CMYK vs. RGB vs. LAB COLOR

Discussion in 'Digital Darkroom' started by david_debalko|1, Feb 16, 2004.

  1. Can anyone give me a brief description of the differences
    of these 3 color choices.
     
  2. That depends a lot on how well you need to understand the difference. They are three different ways to refer mathematically to colors, each a different color "space". Each system can theoretically refer to the same colors (roughly), but the subtle differences between them keep people jumping trying to match colors exactly between different spaces. CMYK (CyanMagentaYellowblacK) and RGB (RedGreenBlue) are more commonly known, because CMYK represents the possible output from a standard 4-color press and RGB represents the way computer monitors build color, but LAB color is the reference space in Photoshop for very good reasons (that I probably can't explain well or quickly). And there's a heck of a lot more to it than that.

    Check out books and articles by people like Bruce Fraser - one of the RealWorld Photoshop guys. Good luck.
     
  3. My line about LAB color could have been much better put in the above post. Suffice it to say that this is quite complicated to understand thoroughly.
     
  4. CMYK is Cyan, Magenta, Yellow and Black. This colour space is used in the world of printing presses where the four inks which go to make up all the other colours are... CMYK. All of your brochures, books, magazines, CD covers etc which are printed in colour are really printed with tiny dots of cyan, magenta, yellow and black ink dots in patterns which help simulate various colours. This colour space is pretty restricted but considering what's going on, its pretty incredible. Despite the fact that your inkjet printer at home has inks called cyan, magenta, yellow and black and probably a few others like light black or "photo magenta" your home printer does not print as a CMYK printer. Therefore, unless you are a professional graphic designer doing layout work, or a photographer supplying photos to a graphic designer who is doing layout work, you won't need to use this colour space.

    RGB is Red, Green and Blue and is a colour space where all colours are created out of Red Green and Blue ink or light. Your inkjet printer and your monitor are both RGB devices. Your printer obviously has other ink colours but the drivers are based on how red, green and blue work together to model all the other colours. sRGB and AdobeRGB and ProPhotoRGB are all models which are based on how Red Green and Blue create colour. They differ on how many colours they include and the number of shades involved. sRGB is the smallest and least desireable colour space from a photographers point of view and ProPhotoRGB is the largest space with the most number of inbetween shades. AdobeRGB lies somewhere in between. Choose too small a colour gamut and you clip colours in the original. Choose too large a colour gamut for a particular photo and you get banding betweeen colour gradations because there are missing steps from the model's point of view in your image between hues. With the advent of 16bit colour in Photoshop you'll be hearing more about ProPhotoRGB

    LAB colour is very different. I believe, but am not sure that this derived from scientific work in the 30s on vision and perception of colour and forms the foundation of colour management theory and in fact the current accepted theory of how our retinas perceive colour. In any case LAB colour considers colour along an L scale for lightness-darkness, an intersecting scale of "a" of red-green and another scale "b" of blue-yellow. This matches our cones which come in 3 "flavours", receptors which respond to darkness-lightness, others which respond to blue-yellow and yet others which respond to red-green.
     
