The Photography & Filmmaking Education Resource
0 Me Search
1 914-347-3300

Learning Center

Color Management Concepts


Browse Library by

BY Phil Nelson September 10, 2014 · Published by Amherst Media

To control color in your images, you need to know how devices handle it. Phil Nelson goes in-depth on color spaces, gamuts, and profiles in this excerpt from his Amherst Media book The Photographer's Guide to Color Management.

This excerpt from The Photographer's Guide to Color Management is provided courtesy of Amherst Media. To purchase the book and learn more about the publisher, visit the Amherst Media website.





If you have wrestled with controlling color in your images and have attempted to try to color manage your photographic workflow, you will probably agree that color management is pretty complicated. What goes on behind the scenes can seem overwhelming. While it is possible to color manage your images without understanding color theory and all the technical aspects of color conversions, it can be very helpful if you know some of the fundamental concepts, especially when it comes to setting up a color-managed workflow, troubleshooting problems, or trying to understand how a new application handles color. This chapter is intended to introduce you to some of the basics.

Figure 11—An arbitrary synthetic color space using the HSL Color Model. The color space is plotted in three dimensions where Hue rotates around the vertical access, Saturation increases from the center core outward, and Lightness values increase and decrease up and down the vertical axis. With values assigned for H, S, and L, a point can be plotted within the sphere that defines a color. This three-component model is typical of the different color models that are used to describe color space.

RGB and CMYK Color Models

A color model defines color values numerically, most often by grouping three or four primary components. HSL, HSV, LCH, CMY, and RGB are all color models that, when associated with instructions for how the three components should be interpreted, will create a color space (figure 11).

RGB. RGB is a basic color model associated with numerous devices that either emit or capture light, like scanners, digital cameras, and computer displays. RGB is referred to as an additive model since the more light that is emitted or captured, the higher the RGB values, and the closer the color is to white. On a computer, in a photo editing application like Adobe Photoshop, we can define colors by assigning individual values of R (red), G (green), and B (blue), typically on a scale from 0 to 255. When assigned in Photoshop’s Color Picker, the maximum values (R 255, G 255, and B 255) will produce white (figure 12). The minimum values (R 0, G 0, B 0) represent no light at all—or black.

Figure 12—RGB is an additive color model. When all three color channels are assigned maximum values, the result is white.

CMYK. The CMYK color model, typically used by printing devices, is a subtractive process and involves applying ink to a medium. As the individual channels of C (cyan), M (magenta), and Y (yellow) inks are combined on paper, different color values can be achieved depending on how the inks are mixed. This is done using a scale from 0% (or no ink) to 100% (or total ink) for each channel. The less ink that is printed, the closer the resulting color will resemble the color of the medium, which is usually interpreted as white. When combining maximum amounts of ink on paper for each CMY channel, the less light can be reflected off the medium, and the closer the printed color will resemble black (figure 13). (However, applying maximum values of CMY does not achieve a dense black; it produces a muddy brown. This is one of the reasons printers added the additional K [black] ink channel.)

Figure 13—CMYK is a subtractive color model. When all three color channels are assigned maximum values, the result is black.

Applications like Photoshop allow you to choose from a number of different color models. Because digital photographic images are in RGB mode at inception, you may spend most of your image-editing and optimization time in RGB. Because of the nature of the offset-printing process, at some point in the workflow, your images will have to be converted to CMYK. This may be a color conversion that you will have to manage yourself—and one that can be problematic if not handled properly. Understanding when and how to make this conversion is a critical factor in the photographic process and can lead to successfully and consistently reproducing images that meet your color expectations.

Color Space and Color Gamut

The term “color gamut” defines the limits of a color space. The sphere in figure 11 (page 18) is an arbitrary color space that is illustrated in three dimensions. Other color spaces describe the abilities of a specific device—a digital camera, computer display, printer, etc.—to perceive, display, or print colors (for more on this, see the following section on device-dependent color).

