Adobe Photoshop works with raster images, such as those acquired from a scanner or a digital camera.
The smallest element in a raster image is a pixel (picture element). Pixels are rendered square, as in a chess board. Adobe Photoshop stores grey-scales as 8-bit images; they are said to have a pixel depth of 8 bits. A bit can have one of two values (0 or 1), which means that an 8-bit pixel can have 28 = 256 different values, corresponding to 254 shades of grey (1–254) plus black (0) and white (255).
In colour pictures, there are three (the additive complementary colours red, green, and blue; RGB, used for screen displays) or four (the subtractive complementary colours cyan, magenta, and yellow plus black; CMYK, used for printing) separate pictures, stored in separate 8-bit channels. Each channel has the appearance of a grey-scale picture taken through a colour filter; displayed in its proper colour together with the other ones, it will give the right colour blend.
Adobe Photoshop also allows the storing of several image layers in each file. A layer may consist of one of more channels.
Brightness and contrast of an image can be modified in several ways in Adobe Photoshop, through adjustment of levels or curves. In the levels control, the pixel values of an image are represented by a histogram, in which the height of each of the 256 bars represents the number of pixels with that value in the image. In the curves control, the user can adjust the shape of a curve defining how each original pixel value of an image is to be translated into a new value. Whereas the curves control allows more flexibility, the levels control is more intuitive and can be recommended for most practical purposes.
The resolution of the image is determined by the original image quality and the pixel frequency. There is no point in increasing the pixel frequency beyond what the actual resolution of the image justifies (except to adjust the magnification of an item in a composite raster image), but the pixel frequency can profitably be adjusted downwards to provide optimal file size. Here are some rules of thumb. For images to be printed with a halftone screen (consisting of black dots with varying sizes; this is the usual mode of printing to paper), the frequency of pixels should be 1.52 times that of the screen. For a 150 lpi (lines per inch) halftone screen, the image should then have a resolution of 225300 ppi (pixels per inch; with respect to the intended final size). Computer screens and web browsers optimally display images with a 1:1 correspondance between image pixels and screen dots. Thus an image intended to take up half the width of an ordinary 640×480 dpi (dots per inch) computer screen should be 320 pixels wide.
The unsharp mask filter is used to restore sharpness to images that have been blurred during the photographic process. It should be applied sparingly, after any necessary transformation of the image (such as resampling of pixels during nonorthogonal rotation or a change of resolution) has been performed. The filter identifies pixels that differ from surrounding ones by more than a specified threshold value, and it increases the contrast by a specified amount. In addition to threshold and amount, the radius of sampled pixels used for the comparison can also be set. Normal values are 12 pixels radius and 50200% amount. Threshold values should be set depending on the nature of the figure; for example, the value can be selected so as to avoid accentuating the grain of the matrix surrounding a fossil.
The effects of using the unsharp mask will typically be more pronounced in a computer screen display than when printed to paper because the halftone screen used for printing will have a lower frequency than the raster image. Thus, when the image is intended for printing, the threshold and amount settings may be set somewhat higher than for screen display.