18

How can I convert an RGB color into a real life color?

If I mix equal amounts of red, green and blue paint together in real life, I will never achieve a white color, in the way you do when setting the maximum values for them in RGB (#FFF), and I'd be hard pressed to get an exact shade of grey.

Looking at an RGB code, how can I know what proportions to mix the colors in to get the same with real paint?

Is RGB to CMYK conversion the answer?

To put it another way:

How can I mix real paint colors together to achieve the same color that I've got on my screen in RGB?

  • 11
    Yup, CMYK - and step 1 to understanding why is understanding the difference between additive and subtractive colours – user56reinstatemonica8 Sep 10 '14 at 15:59
  • Fun example of someone converting digital colours to something physical (Eric Daigh's push pin portraits) – user56reinstatemonica8 Sep 10 '14 at 16:22
  • 10
    I wasn't aware that RGB colors aren't real. – DA01 Sep 10 '14 at 16:34
  • 1
    @DA01 im pretty sure my reality differs from yours, but that may be because im a bit... Hence how can you know what somebody considers real. Unfortunately since i must deal with the abstraction called language this is hard to explain. – joojaa Sep 10 '14 at 17:25
  • 1
    If you want to look how a non-RGB color (anything that can't be perfectly represented in RGB) looks like, then you'll have to get it printed out instead of looking at it through an RGB monitor. – Peteris Sep 10 '14 at 20:48
7

Real color is a quite complex subject. In essence color is something that happens somewhere between your cornea and brain. Simplifying this a bit color is what you sense when some photons interact with the three color sensing structures in your eye. Other definitions exist but they fail on many levels.

Photons can reach your eye trough many processes. They can reach your eye directly from the light source, or they can bounce on the way. When they bounce your actually seeing the color that the surface did not absorb. Whereas when you emit light you see what was sent unless something obscures the light on the way. The photon can also reach your eye trough other processes, some which are weird and obscure but that's another subject.

Because we have 2 primary systems for imaging in design, we also have 2 base models: additive and subtractive. One deals with emissive media like monitors. The other with reflective media like paper, but also works for things like stained glass. These models are inverses of each other: RGB is the inverse of CMY; mixing RG gives you Y and so forth.*

Manually mixing color has some other challenges. Your pigments may not react with each other chemically, or anything can happen. The pigments also need to be pure, and not have too big of a spread in what they remove from the spectrum as the eyes sensors overlap slightly. I find it extremely unlikely that you will find paint that fulfills this requirement easily.

It's also important that your lightning conditions need to be good as it affects how the subtraction happens. It's possible to design colors that change when you go indoors under artificial light, as compared to direct sunlight (This is called metamerism).

Now comes the mindblowingly hard part. In order to mix the right color you need to measure the colors. See, RGB values themselves are just signals; they do not represent colors without you knowing how the device displays the color. Each monitor behaves differently, and so do your pigments.

* Interesting fact: The ordering, RGB and CMY, are not mistakes. Any of the two colors in a triplet combined makes the corresponding color you didn't choose in the corresponding inverse triplet, so G + B = C and C - Y = G. The fact that the colors don't display as pure is because the computer tries to mimic available pigments, failing miserably.

  • I believe additive (RGB), substractive (CMY[K]) and solid (pigment)paint colors are three(!) different worlds. Example: Yellow and Blue gets you green in painterly colors. Additive does not apply (obvious), but substractive print colors (aka CMYK) are not the same either. Neither is paintbox blue anywhere near cyan, nor is the result convingly the same... Similar, magenta is certainly not "red". How come, in a paintbox, red and blue (as in RGB) are primary colors, and yellow (which is a mix color in rgb, a primary color in cmyk). Quite a crude mix... (never got the pigment part myself) – Frank Nocke Sep 11 '14 at 13:40
  • @Fronker cyan is blue just a more pure pigment shade magenta is red of very pure. The pigment coloring is a subtractive on basis that it reflects colors, thats physically the same mechanism. Base color purities may vary and you could have more base colors but its sill subtractive. What matters is how you excite the eye. the purer base pigments you have the brighter colors you can mix. colors dont need to be magenta cyan and yellow. – joojaa Sep 11 '14 at 15:30
  • to add cmy are chosen on basis of simplicity of conversion not universality. in fact if you add green and orange pigment you get a bigger gamut. But this is more of a issue of not having suitable pigments. – joojaa Sep 11 '14 at 15:51
11

Converting to CMYK won't help you unless you find CMYK specific paint, which I'm not sure exists. Commerical paint manufacturers use a variety of non-standardized ways of expressing colour - so try this web based converter:

http://www.easyrgb.com

This will convert your RGB to a paint colour. That's the free option.

If you want to be really picky, you'll have to buy a pantone book:

http://www.pantone.com/pages/pantone/index.aspx

Last time I bought one it was about $200. Pantone is the industry standard for matching colour in printing. If you're trying to match a colour in a logo, ask for the brand guide as it will likely list the Pantone value for the colours in the logo. Every printer owns a Pantone book, and picky printers and designers replace them after a few years when they start to fade and discolour. This is how the design industry references specific colours. And your software should be able to convert an RGB colour to a Pantone colour. However, paint stores don't use this system of colour matching - so you'll still have to take your pantone book with you to visually compare to their paint chips, or to be scanned by their software.

  • 2
    RGB to Pantone color conversion is not going to be 100% accurate as the software will only give you the nearest approximation to a Pantone color. – Joshua34 Sep 10 '14 at 22:27
7

RGB is an additive color model using light directly from its source before it is reflected off of an object. In essence, you start in darkness and because you are directly viewing the light source, the wavelengths can be added to each other to create colors.