If a color falls outside the color space of the device, it is called “out of gamut.” In simpler terms, the device is not capable of perceiving, displaying, or printing the color. Many laser printers, for example, have a relatively small color gamut, especially when compared to the larger color gamuts of many inkjet printers on the market today. There are certain colors that the inkjet printer can print that the laser printer cannot. By knowing the color gamut of a device, it is possible to understand its color capabilities or color limitations.

Figure 15 illustrates two device profiles in comparison with one another. The larger red shape represents the color space of a Microtek ScanMaker 1000XL flatbed scanner. The smaller orange area is the color space of an Epson Stylus Pro 7600 inkjet printer when printing on Epson Premium Luster Photo Paper. Both profiles are mapped on a three-dimensional grid based on the LAB color model. It is apparent from this illustration that the scanner has a larger color space than the printer. This means that the scanner can capture far more color than the printer can print. If we scanned an image that contained color outside the color space of the printer, a color conversion conducted when outputting the scanned file would map out-of-gamut colors to the closest “in gamut” color of the printer, rendering the most accurate representation of the color that the printer can print.

Figure 14—Adobe Photoshop’s Color Picker provides a means to assign RGB and CMYK values to define a color. Notice that equal values of RGB will create a neutral gray while the CMYK values for the same gray tone are not equal.

Device-Dependent Color Space

Color spaces that describe the capabilities of a particular device are commonly called “device dependent.” They describe the device’s color limits in terms of a color gamut. As illustrated in figure 15, every device has unique characteristics.

Figure 15—Two device-dependent color spaces illustrate the capabilities and limitations of different devices.

Different Devices Define the Same Color Differently.

Because different devices have their own color spaces, they define the same color differently. In figure 15, the values for a particular blue color (R 42, G 82, B 171) are defined by the scanner’s unique color space. Because a computer display’s color space is different from the scanner’s, however, it interprets this same blue color with RGB values of R 65, G 117, B 192. To reproduce this same color on an offset press would mean converting the RGB values to C 84, M 45, Y 1, K 0—which, again, is defined by the color space of the press.

Since every device interprets the same color in its own relative fashion, the values R 42, G 82, B 171 in the scanner’s color space would look different than R 42, G 82, B 171 in the display’s color space. Even though the two devices are in the same color mode, RGB, the way their RGB values are interpreted is relative to the device their color model represents. It’s as though the two devices are speaking different languages. They are both saying the same thing, but the words they are using are relative to the language they are speaking and are different from one another.

Figure 16—Each device in the workflow interprets the same blue color with a different set of RGB or CMYK values.

Figure 17—Conversely, they would all interpret the same RGB or CMYK values as a different color.

Device-Independent Color Space

Wouldn’t it be great if all the devices that we work with spoke the same color language? Unfortunately, that is not the case. Because the photographic workflow requires that we move images between numerous devices and applications, we need a way to translate the image’s color from one device’s color space to the next. In order to make the translation, there needs to be an absolute reference that can be used to do the translating. Such a reference exists in the form of a device-independent color space.

The Importance of Device-Independent Color. In the early 1920s, the International Commission on Illumination (Commission Internationale d’Eclairage, or CIE), conducted a study on the perception of color and developed the color model CIEXYZ. Referred to as XYZ, this is a three-dimensional model that includes all the colors that human beings are able to perceive. Because it does not describe the color space of a specific device, it is device independent.

Additionally, because this color model is so large, all the color spaces of the devices that photographers are working with will fit inside XYZ. (Unless you are working with infrared or some color outside of human perception—but it stands to reason that the only color that you are interested in is the color you can see.) Imagine mapping all the colors that you can see into three dimensions using X, Y, and Z axes, and that anywhere in this three-dimensional space, you can plot a point with three specific XYZ coordinates. That point would be the absolute location of a unique color in the XYZ color space. There is no other way to describe this color. XYZ defines all of its colors with unique X, Y, and Z values.

