Adobe Photoshop CS5 One-on-One

Sidebar Title Exploring High Bit Depthsgoes Here

lot of bits. For example, a pixel with 2 bits can be any of 4

colors—00, 01, 10, or 11. Each additional bit doubles the

potential of the pixel, so you quickly progress from 8 colors

to 16, 32, 64, and so on. A total of 8 bits gets you 2^8 = 256

colors. Hence, a typical digital photograph is said to be an

8-bit-per-channel image. Three channels times 8 bits each is

24 bits, which means 2^24 = 16.8 million colors.

The name given to the number of bits assigned to a pixel is

bit depth. So in addition to the height and width of an image,

Photoshop sees a third dimension of bit depth associated with

each and every pixel, as illustrated on the left.

Various digital cameras and scanners support higher bit

depths. A typical digital SLR supports 10 bits per channel,

or 30 bits in all, which translates to 2^30 = 1.1 billion colors.

Other devices capture 12 bits per channel, 36 bits in all, for

236 = 68.7 billion colors. Meanwhile, Photoshop supports

two even higher bit depths: 16 bits per channel (which trans-

lates to 2^48 = 281.5 trillion colors) and 32 bits per channel (or

296 = 79.2 octillion colors).

The benefit of so many colors is somewhat theoretical. First,

your computer’s operating system cannot display more than

16.8 million colors. Second, even an extremely large image—

such as one captured with a professional-quality drum

scanner—might contain at most 100 to 200 million pixels,

and most images contain a fraction of that. Given that

each pixel can display just one color, that leaves at the

very least billions of colors unused.

The advantage: So many potential colors are lying fal-

low that you can apply multiple radical color trans-

formations without harming your image. It’s like a

game of musical chairs in which each contestant has

a million chairs. No one comes within miles of each

other, let alone sits in an occupied chair.

Consider the examples in the upper-right figure on

the facing page. (Hey, it’s an advertisement—the

product name is supposed to be misspelled.) I start

with a photo from a stock image library. Then

I choose the Levels command and adjust the

Output Levels values to dramatically reduce

the contrast of the image so it can serve as

To understand how color works in an image, you have to start
at the very beginning, at the assembly-language level: Your
computer is a binary machine. It thinks and communicates
entirely in the binary digits 0 and 1, known as bits. Pixels
began life as single bits, either 0 or 1, black or white. Hence
the earliest black-and-white images were called bitmaps.

Nowadays, a typical RGB image contains no fewer than 256
luminosity values per color channel, which means a whole

A 24-bit image shown head-on, the way we users see it

The same image with

colors mapped into the third dimension

302 Lesson 9: Pro Photography Tools