The color management problem
A small amount of theory is necessary to understand color management. If you don’t know that there’s a problem, it’s dicult to envisage a solution, so let’s take a glance at the problem color management tries to solve.
Computers are amazing inventions, but they know absolutely nothing about color or tone (or art, truth, or beauty). ey’re very complicated calculators that juggle ones and zeroes to a specied order; everything we do with computers involves representing things—text, pictures, sound, movies—by numbers. e color management problem stems from the way we typically use numbers to represent color.
We’ve come to represent the color of pixels by specifying three values: a red, green, and blue amount, making up the familiar red, green, blue (RGB) color model. An RGB image is made of three grayscale images: one records the red channel, a second records the green channel, and a third records the blue channel. RGB color is appealing because it directly relates to the way we capture color through red, green, and blue lters (or their opposites: cyan, magenta, and yellow) and to the way monitors display images that use red, green, and blue phosphors or light-emitting diodes (LEDs). It’s simple, relatively understandable, and completely ambiguous to actual color appearance.
Red, green, and blue light combine to make white light.
RGB values are basically control signals that you send to devices such as monitors, or receive from devices such as scanners and cameras. e red value in RGB tells a display how many electrons to send to the red phosphors to make them emit a specic amount of red light, and the red value indicates how many photons passed through the red lter of the scanner or camera to record the signal. e problem is that, like people, each device has its own idea of what consti- tutes the color red (and, for that matter, green and blue). Dierent scanners and cameras produce dierent RGB values when they are confronted with the same original or scene. Dierent moni- tors produce dierent colors when given the same RGB values because scanners and cameras use dierent lter sets, and displays use dierent phosphors or LEDs to produce the color.
e rst purpose of color management is to render the ambiguous RGB values unambiguous
by associating them with a specic color as perceived by humans, that is, color appearance. Color management accomplishes this goal by associating a prole with each image. Proles can be quite complex, but all you need to know about them is that they correlate ambiguous RGB, CMYK, and gray device values with numbers in dierent systems that are directly based on human perception. ese numbering systems have names like CIE XYZ and CIE LAB. e CIE numbering systems specify an actual color appearance (for example, a specic shade of red), so they let the proles tell the color management systems exactly which colors a given set of RGB values represent.
e second purpose of color management is to match that specic color appearance as the
image travels through the reproduction chain from camera, to display, and then to print. You can’t match a color’s appearance until you know what the color appearance is, so color manage- ment’s role in specifying color appearance is fundamental.
When photographers shoot lm, color management is simply a useful tool—photographers have the lm itself as a reference for the intended color appearance. But with digital capture, color management becomes a necessity because without it, you don’t know what you’ve captured.
A Color Managed Raw Workflow From Camera to Print