Having an absolute color model makes it possible to translate color from one device-dependent color space to another. This is done by creating an ICC profile for each device to be used in the workflow. These color profiles (described in chapter 5) provide a point of comparison between a color from an absolute, device-independent color space and the color that the device interprets it to be. If there is a difference—and usually there is—it is recorded in the color profile. The profile, then, is a record of how the device handles absolute color values from a device-independent color space and is used in the color-management process as a dictionary of sorts to translate colors
between different device color spaces. The profile is a color characterization of the device that maps the device’s color space and defines the device’s color capabilities.

Imagine the process this way. When making a profile for a scanner, you make a scan of a lot of colors that have absolute XYZ values. Your profilebuilding software knows what these values are. When the scan is complete, you notice that the scanner’s interpretation of a red color is orange. This difference is recorded by the profiling software into the profile. Later, you scan an image of some strawberries. The scanner, as you discovered earlier, makes all the red in the scanned image orange. After scanning you assign the profile you made. The profile knows that all the orange values in the scan are really supposed to be red, and tells the color-management system to make the necessary changes and adjust the colors accordingly.

LAB Color. In the early 1970s, the CIE introduced the color model L*a*b* (or LAB), which is based directly on XYZ and more accurately defines human perceptible color. Like XYZ, LAB is used as the reference model for many profile-making software applications. Unlike XYZ, LAB maps its color using coordinates for luminance (L*) and two color parameters, blue to yellow (a*) and green to magenta (b*).

Figure 18 illustrates the colors from the Macbeth ColorChecker mapped in a LAB color model in three dimensions. The L* parameter, mapped as the vertical axis, describes the luminance of color—its relative brightness. The a* and b* parameters, mapped horizontally (front to back and left to right) describe color. L* 0 defines black; L* 100 defines white. Negative a* values yield green; positive ones yield magenta. Negative b* values yield blue; positive ones yield yellow. Using this coordinate system, it is possible to plot any color that the human eye can perceive as absolute LAB values.

Figure 18—The 24 color Macbeth ColorChecker mapped in LAB.The colored circles plotted in this three-dimensional diagram illustrate how color is mapped in the LAB color model. The L* channel is represented by the vertical axis.The a* channel runs from front to back, and the b* channel extends from left to right. The higher a color circle is positioned on the diagram, the lighter its color. Circles positioned to the left are more blue, and to the right they are more yellow. Colors that are positioned to the front are more magenta. Colors toward the back are more green. Each of the circles has a unique LAB coordinate as do all colors in the LAB color model. Several of them are called out in the diagram. On the L* axis a value of 0 equals black and a value of 100 equals white. Any value in between is a value of gray. Negative values on the a* and b* axes move the color toward green and blue respectively, while positive values move the color toward magenta and yellow. The colored circle at the top of the L* axis has a value of L* 97, b* 0, b* 2. Because the a* and b* values are so close to 0, the color represented by these LAB coordinates is very neutral. The yellow sphere to the right with LAB values L* 81, a* 4, b* 79 sits high on the L* axis and close to the a* axis, but far to the positive side of the b* axis, making it a bright and saturated yellow.

Figure 19—Purplish blue defined in LAB values.This is an absolute definition of this unique color.

Profile Connection Spaces. Figure 20 illustrates a common color transformation from one color space to another. In this example, a scanner scans an image in RGB. The color management system uses the scanner’s ICC profile to map scanner RGB values to LAB. These LAB values from the scanner’s source profile are then passed to the inkjet printer’s destination profile where they are mapped to the printer’s CMYK values. This point of interaction between the scanner source profile and the printer destination profile is called the profile connection space (PCS). XYZ and LAB are commonly used as profile connection spaces. In order for an accurate conversion to take place there must always be a source profile and a destination profile. The result of this conversion should be a print that accurately represents the colors that were originally scanned.

ICC Color Profiles
For a well-managed color workflow to work properly, it is necessary to have a profile for every device in the workflow. To understand why, let’s return to our analogy, looking at an ICC profile as a translation dictionary. Imagine you are trying to speak to a person from Kenya. You speak Russian and the Kenyan speaks Swahili. You have a Russian-to-German dictionary and the Kenyan has a German-to-Swahili dictionary. With these two dictionaries it would be possible to communicate—you could translate the Russian into German and the German into Swahili.

