Digital Processing

Converts the analog scale to a digital one with separate discrete steps between light and dark. These are called gray levels, with a gray scale representing the number of separate steps.

• The more the steps in a gray scale the more accurate can be the reproduction. The price to pay for the accuracy is more processing and storage.

• Thus, it is important to tune the number of gray levels to both the end use and processing power of the system.

• It is generally accepted that the human eye can perceive about 150 separate gray levels at a given time.

• Thus an 8-bit gray scale has been adopted as a standard, since 256 (2⁸) is the smallest power of 2 greater than 150.

• This is especially convenient since a 256 level gray scale can be stored in a single byte.

• The human eye response is not linear in luminance. Thus, there can be some loss of information when a continuous tone image is quantized with equal luminance levels.

• This can be compensated for by either increasing the precision (number of gray levels) or by using different degrees of precision in different parts of the tone scale.

• Varying the precision in different parts of the tone range can create processing difficulties.

• Increasing the precision (to 10 or 12 bits) greatly increases the file sizes and the corresponding amount of data that need to be processed. Nevertheless, some newer desktop scanners support up to 12-bit gray scales (36 bits for color). PostScript can support up to 12-bit gray scales.

• The number of steps in a gray scale is also important in graduated tints and blends. These can be produced in many graphic applications and, if insufficient gray levels are available, can lead to banding (i.e. visible steps).

• For a displayed image, the brightness of each pixel is controlled by the voltage of the electron beam that strikes it.

• The image is generally stored as 8 bits for each pixel on a monochrome display and as 24 bits for color displays (8 bits each of RGB). The minimum intensity is 0 and the maximum is 255. White is represented as (255,255,255) and black as (0,0,0).