Publishers of technology books, eBooks, and videos for creative people

Home > Articles

  • Print
  • + Share This
This chapter is from the book

How color management works

To understand the color management system (CMS), we must first learn the basic steps. This section defines terms and explains the basic CMS. For the sake of clarity and simplicity, I have left out details that will be covered later in this book, such as how to build a profile, choose a rendering intent, and maintain detail in saturated areas of an image. This information is a critical component of the set of common knowledge we must all possess to be truly successful collaborative partners.

Numbers = words

  • Humans use words within a language to express ideas.

    Digital image files use numbers within a color space to express colors.

In digital image files, colors are expressed by numbers, in the same way that in spoken language, ideas are expressed by words. Here's how the numbers work. (For the sake of simplicity, only RGB numbers will be described here. We'll get to CMYK in Chapter 5.)

RGB files are made up of three channels representing the three additive primary colors: red, green, and blue. A color in an RGB file is indicated numerically by the amounts of red, green, and blue that mix together to make that color. The scale of values in each channel is 0–255, in which 0 is the minimum value (black) and 255 is the maximum value (white).

In RGB files, maximum values of red, green, and blue add together to create white (R255/G255/B255), and minimum values of each add together to make black (R0/G0/B0). Colors are defined by various ratios of the three primaries. For example, a light orange might be indicated by R230/G190/B80. The three numbers together are called an RGB triplet.

Color spaces = languages

  • There are many languages in which to express ideas.

    There are many color spaces in which to express colors.

In the same way that a language contains a range of words that are used to convey ideas, a color space contains a range of colors that can be reproduced by specific output devices such as monitors and printers, as well as a range of colors defined by the numbers in a digital image file. Every digital image file, along with everything that renders the file visible, resides in its own unique color space. Color space describes the breadth and depth of color "spoken" by monitors, printers, projectors, and the digital image file itself.

We edit RGB files in, and output them to, particular color spaces. There are two general categories of color spaces: working spaces and output spaces. Color spaces that describe the range of colors our output devices are capable of reproducing are called output spaces. Output devices include monitors, inkjet printers, and offset presses. The monitor's output space is called the monitor space. Color spaces in which we edit RGB files in Photoshop (which actually describe synthetic, idealized output devices) are called working spaces. We explore color spaces in greater detail later in this chapter. For our current discussion, it is important only to know that there are many color spaces in which to express color, just as there are many languages in which to express ideas.

One of the factors distinguishing languages from one another is size of vocabulary. For example, English contains about 250,000 words; Spanish contains about 125,000 words. That difference has nothing to do with the expressive possibilities of the language, but may be a factor in how accurately an idea can be translated from one language to another.

In digital imaging, gamut refers to the number of colors a color space is capable of defining—in essence, its "vocabulary" of colors. However, while size of vocabulary is not the primary factor differentiating languages, gamut is the primary factor distinguishing one color space from another, and is the most important concern when translating color from one color space to another. We will review this issue in detail later in the chapter.

Profiles = identities

  • Determining which ideas specific words express requires knowing which language is being spoken.

    Determining which colors specific numbers describe requires knowing in which color space the digital image file resides.

In verbal communication, a word has meaning depending on the language in which it is spoken. For example, consider the word "pain." In English, it means "hurt." In French, it means "bread." Same word, but two completely different meanings depending on which language it's in.

So, if you see the word "pain," which is it? "Bread" or "hurt"? You can only determine the meaning of "pain" if you know in which language it resides.

Likewise, knowing in which color space a digital file resides is critical to knowing what colors it is expressing. ICC profiles identify in which color space the file resides, or in which color space an output device resides.

An ICC profile (commonly referred to simply as profile) is a small text file that contains the identity of the color space as well as a "dictionary" describing the gamut (number and definition of available colors) of that specific color space. Digital image files have the profile embedded in them. Output spaces (like printers, offset presses, etc.) are also described by profiles, but the actual profiles reside on the user's computer, not inside the device itself.