Figure 20—Converting scanner color to inkjet color. When the conversion takes place, the scanner’s source profile maps its relative RGB values to absolute LAB values.The LAB values are passed to the printer’s destination profile where they are mapped to relative CMYK values for output.

In this example, the dictionaries are profiles, and German is the profile connection space. Now, imagine that a third person enters your conversation—but this person speaks French. In order to communicate with this person, he just needs to have a French-to-German dictionary. In fact, you could continue to have a conversation with anyone who enters the group as long as the person has a dictionary that uses the profile connection space.

The most accurate profiles are the ones that are custom made for individual devices. It is possible to acquire profiles from the manufacturers of input, display, and output devices. Many manufacturers will install ICC color profiles onto your computer when you install their device drivers or applications. While these profiles can be very good, they are built for an entire product line, like all HP Designjet 130s, as opposed to the individual Designjet 130 that you might own. In chapter 5, you will learn how to create custom ICC profiles for all the devices in your workflow.

The International Color Consortium The ICC has been mentioned a lot so far. What is this? The International Color Consortium, or ICC, is made up of a group of like-minded companies that came together to create a standard way of solving “the color problem” when sharing images and files between different hardware devices and computer applications. Founded in 1993 by eight technology companies, the ICC now has an impressive member list that includes all the usual suspects—Apple, Adobe, Sun, Kodak, Agfa, Canon, Quark, GretagMacbeth and X-Rite, R.R. Donnelly, Pantone, Nikon, Nokia, and about sixty more.The ICC developed the methodologies most widely used for color management and promotes their adoption as an open, cross-platform, and vendor-neutral standard. If you look at the device drivers and applications that you are using for photography, you may find that many of them subscribe to the industry standards of the ICC. Unfortunately, each color-management implementation is not always exactly the same when you compare one manufacturer’s device driver or application to another. This has created some confusion with end users. However, understanding the ICC’s color-conversion methods and the process of color management can go a long way in helping you figure out what a vendor is attempting, if it is not initially clear.

Working Spaces

In addition to device profiles, photographers should be familiar with the concept of a working space. Introduced by Adobe Systems, working spaces—like Adobe RGB (1998), ProPhoto RGB, or ColorMatch RGB—are ICC color profiles that do not represent the color capabilities of any given device. These profiles provide a uniform, neutral environment that is well suited for image editing. Usually they have a color gamut that is large enough to give the photographer ample headroom for manipulating color and the flexibility to convert, on output, to numerous different output destinations—like inkjet prints, CMYK color spaces for press output, and the Internet. Working spaces are a great place to hold your color until you are ready to output, at which time you convert to the output color space of the device you are printing to. If you archive your edited images in a standard working space, then you will increase the chances of being able to successfully output to all different types of devices, even those that might not exist today.

Each working space profile has its own unique color gamut. ProPhoto-RGB has an extremely large color space, while sRGB is relatively small. Typically what they have in common is “neutrality,” meaning that they translate equal amounts of RGB as a gray value. This is often also referred to as being “well behaved.” Assigning equal amounts of RGB may not result in a neutral value if you are editing in a device’s color space. Device color spaces are not usually well behaved.

For ideal results, a working space should meet the needs of the photographer’s workflow. Because not all workflows are equal, there are many working spaces available. Some, like Adobe RGB, are widely used. Others can be more obscure—like Don Hutchison’s Don RGB. When choosing a working space for your images, you will want to consider the following:

  1. What is the color space of the output device that you will be sending your image to?
  2. Are you outputting to multiple output devices and mediums, like inkjet, offset, a sign printer, the web?
  3. Are you archiving your images with the expectation that you might be outputting to some unknown device in the distant future?

