I have not come across any option in Photoshop that can separate colors strictly according to theoretical principles.
Here’s the basic theory that should ideally be followed:
1. Tertiary colors: These are created by mixing primary and secondary colors, as shown in image (A). For example, orange is a mix of red and yellow. If red is completely removed, as in image (B), it should also be removed from orange, leaving only a shade of yellow.
2. Complete removal of primaries: If red, green, and blue are entirely removed from an image, as in image (C), the affected areas should appear gray. The luminance values should follow the correct order: blue (lowest), red, then green.
According to the standard grayscale conversion formula Y = 0.299R + 0.587G + 0.114B, the expected luminance values are:
-Blue: 29
-Red: 76
-Green: 150
So far, I haven’t found any option in Photoshop that truly separates colors according to theory. Is there a recommended approach, workaround, or third-party solution that achieves this?
Explore related tutorials & articles
17
Replies
I wonder if any of the Luminosity Masks would come close.
I only have Sven Stork's Luminosity Mask, but the Tony Kuper extension is supposed to be the goto tool. You can contact Tony here
Probably not exactly what you need, but still.
Thank you for your reply. I have already tested a similar approach in Photoshop. The result looks close to what we want, but only because the example uses solid colors that are easy to identify. In practice, this method isn’t reliable—especially when applied to gradients, as shown in the example below—and it becomes even more challenging when the colors are less defined or difficult to identify.
I expected someone to point out an option in Photoshop for this, but am I asking for something that’s too difficult to achieve?
…since it seems like a very basic and important feature?
By @vinay kumar vishwakarma
What are some examples of its importance?
I’m asking only to understand the benefits of it, because apparently such a feature has not been considered a high priority during the over a third of a century that Photoshop has existed. Photoshop was originally designed as a photo editor (later extended for graphic design and other uses), so in just about all intended use cases Photoshop is supposed to support a creative workflow. From that point of view, the most important criteria of Photoshop color adjustment have been:
Best possible perceptual color. Objectively achieving numerically accurate color is important as a baseline for editing. But for final delivery of color, users and their clients place a much higher priority on color that meets aesthetic design goals and is pleasing. This drives color processing that is often perceptual and adaptive, taking into account the adaptive, nonlinear quirks of human color perception.
Best translation to delivery color. A major goal is to convert original color values to the color values that must be delivered to clients or audiences. That involves a mandatory translation to the color values in the final media, and correctly optimized for how they reproduce colors. This drives processing that translates color into the values that work for specific combinations of ink, paper, and printer; or specific color gamuts and specifications for web, mobile, or pro video standards. This very much takes the goal of color processing away from the theoretical/academic, and toward the real world.
Both of those criteria also led to the development of ICC color profiles, which again filled a real-world need to represent theoretical/idealized color values consistently across varying real-world constraints.
Although there are a few people who use Photoshop for academic/scientific color analysis, in general Photoshop color tools have not been built for that, they’ve been built for human perception and to adjust color values to look their best on specific delivery media. That’s what leads to my question, how would the type of adjustment you need be important for commonly done color tasks?
Also, is there a chance Adobe might add a simpler option for this in the future, since it seems like a very basic and important feature?
As others already hinted at: This seems neither basic nor important (to some Photoshop users at least).
Please explain the actual scenario in which this matters to you.
Colors are subjective. In theory, you could remove red by setting the R component of every RGB pixel to zero. In practice, that's going to be a mess, especially when you factor in different colorspaces.
By @ExUSA
I thought about this too. For RGB, many here would interpret the phrase “Complete removal of primaries” to mean the values of all RGB primaries being set to 0, which means black, but that isn’t what’s being requested.
The first post seems to actually request removing all color components (not primaries), and leave the luminance component? In that case it isn’t a straight RGB edit. Instead, it might be a question about which desaturation method to use out of the several available in Photoshop.
Answer to the question "Please explain why this is important and basic?"
Why Basic?
The very first thing we learn is that RGB are the primary colors, and CMY are the secondary colors. Together, primary and secondary colors form tertiary colors.
For example:
Orange = Red + Yellow
Violet = Blue + Magenta
If we remove the primary colors, we are left with shades of gray. According to the standard grayscale conversion formula: Y = 0.299R + 0.587G + 0.114B,
the expected luminance values are approximately:
-Blue: 29
-Red: 76
-Green: 150
This is basic knowledge because even people who are not experts in image processing know it.
Why Strict Theoretical Color Separation is Important
1. A perfect way to identify the problem.
When beginners start learning color correction in Photoshop, it is unrealistic to expect them to produce perfectly color-corrected images from day one. Developing a good sense of tone and color perception typically takes time—sometimes weeks, months, or even years.