Same RGB triplet, different colors

Just like our bread/hurt example, in digital imaging the specific color an RGB triplet defines is dependent on the color space in which the file is being edited or the color space to which it's being output. For example, the RGB triplet noted earlier—R230/G190/B80—will yield three different shades of light orange in three different color spaces. (See the "Why doesn't a particular number describe a particular color?" sidebar.)

Same RGB triplet of numbers, but resulting in three different colors depending on which color space it's in.

Isn't this a Problem?

Yes. Big problem.

Each RGB color space uniquely expresses a given RGB triplet. Our light orange example—R230/G190/B80—will be a very bright orange in the ProPhoto RGB working space; a medium orange in the Adobe RGB (1998) working space; and a brownish orange in the sRGB working space. And, sent to an inkjet printer or an offset press, it will vary even more widely in how it is output (see Figure 4.2).

Figure 4.2

Figure 4.2 The same RGB triplet will express two different colors in two different color spaces. Here, we see a comparison of the same RGB triplet in ProPhoto RGB and Adobe RGB (1998).

As we shepherd a file through the RGB-to-CMYK workflow, it will move through multiple color spaces. The photographer will edit the image file in Photoshop in one color space (his working space), view it on his monitor in another color space (his monitor space), proof it on his inkjet printer in yet a third color space (the printer's output space), and then hand it off to the designer, who will run it through several color spaces as well. Eventually, the file will end up at the offset printing plant where it will take final form as a press sheet in the offset press's color space (the final output space).

Because all of these color spaces each speaks a unique "language," the set of RGB numbers in the image file will yield different colors in each of the different color spaces along the way. This does not sound like the way to achieve "predictable color."

It's like broadcasting a message in one language to listeners who each speak a different language and hoping that the message will be clear to each of them. Or, like that backyard game of "Telephone" kids play, where the more times the message is passed along, the more unrecognizable it becomes.

We need a way to translate color from one color space to another without losing the accuracy of the color. In other words, "How can I get my printer to match my monitor?"

Universal color space = universal set of ideas

  • Unambiguous communication between two languages requires a translation that references the universal set of all ideas/concepts known to man.

    Unambiguous communication between two color spaces requires a translation that references the universal set of all colors visible to the human eye.

Accurately translating between two languages requires a dictionary of definitions for each language and a knowledge of the idea being translated. It is the idea that is paramount; the words used to convey that idea are simply the means to the end.

Accurately translating between two color spaces is a very similar process. It is the color that is paramount; the RGB triplet numbers are simply the means to the end. In color management, translating is called converting.

In language, dictionaries define words in reference to universally understood ideas and concepts. For example, languages have words to express concepts such as "sky," "fast," and "happiness." They are different words in different languages, but they have the same meaning.

In digital imaging, profiles define colors in specific RGB color spaces in reference to a universally accepted understanding of absolute color. That universal understanding is contained in a unique color space called CIELab (or Lab, for short). Unlike RGB color spaces, which define color based on the specific behavior of output devices like monitors and printers, Lab defines color based on normal human vision.

The set of color definitions in Lab is not dependent on a particular device, but is absolute and universal. It is as though there existed one universal dictionary for all concepts in all languages. As such, Lab becomes the hub for all translation, or conversion, from one color space to another. Lab is our absolute and unambiguous reference space. In color management parlance, it is referred to as the profile connection space (PCS).

Although many words can be exactly and directly translated from one language to another, there are others for which there is no direct corollary and result in a more ambiguous translation. Likewise, while most colors can be directly converted from one color space to another, there are others that cannot, resulting in colors that do not absolutely match each other in two different color spaces. For example, highly saturated blues that exist in a large RGB color space like ProPhoto RGB do not exist in a much smaller working space like sRGB. Converting that blue from ProPhoto RGB to sRGB will not result in an exact color match. I'll address this issue of accuracy later in the chapter when we discuss gamut.

Conversion: the basic mechanics of color management

A digital file will be expressed in many color spaces along its way in our RGB-to-CMYK workflow from camera to offset press. As we discussed earlier, each partner will view the image on a monitor in the monitor's color space, edit it in Photoshop in a working space, and proof it on an inkjet printer in the printer's output space before it reaches its ultimate destination on a CMYK press sheet in the offset press's output color space.

