The aliasing problem in computer-synthesized shaded images.

Abstract : This paper describes work toward improving the quality of computer-synthesized shaded images. Current (practical) hidden-surface algorithms produce an image whose precision is strictly limited by the number of picture elements. Problems caused by this limitation are described and explained in the text. Other image production media such as television and screen printing processes exhibit the same problems, but to a far lesser extent. Therefore, current image production media are surveyed and compared to computer-synthesized images with a view toward establishing a criterion for acceptable image quality. The problems to be seen in computer-synthesized images are predominantly caused by a phenomenon known as aliasing. Aliasing is an artifact of the process of producing an image made of discrete dots. A well known cure for aliasing is motivated and explained in the text. The notions of sampling and convolutional filtering are introduced to assist the explanation. Based on these techniques, an algorithm for computing the intensity of a single dot of an image is developed. The algorithm is applied to a test pattern which serves as a worst case for the problems caused by aliasing.