Rendering, Complexity, and Perception

Over the last ten years, computer graphics researchers have labored intensely to strengthen the physical foundations of computers graphics. We now need to step back and examine the nature of scenes our end-users wish to render, and what qualities these rendered images must possess. Humans view these images, not machines, and this crucial distinction must guide the research process, lest we become an increasingly irrelevant enclave, divorced from the users we profess to serve. In Is Image Synthesis a Solved Problem?[Coh92], Cohen examined the degree to which current realistic rendering methods solve problems encountered in real applications. He concluded that these methods do not adequately deal with complexities in scene geometries, re ection properties, and illumination; and with issues involving human perception. In his informal survey of rendering researchers, the two problems of image synthesis considered the \least solved" were environmental complexity and perceptual issues. In this paper, we propose a rendering system that is designed around these two problems. We do not claim to have a solution, rather we have a partial solution made from current technology, and a direction for future development. In Section 2 we examine the qualities required by typical computer graphics applications. Section 3 discusses how these qualities impact our modeling techniques. In Section 4 the importance of perception is expanded. Section 5 summarizes the the requirements for future renderers. In Section 6, we present our framework for a renderer satisfying these requirements. Section 7 presents directions for further work.