  5. LAB colour is very different. I believe, but am not sure that this derived from scientific work in the 30s on vision and perception of colour and forms the foundation of colour management theory and in fact the current accepted theory of how our retinas perceive colour. In any case LAB colour considers colour along an L scale for lightness-darkness, an intersecting scale of "a" of red-green and another scale "b" of blue-yellow. This matches our cones which come in 3 "flavours", receptors which respond to darkness-lightness, others which respond to blue-yellow and yet others which respond to red-green.
    The last parts of the above is wrong. First to address the final part. The cones in the average human eye are sensitive to electromagnetic radiation and they correspond to Red, Green, and Blue color perception. These cones are actually sensitive in differing amounts to different frequencies of electromagnetic radiation. The specific frequencies the different types of cones are sentivive to overlap with the strongest sensitivity for each type of cone at a different frequency as is illustrated here . This is the first physical layer of human vision. The secondary layers and beyond that occur in the brain are assumed to transform this trichromatic (RGB) representation into other forms. To learn more about this I would suggest reading this to learn some keywords and basic concepts and progress from there.
    Whereas, the LAB color space was is an attempt based upon perceptual measurements of humans to create a perceptually uniform space. Roughly, this means that if two colors look to be different by about the same amount, then the numbers that describe those colors should differ in the same way. This has no simple correlation to cones in the human eye, but is instead statistically correlated to high level perception of color by humans. LAB is more correctly called L*A*B* or CIELAB.
    Beyond that, addressing the original question, LAB is an absolute reference space for color. Whereas both RGB and CMYK are relative spaces and can have various different definitions. This is where discussions of Adobe RGB (1998), sRGB , ColorMatch RGB, and etcetera stem from as they are all RGB spaces, but they are not all capable of rendering the same colors.
    It should also be noted that the CMYK space is of use to people who never do prepress because the addition of a 4th color channel (blacK) allows a skilled operator to generate color separations that allow easy access to a single feature in an image. This is because there is an extra degree of freedom in the selection of the CMYK representation of a color which allows for an additional level of control.
    Lastly, I will note that RGB and CMY are theoretically equivalent. But, in practice it often turn out that mixing Cyan, Magenta, and Yellow ink to get black often leaves a color cast so a pure black ink was useful. The usage of black ink also reduces the amount of ink that needs to be layed down on the page (1/3rd as much) which prevent overinking issues.
    This is a complex subject area and there are few who really understand it. Most of the "experts" are neophytes on the technical aspect, but highly skilled users, from what I have seen of their writing. Personally, I have limited my studies of it to awareness of what to research as I am more interested in creating images than I am interested in creating tools to create images.
    A webpage with some helpful graphics and accessible explanations.
    my $0.02,
    Sean
     
  6. Here's a fast overview. I'll do my best to limit the technical jabbering and concentrate on useful info.

    RGB is an additive color model. Red + Green = Yellow, etc. It is also closest to how our eyes see color. This makes RGB is an intuitive color model to work with. It's also how digital cameras and scanners (via their filters) see the world.

    CMY(K) is a subtractive color model. In other words, how inks or dyes absorb light. Cyan absorbs (subtracts) red, magenta subtracts green, etc. Combine all three and you (theoretically) get black. That works with the careful balance of photo paper, but not with ink. Adding a black (K) plate gives the advantages Sean mentions. It also allows editing work on the black plate alone - giving fine control over how shadow details are rendered.

    L*AB is a perceptual model of human vision. As Sean points out, it also precisely specifies an actual color. RGB and CMYK numbers are just that - numbers. It takes an associated color space to give them meaning. Without a color space all you have is RGB or CMYK mystery meat. A LAB value, by contrast, correlates to an exact spectral color.

    Working in LAB is counterintuitive at best. All the brightness information is in the L channel while color is encoded in the a and b channels. Minor changes to the latter make for huge changes in the appearance of the image. Some tricks, however, work very well in LAB. For example, sharpening is nothing more than highlighting contrast (brightness) differences. Ideally you do not want to create color casts in the process. Sharpening the L channel alone accomplishes this. Digital noise, on the other hand, is largely confined to the color channels. A gaussian blur applied to either the a or b channel (pick the one with the most noise) can eliminate much of the visible noise. Almost all the image detail is carried in the L channel, so this type of noise reduction does not degrade sharpness.
     
  7. Hi,


    Yes! Im another newbie (or should i say totally confused?) !to planet colourspace! I've been rading intently about colourspace and noticed that people are makeing a distinction between 2a COLOUR MODEL, which i presume they mean, RGB and CMYK"? and COLOURSPACE.

    1. What is the difference?
    2. And if there is a difference, Can you use a CMYK or RGB COLOURMODEL in either sRGB or Adobe 1998 COLOURSPACES? Sorry, I just got lost in this discussion.

    Thanks in advance.
     

Share This Page