CMYK and 'real life colors' both use a subtractive method to display color. In essence, you begin with light and the color wavelengths which are not absorbed by a pigment, object, etc. are then reflected into your vision as different colors.

Yes CMYK conversion will make this much more achievable but understand that your monitor's RGB color profile will display one color, while CMYK colors printed (or mixed!) can look very different depending on your lighting, substrate used, and pigments, among other things.

  • Both additive and subtractive models exist and can be seen every day in real life, not just subtractive as you claim. Additive means you add light, and the more you add the "brighter" the color becomes, with white being the brightest you can get. With subtractive, you're adding paint, which prevents light from reflecting, and the more paint you add, the less light will be reflected, and the "darker" the color will result. If one wants to simulate paint mixing, a subtractive method is more appropriate, but if you want to simulate light mixing, an additive method is more appropriate. – Panda Pajama Sep 11 '14 at 2:45
  • 'real life' was used in the same context of the OP, I suggest you re-read his OP. It seems obvious he meant something tangible, i.e. paint, rather than recreating a RGB light source. BTW why didn't you just post as your own answer rather than as comment to mine? – Joshua34 Sep 11 '14 at 2:50
  • @PandaPajama "Both additive and subtractive models exist and can be seen every day in real life, not just subtractive as you claim" Did I really claim that? All I claimed was that RGB used the additive model, I never claimed it was the only user of the model nor that the additive model couldn't be found in real life. You are free to edit or delete your comment. – Joshua34 Sep 11 '14 at 2:56
  • That is much more clear now after your revision to your answer. – Panda Pajama Sep 11 '14 at 3:00
4

if I take equal amounts of red, green and blue paint in real life I will never obtain the white color or some shade of grey

That does not seem true. You will actually get a dark shade of grey such as this one in the center (from the subtractive color model), also known as the imperfect black in CMYK:

subtractive colour

As for your main question, i.e. “how do I know what will it look like”, a precise answer will require you to know more about the light, about the colour of your red, green and blue pigments and how exactly they mix. If you want to actually experience the colour, for instance see it on a screen, you will need a properly colour-calibrated pipeline (I suggest Abhay Sharma’s Understanding Color Management for an accessible book on the subject) and any piece of software that understands colourspaces.

However, if you only need an approximate answer, here is a little tip: the hue of the final colour will be the same as the one you see on screen (because for hue calculation it does not matter whether you’re doing additive or substractive colour mixing); if your colour is #ff7f00 or [255,127,0], which is orange on an RGB screen, then you know the final colour hue will be orange (albeit quite a dark one).

3

as an addition to what people are saying about light. Red, blue and green light do infact make white. the rainbow (R,O,Y,G,B,I,V) is infact a white light that has been separated via refraction and divided into the R,G,B colors. the other colors are the overlapping of two light colours

3

I think a good caveat to all these answers is to keep in mind that color management in the digital world is a "close enough" science. When you calibrate and profile a monitor, you're not getting an accurate plot of all 16+ million 24-bit RGB colors. You're getting an accurate plot of a very minor subset of those colors, with every other color between them interpolated to a 'best guess.' Even if you're working in 48-bit color, the monitor is still only providing 'best-guess' color, and in fact is having to guess even more due to the finer color depth.

So you'll run into inaccurate color representation when comparing specific 'real world' colors to the equivalent monitor colors all the time. Many times the inaccuracy will be very minor, but at times, it will be surprisingly inaccurate.

Factor into this the facts that humans don't all see color in the same way, and that each eye will have slight color-sensitivity variations, and we're talking about a very inaccurate science that has a lot of room to grow.

  • Well the cells themselves have a quite uniform response, the problem is in procrssing and some people missing cells. And horror some women have a fourth color pair making the color wheel obsolete. Some humans are also missing their natutal uv filer so their eyes see into uv ranges. But color correction is insanely hard yes. Bit depth is not the problem possible gamut is the limiting factor. – joojaa Sep 11 '14 at 15:42
  • The hardware and software are the real problems when it comes to digital color. Gamuts are just containers of color. Yes, almost all gamuts will lack certain colors that fall outside of them. But some of the colors they do contain will be inaccurately displayed as long as graphics card lookup tables and profiling software don't plot all RGB values. Even if that level of granularity were achieved (which would be an insanely long process, BTW), the physics involved in converting digital color information to actual light rays emitting from the monitor will continue to be a limiting factor. – digijim Sep 11 '14 at 16:02
1

There is an online service which calculates artist paint mix recipes from rgb values here: http://sensuallogic.com/paintmaker/OnlinePaintMixer.html

It works both with oil and acrylic paints.

  • And the brands they cover is:
  • Golden Heavy Body acrylics
  • Liquitex Heavy Body acrylics
  • Rembrandt Fine Artist Oil colours
  • Gamblin Conservation colours

We're looking for long answers that provide some explanation and context. Don't just give a one-line answer; explain why your answer is right, ideally with citations. Answers that don't include explanations may be removed.

  • 2
    Welcome to Graphic Design SE. While your answer seems to address the question, you also seem to be behind the website you are linking. Please disclose your affiliation. See How to not be a spammer. – Wrzlprmft May 24 '15 at 9:19
0

RGB, CMYK, HEX, values don't describe really a color. They name something not so far around.

Why? They forget the background, on what the "real" paint or ink pigment is applied.

Only CIE Lab* Values describe the real color as they appear under certain conditions (Observer 2° or 10°, Light D50 / D65 / ..., measured with or without polarisation filter, Light with UV content or without).

So, use Lab Values with measurement condition for really fix a color, that is truely independent of all other influence.

From this you can go back to cmyk, rgb or other. But the good quality measurement of Lab-Values is the base. Nothing else counts.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.