As we said earlier, a pixel can have a value of 165, but that doesn't mean anything until you know what image mode the image lives in. That 165 could represent a level of gray, or a particular color, or it might be only one member of a set of three or four other 8-bit values. Fortunately, Photoshop makes it easy to see what image mode an image is in, as well as to convert it to a different mode, if you want.
Ultimately, an image mode is simply a method of organizing the bits to describe a color. In a perfect world, you could say to a printer, "I'd like this box to be navy blue," and they'd know exactly what you were talking about. However, even Bruce and David can't agree on what navy blue looks like, much less you and your printer. So color scientists created a whole mess of ways for us to describe colors with some precisionto each other and to a computer.
Photoshop reads and writes only a handful of the many different color modes they came up with. Fortunately, they're the most important of the bunch, at least for those in the world of graphic arts. Each of the following image modes appears on Photoshop's Mode menu. Note that the mode your image is in determines the file formats you can save in. For instance, you can't save as JPEG if the file is in Lab mode. We'll talk more about this in Chapter 13, Image Storage and Output.
David really wishes that Adobe had picked a different word for this image mode. He insists that all images in Photoshop are bitmapped, but only "flat" black-and-white images, in which each pixel is defined using one bit of data (a zero or a one), are bitmaps.
One-bit pictures have a particular difference from other images when it comes to PostScript printing: the white areas throughout the image can appear transparent, showing through to whatever the image is printing over. Ordinarily, images are opaque, except for the occasional white silhouetted background made with clipping paths (see "Silhouettes" in Chapter 12, Essential Image Techniques).
The other major difference between the other image modes and Bitmap mode is that you're much more limited in the sorts of image editing you can do. For instance, you can't use any filters, and because there's no such thing as anti-aliasing in 1-bit images, you just cannot use tools that require this, such as the Smudge tool, the Blur tool, or the Dodge/Burn tool.
Bilevel bitmaps are the most generic of images, so you can save them in almost any file format.
Grayscale files in Photoshop are always either 8- or 16-bit images: Anything less than 8-bit gets converted to 8-bit; anything more than 8-bit gets converted to 16-bit. Eight-bit is still more common, although most scanners now allow you to bring more than 8 bits into Photoshop.
With 8-bit grayscale, each pixel has a value from 0 (black) to 255 (white), so there are a maximum of 256 levels of gray possible. With 16-bit grayscale, each pixel has a value from 0 (black) to 32,768 (white), for a theoretical maximum of 32,769 possible gray shades.
Few capture devices can actually deliver all those gray shades, so 16-bit files usually have rather a lot of redundancy. But that redundancy translates into editing headroom, so if your camera or scanner can capture 12 or more bits per pixel, it's often worthwhile bringing the high-bit data into Photoshop.
Eight-bit grayscale images are pretty generic, so you can save them in almost any format this side of MacPaint. You can save 16-bit grayscale images in a number of formats, but if you add layers, your choices are limited to Photoshop, Large Document Format, PNG, PDF, and Photoshop Raw, and TIFF.
When you print a grayscale image on a printing press, those 256 levels of gray often get reduced to 100 or so because of the limitations of the press. You can counter this flattening effect considerablyincreasing the tonal range of the printed imageby printing the image with more than one color of ink. This is called printing a duotone (for two inks), a tritone (for three inks), or a quadtone (for four).
The key is that the extra colors aren't typically used to simulate colors in the image; rather, they're used to extend the dynamic range of the underlying grayscale image. Those expensive Ansel Adams books on your coffee table were very likely printed using three or four (or even five or six) different black and gray inks.
Photoshop has a special image mode for duotones, tritones, and quadtones, and even though the file may appear to be in color, each pixel is still saved using only eight bits of information. The trick is that Photoshop saves a set of contrast curves for each ink along with the 8-bit grayscale image. Creating a good duotone is as much art as science.
Note that if you want to place duotone images in a page-layout application for spot-color separation, the safest choice is still to save in EPS format, though InDesign supports Photoshop and PDF duotone-mode files. For more information, see Chapter 10, Spot Colors and Duotones.
As we said, each pixel in a grayscale image is defined with eight bits of information, so the file can contain up to 256 different pixel values. But each of those values, from 1 to 256, doesn't have to be a level of gray. The Indexed Color image mode is a method for producing 8-bit, 256-color files. Indexed-color bitmaps use a table of 256 colors, chosen from the full 24-bit palette. A given pixel's color is defined by reference to the table: "This pixel is color number 123, this pixel is color number 81," and so on.
While indexed color can save disk space (it requires only 8 bits per sample point, rather than the full 24 in RGB modesee below), it gives you only 256 different colors. That's not a lot of colors, when you compare it to the 16.7 million different colors you can get in RGB.
Another major limitation is that most editing tools won't work in Indexed Color, because they almost all rely on the numeric values having a relationship to how light or dark the pixel is. Therefore, you should always do your image editing in RGB mode and then convert to Indexed Color mode as a last stepthe relatively tiny size of Indexed Color images makes them useful for Web graphics, but not for many other uses.
You can save indexed-color images in Photoshop, GIF, PNG, PICT, Amiga IFF, or BMP format (see "Niche File Formats" in Chapter 13, Image Storage and Output).
Every color computer monitor and television in the world displays color using the RGB image mode, in which every color is produced with varying amounts of red, green, and blue light. (These colors are called additive primaries because the more red, green, or blue light you add, the closer to white you get.) In Photoshop, files saved in the RGB mode typically use a set of three 8-bit grayscale files, so we say that RGB files are 24-bit files.
