Images must be digitized and stored in a suitable form before any computer manipulation is possible. This is done by dividing the image area into a rectangular grid of small discrete cells known as picture element, or pixels. Each pixel is capable of capturing only a single colour, so any colour variation existing within the relevant area is lost. However, the dimensions of each pixel are so small that the discrete nature of the image should not be visible to the human eye. Other pixel shapes and formations, notably the hexagonal grid, are sometimes used.
Pixels determine the size of an image as much as its quality. Large numbers of very small pixels may produce an unacceptable level of electronic noise, so a better-quality image might be produced by fewer, larger pixels. The quality of an image is therefore determined by not only the lens, the camera's processor and the number of pixels, but also by the quality of the pixels themselves. Nevertheless, a minimum number of pixels is required to avoid pixelization, or jaggies, a visible step-like appearance in a smooth curve or angled line. Broadly speaking, it is necessary to have 300 pixels per inch, both horizontally and vertically, to achieve acceptable image quality. In other words, given sufficient magnification, it should be possible to count 300 pixels along each edge of every square inch within an image.