Color depth is the actual number of different colors that you have available. (Remember that each pixel can be only one color at any one time.) When you work in 8-bit/channel color (simply called 8-bit color), each of the component colors is recorded with exactly 8 bits of information in the computer file. (At the beginning of this chapter, I mention digits. These are the actual numbers — the zeros and ones recorded on the hard drive to track each pixel's color.) In an 8-bit RGB image, each pixel's color is recorded with three strings of eight characters. When you work with 16-bit/channel (or 16-bit color), each of the component colors is recorded with 16 characters. The larger numbers mean more possible ways to record each color, which means more possible variations of color (as well as files that take up more space on your hard drive).
What that means to you, in practical terms, is possibly a better-looking image when working in 16-bit color. You'll have smoother transitions between colors throughout your image, no banding in gradients (those annoying areas in a gradient where you can actually see one color stop and the next color start), and no splotchy shadows. Posterization, which I explain in Chapter 5, is the degradation of your image's appearance when similar colors are forced to the same color, making transitions between colors more abrupt. Many tonal and color corrections that produce posterization in your 8-bit images won't harm a 16-bit image in the least. Take a look at Figure 6-5. A Levels adjustment is increasing the image's tonal range. When the same adjustment is applied to an 8-bit version of the image, some rather substantial posteriza-tion becomes visible in the Histogram palette (represented by the empty white gaps in the histogram).
So, should you use 16-bit color all the time? No. You can't post a 16-bit image on the Web, and 16-bit color is rarely used for CMYK images. Digital photos taken in JPEG format are 8-bit (because that file format doesn't support 16-bit color). And with most inkjet printers, you won't see any improvement in the final print. (You might, however, see a dramatic increase in printing time because there's twice as much image data to process.)
If you shoot in 16-bit color, whether TIFF or Raw, it makes sense to process the image in 16-bit color. When the image is perfect, you might want to convert a copy of the file to 8-bit color for printing (ImageOModeO8-Bits/ Channel). You might find some situations in which you'll convert an 8-bit image to 16-bit. Switching from 8-bit to 16-bit doesn't help you avoid posteri-zation, but it might reduce banding if you add a sizable gradient.
One other note on color depth: Photoshop CS2 can work with 32-bit/channel images. These monstrous files are called high dynamic range (HDR) images and are typically constructed by combining different exposures of the same photo. In Photoshop CS2, you can work with a pair of exposures in the Merge to HDR feature (presented in Chapter 10). You probably will never be called upon to create an HDR image for any of the esoteric 3-D or video programs that use them, but you might find some benefit in combining multiple exposures of a difficult shot. Afterward, so that you have a file of reasonable size and access to more of Photoshop's features, use the ImageOMode menu to convert to16-bit color or even 8-bit color.
Was this article helpful?