Home/Technical/The confounding world of Digital Color

The confounding world of Digital Color

Artists need to photograph their work. In the modern world of digital cameras and printers, what could be easier? Take that piece outside, lean it up against the wall and pull out the old ‘point and shoot’ camera (or even your iPhone) and snap away. When you download the image and pull it up on the monitor, you discover that the colors are all funny. Put a piece of photo paper in the printer and print, and the colors are still all funny, just a different funny.

What is going on?

Welcome to the very complex, powerful and sometimes bewildering world of digital color.

The bad news is that there are so many ways you can mess yourself up in this world it sometimes seems a miracle that you can ever get a decent photo of your work. The good news is that unlike the days of silver based film, many artists have powerful tools available to them to help them get the color right.

I am not going to give you all the answers here, but I will try to explain a little how the world of digital color works. It should help you make better sense of your tools and give you the ability to ask the right questions should you decide to seek professional help.

Let’s start with the camera. The picture file that comes out of your camera (usually a .jpg) in its simplest abstraction represents a stream of pixels, each represented by three numbers, representing the Red, Green and Blue components of a color. Specifically, it represents the nearest color in the target color space. The default color space for most cameras is sRGB. It defines a finite set of possible colors (called the gamut). Note that not all colors exist in this set. Other profiles exist that have a different gamut, that is a different set of possible colors. (two common ones are Adobe RGB and ProPhoto RGB). The image processor in the camera knows how to take the raw numbers that come from the sensor and map them into this standard space. A simplistic way to think about it is to imagine that there are three numbers (small positive or small negitive values) that get added to each of the Red Green and Blue values from the sensor to transform them into the correct values in the sRGB space. Except it is more like a set of numbers for each possible combination or RGB values. What actually happens is a good deal more complex than that, but this abstraction will do. This set of correction values is specific to the design of the sensor. It is called the profile for the sensor, and it is built right into the camera (if you are using a digital SLR and shooting in .RAW mode, the profile is built into the program you use to create the .jpg on your computer, such as Adobe Camera Raw which is part of Photoshop).
So your camera has now created the perfect image, right? Well, close. One of the important pieces of information that is missing is the exact color of the light that was illuminating the scene when the camera took the picture. This is the white balance or color temperature setting (expressed in degrees Kelvin, eg. 6500°K). When your camera is in auto mode, the image processor takes a guess, based on what it sees, of what the correct value is and stores this in the image file. The problem is, that this is at best a guess and is often wrong. You can usually influence this using camera settings, or you can set it after the fact using Photoshop. In order to do that accurately, however, you must have something white (or neutral grey) in the photo. Taking two shots, one with a piece of printer paper in the view will give you the info you need. Camera Raw will then allow you to point at that part of the image and say, in effect, this is white. In my experience, getting this set right is the single most important part of color correction.
After you load your image onto your computer, you want to look at it. So the operating system reads the RGB values out of the image file and using a different profile, sets the values of the pixels that will be displayed by the monitor. This profile may be a generic one or it may be one provided by the manufacturer of the monitor. In either case, what is happening is that the process that went on in the camera is reversed. Here you are tweaking the values so that the monitor will display the color closest to that specified by the color space. The problem is that this profile may not be very accurate. Have you ever seen the same file look radically different on different computers? The profile is the problem. The generic profile usually isn’t even close. The manufacturer profile gets you in the general neighborhood, but doesn’t account for variations between different monitors of the same model or performance under different lighting conditions. So, if the color doesn’t look right on the screen, is the problem in the file or the profile? The only way to know is to calibrate your monitor.
Monitor calibration is done with a device that reads the colors that the monitor is actually putting out (and, perhaps, the ambient light conditions as well) and uses a program to create a custom profile for your monitor. The operating system then uses this profile to adjust the values, resulting in the truest representation of the colors represented in the file. Then if the color on the screen doesn’t look right, the problem is in the file and you can use Photoshop or some other program to make the necessary adjustments. These calibration devices vary widely in price and accuracy. The Pantone huey Pro is a good entry level tool that usually sells for less than $100USD. At the high end, it is like stereo gear: the sky is the limit. The important thing to realize is that ANY calibration is better than none, and the simplest calibration will get you 80% or more of the possible improvement.
The third step in the chain occurs when you go to print a copy of your image. Your trusty photo printer comes with a set of profiles that tell the print driver how to tweak the values to get the right color on the page. It is important to remember, however, that the profile is specific to the ink and the medium. To get the best color, it is important to make sure that you have the correct paper type set in the print driver before you print. Glossy photo paper is not mat photo paper.
What actually goes on in the printer driver is much more complex. The image file is expressed as RGB values which are additive primaries. This is great when dealing with light, but when dealing with ink, we work with subtractive color. The primaries for that are Cyan, Magenta and Yellow (usually with Black added in). This is CMYK color in computer speak. So what actually happens is that the driver first maps things from the sRGB color space into the CMYK color space and then adjusts things based on the profile. Here again, this is a simplification, but you get the idea.
So… If you want a good image, you can usually get it by doing three things:
  1. Calibrate your monitor. Recalibrate it every once in a while as well.
  2. Set the white balance correctly, either in the camera or afterwards with Photoshop
  3. Make sure that you tell the printer driver what media you are using to print on.

Easy, right?

About the Author:

Dean Russell Thompson is an Artist and Printmaker working in Loveland, CO.