The "Quest for Realism" has motivated much of the history of rendering, the process of creating synthetic imagery with computer graphics. The earliest work in this area concerned the development of plausible "local illumination" models, the study of how light reflects off a surface. Later work concerned the problem of solving for the equilibrium solution of light reaching all surfaces as the light reflects about an environment, a problem known as "global illumination." The careful characterization of these problems as physical processes that can be simulated with ever-increasing speed and accuracy ranks among the great successes of the field of computer graphics.
However, with this ability to simulate scenes of ever-increasing realism comes a new problem: the problem of depicting and visualizing these complex scenes in a way that communicates as effectively as possible. Thus, over the past decade, a new type of quest has emerged-a quest more subtle and actually more interesting, in my opinion, than the quest for realism. This new (and in some sense larger) quest has more to do with creating imagery that is useful, first and foremost, and also beautiful-rather than just physically realistic. To this end, we can no longer turn to the physical sciences. Instead, we must look to the cognitive sciences, as well as to the fields of art, graphic design, and traditional illustration, where the challenges of structuring and abstracting information so that it can be communicated most effectively-and attractively-have been most carefully studied.
This new area of endeavor, which by way of contrast with the earlier "Quest for Realism" has become known as "Non-Photorealism," or "NPR" for short, has provoked a tremendous level of interest in recent years. Indeed, there has been an absolute blossoming of fascinating papers and techniques in the research literature: from "artistic screening" methods for printing images using microdots with meaningful shapes that might deliver their own message; to techniques for rendering images in the style of pen-and-ink or watercolor or engraved etchings; to procedures for lighting and even distorting three-dimensional models in order to clarify shapes or direct a viewer's attention. The variety and cleverness and even audacity of these manifold techniques never cease to amaze me as I see each new one presented for the first time at SIGGRAPH (the premier computer graphics conference), or at some other research forum.
Now, a great number of these remarkable techniques have been assembled, organized, and presented comprehensively for a larger audience-in the form of this book that you have in your hands. This book provides the most systematic and in-depth study of the field of NPR that has been published to date, and I believe it will go a long way toward making the field accessible to practitioners and researchers alike. By disseminating the many early research results in NPR to a much larger audience, my sincere hope is that the book will also play a pivotal role both in enticing practitioners to refine these approaches-making them really practical for computer-graphics production-and in inspiring researchers to develop ever more creative and audacious techniques.
David H. Salesin
Senior Researcher, Microsoft Research
Professor, Department of Computer Science & Engineering, University of Washington