These files can include up to approximately 16 million colorsmore than enough to qualify as photographic quality. This is the mode in which we prefer to work when editing color images. Also, most scanners save images in RGB format. High-end drum scanners include "color computers" that automatically convert files to CMYK mode (see below), but RGB scanning is becoming more common even in shops with these scanners.
If you're producing images for multimedia, or you're outputting files to a film recorderto 35 mm or 4-by-5 film, for instanceyou should always save your files in RGB mode.
Tip: To RGB or to CMYK
A great philosophical debate rages on whether it's better to work in RGB or in CMYK for prepress work. As with most burning philosophical questions, there's no easy answer to this one, but that doesn't deter us from supplying one anyway. If someone gives you a CMYK scan from a drum scanner, work in CMYK. In all other cases, we recommend staying in RGB for as long as possible. We discuss this question in much more detail in Chapter 6, Image Adjustment Fundamentals.
You can save 24-bit RGB files in Photoshop, EPS, TIFF, PICT, Amiga IFF, BMP, JPEG, PCX, PDF, Pixar, Raw, Scitex CT, or Targa format, but unless you have compelling reasons to do otherwise, we suggest you stick with Photoshop (PSD), TIFF, PDF, or EPS.
Photoshop also lets you work with 48-bit RGB files, which contain three 16-bit channels instead of three 8-bit ones. Layered 48-bit images offer great editing flexibility, so we're using them more often now.
Traditional full-color printing presses can print only four colors in a run: cyan, magenta, yellow, and black. Every other color in the spectrum is simulated using various combinations of those colors. When you open a file saved in CMYK mode, Photoshop has to convert the CMYK values to RGB values on the fly, in order to display the image on your computer screen. It's important to remember that when you look at the screen, you're looking at an RGB version of the data.
If you buy high-end drum scans, they'll probably be CMYK files. Otherwise, to print your images on press or on some desktop color printers, you'll have to convert your RGB images to CMYK. We discuss Photoshop's tools for doing so in Chapter 5, Color Settings.
You can save CMYK files in Photoshop, TIFF, PDF, EPS, JPEG, DCS, Scitex CT, and Raw formats, but the first four are by far the most common.
The problem with RGB and CMYK modes is that a given RGB or CMYK specification doesn't really describe a color. Rather, it's a set of instructions that a specific output device uses to produce a color. The problem is that different devices produce different colors from the same RGB or CMYK specifications. If you've ever seen a wall full of television screens at a department store, you've seen what we're talking about: The same imagewith the same RGB valueslooks different on each screen.
And if you've ever sat through a printing press run, you know that the 50th impression probably isn't exactly the same color as the 5,000th or the 50,000th. So, while a pixel in a scanned image may have a particular RGB or CMYK value, you can't tell what that color really looks like. RGB and CMYK are both device-specific color modes.
However, a class of device-independent or perceptually based modes has been developed over the years. All of them are based, more or less, on a color space defined by the Commission Internationale de l'Éclairage (CIE) in 1931. The Lab mode in Photoshop is one such derivative.
Lab doesn't describe a color by the components that make it up (RGB or CMYK, for instance). Instead, it describes what a color looks like. Device-independent color spaces are at the heart of the various color management systems now available that improve color correspondence between your screen, color printouts, and final printed output.
A file saved in Lab mode describes what a color looks like under rigidly specified conditions; it's up to you (or Photoshop, or your color management software) to decide what RGB or CMYK values are needed to create that color on your chosen output device.
Photoshop uses Lab mode as a reference when switching between CMYK and RGB modes, taking the values in your RGB Setup and CMYK Setup dialog boxes into account (see Chapter 5, Color Settings, for more information on this conversion). You can save 24-bit Lab images in Photoshop, DCS, EPS, PDF, TIFF, or Raw format. (You can only save 48-bit Lab images as Photoshop, Large Document Format, PDF, TIFF, or Raw.)
Lab is considerably less intuitive than the other color modes. The Lightness channel is relatively easy to understand, but the a* and b* channels (pronounced "ay-star" and "bee-star") are less so. The a* channel represents how red or green a color isnegative values represent greens, positive ones redsand the b* channel represents how blue or yellow the color isnegative values represent blue, positive ones yellow. Neutrals and near-neutrals always have values close to zero in both channels. Most hardcore Photoshop geeks have a few tricks that rely on Lab mode, but many of them can be accomplished more easily by using blend modes instead. Luminosity blending, for example, produces extremely similar results to working on the Lightness channel in Lab mode.
The last image mode that Photoshop offers is Multichannel mode. This mode is the generic mode: like RGB or CMYK, Multichannel mode has more than one 8-bit channel; however, you can set the color and name of each channel to anything you like.
This flexibility can be a blessing or a curse. Back in the days when color scanners cost a fortune, we used to scan color photographs on grayscale scanners by scanning the image three times through red, green, and blue acetate, combining the three images into a single multichannel document that we then turned into RGB. Ah, those were the days.
Today, many scientific and astronomical images are made in "false color"the channels may be a combination of radar, infrared, and ultraviolet, in addition to various colors of visible light. Some of our gonzo digital photographer friends are using Multichannel mode to combine infrared and visible-spectrum photographs into composite images of surreal beauty.
We mostly use Multichannel mode as an intermediary step or for complex spot color images. For instance, you can use it to store extra channels for transparency masks or selections in other images. Your only options for saving multichannel images are the Photoshop, Raw, Large Document Format, and DCS formats.