Workflow

A workflow specifies the steps you take from opening an image in Photoshop to exporting the image for the web or printer. The following steps are recommended:

Not all steps are required for every image. For example, you typically don't need to correct for lens distortion when working on a portrait, or remove noise for low ISO images. The rough crop removes extraneous and distracting detail. Not only does this make it easier to edit, but extraneous detail can skew results for commands such as Auto Levels.

Use Levels and Curves to bring out blemishes and artifacts before retouching with the Rubber Stamp or Healing Brush. Resizing after retouching helps smooth out retouching imperfections. Sharpening should be the last step, as sharpening parameters depend on target resolution.

After retouching, and before the final crop, archive a copy of the image. Be sure not to over-write the original. This is your digital "negative", and you may want to redo things later.

8-bit vs. 16-bit

Workflow is affected by whether you choose to edit in 8-bit or 16-bit mode. An 8-bit RGB image has 8 bits dedicated to each color channel. That is, each color ranges in value from 0 to 255. For a grayscale image, where all three channels contribute equally, this results in 256 steps or tonal values. For a color image this implies 256x256x256 or 16,777,216 possible colors.

A 16-bit image has 16 bits dedicated to each color channel. Each color ranges in value from 0 to 65535. When a 16-bit file is opened in Photoshop you'll notice that the maximum pixel value is 32768. Photoshop maps pixel values to a 0..32768 value range for efficient operations. When saving a file the pixels are mapped back to 0..65535. Thus 16-bit images are closer to 15-bit images in Photoshop. At any rate, the extra bits give the potential for more precise color rendering. Consider the following two 8-bit grayscale patches:

The left side is (128,128,128) and the right side is (129,129,129). The difference would be difficult to detect unless the two patches were adjacent to each other. In isolation one would be hard-pressed to distinguish the 128-patch from the 129-patch. It's even more difficult when textures are involved. Pixel values in one half of the following image have been raised (lightened) by one 8-bit density value. Which side is lighter?

The dividing line went right down the middle of Buzz's suit so that's a good place to examine. If you have a hard time determining which side is lighter then perhaps the extra colors obtained in 16-bit mode are not relevant to your work. Incidentally, the left side is lighter.

Although the extra tonal values offered by 16-bit mode cannot be seen on a display or in print, they do offer an advantage during processing. If your edits involve strong adjustments to image contrast or color then banding, or discrete jumps in color, can be avoided by processing in 16-bit mode. If you wish to archive 16-bit files choose the TIFF format (JPEG only supports 8-bits) or the JPEG 2000 format to save space. Not all tools are available in 16-bit mode. This state of affairs improves with each release of Photoshop.

Output devices such as printers and monitors require 8-bit images so you'll have to convert to 8-bit mode before you export. For my purposes I find 8-bit mode is sufficient unless I observe banding. Then I switch to 16-bit mode, edit the image, then switch back to 8-bit mode.