However, should we expect a beginner to identify color issues in an image from the very beginning?
Surprisingly, the answer is yes—if the color in the image is separated based on a strict theoretical approach. Even someone new to color correction can quickly analyze and understand the components of the image, far more effectively than by relying on histograms or RGB channel observation alone.
Let’s consider a simple example: In the left image below, there's a noticeable color variation on an elderly man's face. A beginner may struggle to determine whether this is a single color, a combination of two, or even three different colors blended together. But if a theoretical color separation tool is available, they can isolate and remove colors one by one. This makes it significantly easier to identify the color components and select the appropriate correction tool.
Importantly, the correction in this image was achieved without any masking or selection—thanks entirely to theoretical color separation. (Could you please try this in Photoshop or Lightroom without using selection and masking? Feel free to use any other application or plugin that you prefer.)
2. Helps in developing image processing algorithms.
Below is an image of a red car with visible reflections. The objective in this case is to reduce the luminosity of the red color.
Image A shows the original, unaltered image.
Image B illustrates a reduction in red luminosity using the Selective Color (Relative) adjustment method.
Image C attempts to achieve a similar effect by theoretically isolating the red.
During experimentation, it was observed that the gradation in shadow areas appears smoother and more natural, suggesting improved tonal transitions in those regions.
3. It can be revolutionary in biomedical research.
The image below illustrates a brain mapping analysis. Image 'A' represents the original, while Image 'B' shows the same image with red color filtered using the traditional hue range method. Image 'C' depicts the same image, with red filtering applied using the theoretical color separation.
A comparison reveals that the traditional filtering method results in the loss of subtle red information, particularly lighter shades of red. In contrast, the theoretical color separation method preserves all red components, including those within the orange spectrum.
----------------------------------------------------------------------------------------------------------------------------------------------------
I have included the section on biomedical research, as several individuals expressed interest in understanding the relevance of Theoretical Color Separation.
----------------------------------------------------------------------------------------------------------------------------------------------------
Some technologies reveal their true potential only when we experiment with them.
Thanks for the elaborate answer.
Alas, I am not sure there might not be misunderstandings on either side.
I suspect you might mix up concepts from different Color Models and you might be disregarding Color Management completely. (Though the latter might indeed be irrelevant in your use cases.)
(Edit: Or I might have misunderstood your explanations.)
As for medical imaging I am afraid your argument seems pointless to me.
As far as I know CT, MRI, PET, … produce one channel images to begin with and the images you are showning have already be processed/colorized, so the prudent approach to isolate regions of specific luminosities would probably be to start with the original grayscale images.
As for my last test one could change the Curves even more harshly – see screenshots; I have not drawn the red curve numerically exactly but if someone is willing to do the math and do it manually (or via a Script) it might be possible to get a more exact match.
1. I can understand your concern about the different color models and color management, as you are using HSB, HSL, and Hue/Saturation. However, for the simple task of separating colors, you’ve made things unnecessarily complicated. There is no need to use HSB, HSL, or Hue/Saturation.
2. Just to clarify, I was referring to biomedical research, not specifically to medical imaging.
The example image used in the biomedical research section is from QEEG (Quantitative Electroencephalography). To the best of my knowledge, in practice, most clinical and research QEEG reports use color-coded brain maps rather than grayscale images.
After reviewing several research papers on QEEG, one common observation is that colors like red and blue are often selected using a range in hue, as shown in the example below.
Moreover, similar use of color can be seen in pathology slides, where tissue samples are stained with color dyes. There are several such examples across biomedical research.
The key point is that whenever color is involved and traditional methods—such as hue-based selection—are used to separate colors, theoretical color separation can play an important role.
3. I'm a bit concerned that even if the curve is adjusted accurately, it might remove the tint, though it may not be effective on practical images. For now, it’s perfectly fine to ignore the red and green tint. Please try your algorithm on the image below. From my observation, it might face challenges in areas with color gradation or where there is a transition in color. Kindly give it a try and let me know.
I am not familiar with QEEG and what imagery it produces, so if the raw data is colorful that certainly is different from the imaging processes I had in mind (based on »3.png«) and tinted slices are naturally a different issue yet again.
As for the photograph with the violet spot I don’t quite understand what you are talking about, but please don’t feel obligated to elaborate.
If I miss the benefit of your approach that’s not your problem.
P.S.: After more thought I suppose one can isolate a color range via the R-Channel after applying the HSB-Filter, technically without creating a Layer Mask but a Clipping Mask.
AdChoices