The days of using density may be numbered, however. CIELAB (full name is CIE 1976 L*, a*, b*) is gaining in popularity and is increasingly being used throughout the workflow from prepress to press. Software, such as Photoshop and QuarkXPress, use CIELAB. Color management via ICC profiles is based entirely on CIELAB. CIELAB can be used to measure color on all the different media used today ― not only color from film, plates and press sheets, but also from inkjet proofs and LCD and CRT computer monitors.
Density is specified by the American National Standards Institute (ANSI) while CIELAB is specified by a separate organization, the Commission International de l'Eclairage (CIE). As a CIE system, CIELAB makes color communication easy because you can call somebody up and talk to them in terms of numbers, rather than saying, “This print is too blue” or “This image has a magenta cast.” And because CIE systems are internationally recognized, these numbers refer to the same color to any user anywhere. CIELAB is the most common of all the CIE color systems.
Defining saturation, lightness and hue
Color has three basic attributes ― saturation, lightness and hue ― which makes it convenient to represent color in terms of a three-dimensional volume (see Fig. 1). Near the center of this sphere, colors are desaturated or neutral. As you trace a color outward from the center of the sphere, the color can change in saturation. Note that the color itself doesn't change ― that is, a green stays a green ― but it becomes more saturated or vivid as you move outward in the sphere.
The vertical axis corresponds to a change in lightness, with light colors near the top of the sphere and dark colors toward the bottom. Along the central vertical axis, we can imagine colors going from white to gray to black.
Finally, it is possible to change the main or “starting” shade of a color by moving around the sphere's circumference. This alters color's third attribute, the hue.
In the CIELAB system, a color is represented by three numbers, which specify its position in the three-dimensional volume. The first number, the L value, defines how light or dark a color is. The “a” and “b” tell us about the “color.” A CIELAB of 50, 75, 5, for example, is a red, while a CIELAB of 50, -75, 5 is a green. A yellow sample would perhaps have CIELAB of 70, 0, 80. Two samples that are the same color and change only in lightness would be, for example, 50, 50, 50 and 70, 50, 50.
Because colors are specified in terms of numbers, it is easy to go one step further and describe the difference between two colors with a numerical value. This difference, expressed as the numerical difference between the two CIELAB numbers, is called Delta E (3E). 3E can be used to express the difference between a press sheet and a contract proof, for example. A 3E of 1 means the difference in color is barely distinguishable. In commercial printing, a color difference of 4 3E to 8 3E may be acceptable.
Mimicking how humans see color
One of the biggest differences between density and CIELAB is that CIELAB is closely related to the way humans see color and so performs better when measuring press sheets. Density can sometimes be misleading because it doesn't deal with all colors equally.
Consider this example: We measured some samples from an offset sheetfed press using a spectrodensitometer from a reputable vendor. First we measured two dark samples; the densities were 1.73 D and 1.92 D. These patches had a measured density difference of around 0.2 D but were visually identical. We then measured two lighter samples, which were 0.3 D and 0.52 D. These also had a density difference of around 0.2 ― but this time the color difference was clearly visible.
What does this mean in practice? Suppose, as is the norm, density is used to measure a press sheet. The pressroom manager could ask the operator to compare the press sheet to the contract proof for a given color. “Are you hitting the target density?” The press operator may answer, “We are 0.2 D away.”
If this discrepancy is measured in a dark area, the customer would see no visual difference. If, however, the same density difference occurs in a neutral, midtone color, the job would have to be re-printed.
Our team then remeasured the same press sheet, this time in terms of CIELAB. We obtained a 3E of 3 for the dark set of patches and a 3E of 14 for the lighter patches. When there is little visual difference between color samples, the 3E is low. When there is a large visual difference, the 3E is large.
Analyzing all 3 dimensions
As previously explained, color has three basic attributes and is therefore fundamentally a three-dimensional concept. What does this mean? We measured a cyan, magenta, yellow and black “step wedge” printed on an Epson 5000 inkjet proofer. We measured the sample using density and CIELAB and compared the results.
Fig. 3 shows each of the four color scales (a), plotted in terms of density (b) and CIELAB (c). The three-dimensional CIELAB plot provides information that is not evident in the density graph. Note, for example, the grayscale. Correct reproduction of neutrals is important, as the human eye is extremely sensitive to even the smallest cast in gray colors. The density plot doesn't indicate whether the grayscale is neutral or if it has a color cast. In the CIELAB diagram, though, we can clearly see that the grayscale does not go straight up the central axis but in fact twists round, exhibiting a different color cast in the highlights, midtones and shadows.
Also examine the magenta scale in the CIELAB diagram. As the color gets darker, the patches move away and down from the apex as expected. But then the magenta patches twist, indicating that the darker areas are too red. Magenta at a higher dot percentage exhibits a significant color cast.
Why is there such a difference between density and CIELAB? In a densitometer, measurements are made through separate color filters and the data is not merged. When calculating CIELAB, however, the measuring instrument's hardware and software does considerable data integration and post-processing.
Color and paper stocks
Most printers are aware of the effect of the substrate on the printed product. It is generally accepted that coated stock will produce higher-quality results than uncoated stock; that glossy paper will hold more color than newsprint. How do we quantify this effect and relate it to images?
Using CIELAB, it is possible to see whether the colors of an image will be contained within the gamut of the print process. The colors from the mailbox image (Fig. 4a) are displayed alongside the color gamut (color capability) of uncoated printing paper (b) and of coated printing paper (c). Note that many of the image colors lie outside the gamut of the uncoated paper. If uncoated paper were used for this job, many of the yellow and red portions would not be reproduced properly. The gamut of coated paper, on the other hand, totally encompasses all the image colors and can easily reproduce every part of this image. Show these diagrams to your client to justify printing their job on more expensive coated paper!
Printers can also use these diagrams to compare the gamuts of different devices. Suppose you were seeking a relatively inexpensive proofer. You could, with CIELAB, compare the gamut of an inkjet printer with the intended press to see if all colors that are reproducible on press can be simulated on the inkjet proofer.
Conclusion
Density has served the printing industry well. Nevertheless, with the continued introduction of new technologies and new ways of working in the graphic arts, such as digital cameras, screen-based proofing and inkjet proofing, press operators need new ways to measure and control color. It is often necessary to use more advanced and visually relevant color metrics than simply density. The quality and affordability of today's color-measuring devices means that there has never been a better time to start using CIELAB to control color in digital color imaging.
Instruments for measuring color
Once you have decided to go with CIELAB to measure and communicate color with your clients, you must choose the appropriate measuring instrument. Measuring devices can be divided into three main categories: densitometers, colorimeters and spectrophotometers.
Densitometers
Densitometers are the simplest type of color-measurement device, but they only report density. They are unable to calculate CIELAB.
Colorimeters
Colorimeters measure a sample and compute CIELAB values from the data. They are light, compact, reliable and inexpensive. In the graphic arts, colorimeters are most commonly used to measure computer monitors.
Spectrophotometers
The most sophisticated color-measurement instrument is a spectrophotometer (Fig. 2). A spectrophotometer measures the spectrum of a sample, reporting the reflectance or transmittance at regular intervals. The spectrum is the most complete description of a color, and it can be used to calculate all other metrics, including density and CIELAB. Because a spectrophotometer can do the job of a densitometer, some devices are called spectrodensitometers.
Today, neither the price nor the size and ease of use is a significant differentiator between a colorimeter and a spectrophotometer, so in most instances a spectrophotometer is the logical choice. You may want to invest in some visualization software along with your hardware purchase.
Color-management basics