Future Trends in Computer Graphics: How Much is Enough?

Over the forty-year history of interactive computer graphics, there havebeen continuous advances, but at some stage this progression mustterminate with images being sufficiently realistic for all practicalpurposes. How much detail do we really need? Polygon counts over a fewmillion imply that on average each polygon paints less than a singlepixel, making use of polygon shading hardware wasteful. We consider theproblem of determining how much realism is required for a variety ofapplications. We discuss how current trends incomputer graphics hardware, and in particular of graphics cards targetedat the computer games industry, will help or hinder achievement of theserequirements. With images now being so convincingly realistic in manycases, critical faculties are often suspended and the images areaccepted as correct and truthful although they may well be incorrect andsometimes misleading or untruthful. Display resolution has remainedlargely constant in spatial terms for the last twenty years and in termsof the number of pixels has increased by less than an order ofmagnitude. If the long-promised breakthroughs in display technology arefinally realised, how should we use the increased resolution?