One major drawback of raytracing is the large amount of computer time it consumes. Even on a powerful minicomputer, such as the VAX 11/785, the time required to generate an image containing just 10's of objects is generally measured in hours. The time required for an image approaching the complexity of a normal natural scene would be measured in days, weeks or even months of computer time.
Several attempts have been made to speed up raytracing. Most of these attempts have focused on reducing the number of costly ray object intersection calculations required to generate an image. The work of two Japanese workers, Fujimoto and Iwata [FUJI85], stands out among attempts to speed up raytracing. By the use of a 3 dimensional extension of the well known "digital differential analyser" algorithm, used extensively in drawing lines on raster displays, these two workers achieved some very spectacular results. One picture that took 2 hrs and 15 minutes to generate, using the system they developed to test their algorithm, would have taken an estimated 40 days by classical raytracing.
At present raytracing is an active area of research in many Universities and other organisations all over the world. In recent years, most effort focused on raytracing has been directed at two main areas:
(1)
Speeding up raytracing and
(2)
Extending the range of special effects that raytracing
can produce, such as textures.
There are several journals that publish the bulk of raytracing research. One of these is "Computer Graphics and Applications" which is published by IEEE. IEEE stands for Institute of Electrical and Electronic Engineers. Another worthwhile journal is "Computer Graphics, Proceedings of SIGGRAPH". SIGGRAPH stands for "Special Interest Group in Computer Graphics" and this organization holds an annual conference which usually has at least one session devoted to the subject.
Amanatides's review of Realism in Computer Graphics [AMAN87] provides a good overview of where raytracing fits in with other research on producing realistic images.