simply that I need to set the working space I want and assume tools operate on those RGB values unless otherwise documented?
Originally by Photography Stack Exchange contributor. Source · Licensed CC BY-SA 4.0
Photography Stack Exchange contributor
9y ago
2 Answers
6
Most editing software cares very much about colorspace. The issue is that users don't usually understand it (because it's very complicated), and even when they do, they often expect contradictory results from their software.
For example, Adobe Photoshop allows a user to set the working color space for manipulating their images. The vast majority of PS users probably want the results they've always gotten from the application, even before it had the functionality of setting the working colorspace. That means working in sRGB, even if it's not physically representative of how light works. In fact many of the blend modes, for example, are tailored to imitate linear light blend modes with non-linear RGB values. Screen in sRGB gives similar results to Add in linear RGB.
Fast forward to present day where many apps let you set the working space. Let's say a user thinks they want to use linear RGB to be more physically accurate. Suddenly nothing works like it used to! The Screen blend mode produces weird results! Contrast looks completely different! No amount of telling users, "This is what you asked for" will sell more copies of that software if it produces unexpected results.
And when a particular function works in a different colorspace internally, there's rarely a need for the user to know that. If it produces the results the user wants, then does it matter if it's HSV internally or HLS? or YCbCr? Or L * a * b?
Originally by user22895. Source · Licensed CC BY-SA 4.0
user22895
9y ago
0
Generated from our catalog & community — verify before relying on it.
Most editors do care about color space, but they often hide the details because most users neither need nor want that complexity.
General rule: many tools operate on the image’s current RGB values in the document/working space, unless the software or specific tool says otherwise. If you want a particular behavior, set your working space accordingly. In Photoshop, for example, the RGB working space is user-selectable, and the document profile can be displayed in the interface.
Some operations are not intended to be physically accurate light math. Blend modes are a good example: in Photoshop, many were designed to give familiar visual results in gamma-encoded RGB rather than strict linear-light behavior. So “screen” in sRGB may resemble “add” in linear RGB, even though they are not the same thing.
HSL/HSV controls are best thought of as alternate ways of selecting/manipulating RGB values, not independent color spaces with their own physical meaning.
So yes: if linear-light or a specific RGB space matters for your workflow, you usually need to configure that deliberately and check the app’s documentation/UI for profile readouts. The reason this is not always prominent is likely simple demand and usability.
Recommended products
UniqueBot
AI9y ago
Your Answer
Related Questions
In UFRaw, what are the base curve and correction curve used for?
Can Lightroom apply a precise numeric gamma correction, such as converting linear scans from gamma 1.0 to 2.2?
How do RGB values on a QPcard 201 gray patch translate to percent gray?
How should lens flare be composited: additive blending or something else?
Is gamma-correct resizing really an issue, and is the article's test valid?