RGB working spaces have a relatively large color gamut when compared to the color gamut of many printing devices. For example, Adobe RGB completely encompasses the CMYK color space based on the SWOP specification for offset printing. Ideally, a working space should contain all the color from the input device and encompass the color space of all the output destinations you may be sending your image to. This is a pretty tall order considering that many input devices today can capture a very large amount of color data, and the numerous output devices that you might send your image to could vary dramatically. 

So why not just use the largest working space available, like ProPhoto-RGB? Simply put, in a color space as large as ProPhoto RGB or Wide Gamut RGB, you may be able to assign highly saturated color that current printing technology could never reproduce. When converting for output, the color-management system will do its best to map the saturated ProPhoto RGB color to the closest in-gamut color that the printer can reproduce. These colors could be very far apart, and the end result could be a very drastic shift in color that is far from meeting your expectations.

Figure 21 (below) shows an an extreme example, comparing ProPhoto RGB to the CMYK color space U.S. Sheetfed Coated. As you can see, a color assigned at the periphery of ProPhoto RGB would have to undergo a significant transformation to be output in this CMYK color space. The transformed color would be quite different from the original. Additionally, editing an 8-bit per pixel image with only 256 tonal levels in a large working space like ProPhoto RGB, where the colors are spread out over a wide space, can easily introduce posterization into an image. This creates visible bands, especially in areas of subtle gradation. It is better to edit 8-bit images in smaller working spaces where the color is closer together and where the potential of introducing banding will be less of an issue.

Figure 21 (left)—This chromaticity diagram illustrates the extreme difference between two color gamuts. The larger, ProPhoto RGB, is so large that it extends outside of what the human eye can perceive. Converting from this large space to a much smaller one, like U.S. Sheetfed Coated, could drastically alter the highly saturated colors that reside at the periphery of ProPhoto RGB as they are transformed to the closest match inside the printer’s color space. Simply put, there are a lot of colors in ProPhoto RGB that the printer simply cannot print. Figure 22 (right)—Putting image color in a working space of a more moderate size, like Adobe RGB (1998), will provide ample editing headroom and will help minimize color transformation on output.

For these reasons, it is a good idea to try to convert the large amount of color that you might capture with your digital camera or scanner into a working space that will not drastically alter your color and will give you enough headroom to make color adjustments without pushing your color out of gamut. Additionally, you will want to be able to convert your color, upon output, to devices that have the majority of their color space contained within your working space. Ideally, you would want to have all of the printer’s color contained within your working space, but that is not always possible—especially when you plan to output to numerous output devices or mediums.

Figure 22 shows the output color space for an Epson 2200 inkjet printer when printing on Moab Entrada paper. The printer’s color space is fully contained within Adobe RGB. It is large enough that any colors assigned within Adobe RGB that are outside the 2200’s color gamut will not have to go through a drastic transformation when converted for printing.

Working spaces are good places to “hold” color if you are not sure what your output device will be—especially if you are archiving your images for future use. Because most high-end digital cameras provide RAW capture into 16-bit files, many photographers are putting their images into large color spaces like ProPhoto RGB. Not only does a 16-bit file, with a potential of 65,536 tonal levels, give the photographer enough image data to edit his images without the risk of introducing banding, but archiving into a large color gamut retains a maximum amount of color data for future output. Anticipating that future printers will have much greater color output capabilities with a much larger color gamut, it makes good sense to retain as much color data as possible.

In this example, putting image color in a large working space ensures that the photographer will be able to take full advantage of future printing technology. Even if the photographer needs to provide his image today to a client or outside service provider, spinning off an interim conversion to a smaller working space like Adobe RGB while maintaining the original in a larger working space is a viable solution. Figure 23 illustrates this workflow.

Figure 23—Converting a RAW image into a color space like ProPhoto RGB preserves most of the color data because ProPhoto is a relatively large color space. When preparing a file for a client, it would be wise to first run an additional conversion to a smaller RGB color space.


Color Management and Workflow

Back to list

to top

New to PhotoVideoEdu? SignUp now to see EDU discounts!

Log In
With Social Account
You can use your social services accounts to login to our system, but if you're logging in the first time please select if you are a



With E-mail