To maintain accurate, predictable color, a conversion from one color space to the next must take place each step of the way. The basic mechanics of color management lie in that conversion, and the process is the same every time:

  1. The user tells the color management system (CMS) where the image is coming from (the source space) and where it's going to (the destination space).
  2. Using the profile of the source space, the CMS "translates" the RGB triplets in the source file to universal Lab values.
  3. The CMS then translates those Lab values into new RGB triplets that "speak" the "language" of the destination space using the destination profile.
  4. The CMS then converts the file to the new RGB triplets, resulting in a file that will yield the same color as the source file—different set of numbers, same color.

To make this a bit more concrete, here are two examples of the basic Source-to-Lab-to-Destination color management move in action: printing a digital file to an inkjet printer, and viewing a digital file on a monitor. In each case, a source space and a destination space are identified by the user, then the color management system performs the conversion behind the scenes using the Lab reference space.

Example #1:

Let's say you want to make a print of our "light orange" color. The digital image file is in the Adobe RGB (1998) working space and you'll print that to an Epson 3800 printer onto Epson Proofing Semimatte paper. To achieve accurate and predictable color, you will convert from the source space (Adobe RGB [1998]) to the destination space (Pro38 PPSmC). Each of these color spaces is described by a profile. Here's how this conversion works:

  1. The source is a file you're editing in Photoshop in the Adobe RGB (1998) working space. The color you use has an RGB triplet of R230/G190/B80.
  2. You use the Print command and identify the destination by telling Photoshop that you are printing to the Proofing Semimatte paper (PPSmC).
  3. The CMS reads the source profile (Adobe RGB [1998]) and determines that the RGB triplet of R230/G190/B80 refers to an absolute Lab color of 81L/11a/66b.
  4. The CMS then reads the destination profile (Pro38 PPSmC) and determines that the absolute Lab color of 81L/11a/66b can be produced by the printer using an RGB triplet of R242/G202/B125.
  5. The new converted RGB triplet of R242/G202/B125 is sent to the printer, and an accurate "light orange" matching your expectations is printed—different set of numbers, same color (see Figure 4.3).
    Figure 4.3

    Figure 4.3 In this example, a conversion takes place from Adobe RGB (1998) (source space) to Pro38 PPSmC (destination space).

Example #2:

Let's consider how you accurately view that "light orange" color on your monitor. Again, a conversion needs to take place between source and destination. In this case, the source is the Adobe RGB (1998) file you are editing in Photoshop, and the destination is your monitor color space as defined by your monitor profile. Here's how this conversion works:

  1. The source is a file you're editing in Photoshop in the Adobe RGB (1998) working space. The color you use has an RGB triplet of R230/G190/B80.
  2. In your computer operating system, you have identified your monitor profile.
  3. The CMS reads the source profile (Adobe RGB [1998]) and determines that the RGB triplet of R230/G190/B80 refers to an absolute Lab color of 81L/11a/66b.
  4. By reading the destination profile (monitor profile), the CMS determines that the absolute Lab color of 81L/11a/66b can be produced by the monitor using an RGB triplet of R228/G192/B65.
  5. The new, converted RGB triplet of R228/G192/B65 is sent to the monitor, allowing the monitor to accurately and predictably display the "light orange" color. Again, different set of numbers, same color (see Figure 4.4).
    Figure 4.4

    Figure 4.4 Every conversion from one color space to another takes the Source-to-Lab-to-Destination path. Lab is used as the reference color space through which RGB triplets are translated in order to convey accurate, predictable color. A successful conversion results in different numbers but the same color.

Basic Move, No Judgment Required

Both of these examples demonstrate the basic mechanics of the Source-to-Lab-to-Destination process that takes place in every conversion. There is no judgment required to perform this maneuver. Judgment comes into play when determining how to do a conversion in order to deal with the issue of gamut.

  • + Share This
  • 🔖 Save To Your Account