Monday, October 8, 2007
Curving RGB Color
Color correcting by numbers is a combination of art and science
The digital darkroom offers more control over the imaging process than ever before with the promise of higher quality and quicker, easier results. The price is a steep learning curve and a plethora of creative choices that often leave us scratching our heads wondering where to begin. Mostly, we need to begin, after the image capture, with color correction. This article will help you develop a strategy for correcting images using an analytical approach—by the numbers! (I suggest selecting one of your own photographs and use it to follow along with this tutorial.)
Images captured by a digital camera are made up of numbers that define pixels in terms of varying amounts of red, green and blue brightness. These numbers allow you to analyze image colors in an objective manner. Let's go through the process with this image. At first glance, there doesn't appear to be too much wrong with this photo, but is it as good as it can be (Figure 1)? The Info palette in Photoshop is the most important palette in the application, yet this modest little window on the color of the image is often left closed by most users. Here, you can gather clues to help decide which way to push the colors in the image to optimize a print. Start by making sure you have the appropriate options set for the palette. Select Palette Options from the flyaway menu at the upper right (Figure 2). Make sure you select Actual Color for the First Color Readout and CMYK Color for the Second Color Readout (Figure 3). Though the file is in RGB, you'll be using CMYK values to analyze the color for the best skin tone—more on this later.
When preparing images for print, we must remember that the contrast ratio of the print will never be close to the extremes found in nature or near the range offered by a backlit monitor. You have to make sure that the image fits the widest contrast that the paper/pigment offers to deliver an image that conveys the most depth. Most of the time, RGB images should “place” the darkest shadow values at levels of 10, with highlight values at 245. (Note: The ideal numbers will vary depending on the type of output. The numbers given here are conservative, but practical for a wide variety of paper-based output. It's always best to test the upper and lower limits for any specific output.) These values will generally result in the widest practical range of tones from black to white and still preserve image detail and a sense of texture at the ends of the scale. We'll see that the original values for this image don't come close.
The first task is to determine the location of the darkest and lightest values for this image. To easily visualize where the real light and dark points are, let's utilize a simple trick. First, go to the Layers palette and select Threshold from the new adjustment layer icon at the bottom of the Layers palette (Figure 4). The resulting high-contrast adjustment turns the image into pure black and white areas, with 128 as the midpoint between black and white. To find the darkest area in the image, push the slider to the left until most of the image turns white and only the darkest parts of the image remain black—in this case, at the threshold level of 14. No real surprise that the darkest shadow is in the crevice between the subject's lower back and the red satin robe.
Now place a color sampler on this point: Go to the Tools palette, click on the Eyedropper tool and select the Color Sampler tool from the flyaway menu (Figure 5). Click directly on the darkest patch in the Threshold palette. This engages a circular target, which we'll use later when we return to our color image to make adjustments.
Next, return to the Threshold adjustment by double-clicking on the adjustment icon in the Layers palette—the Threshold dialog will reopen. This time, push the slider to the right until most of the image turns black and only the lightest parts of the image remain white.