3D Real-Time Rendering Technology: Research And Applications

In the past three decades, 3d computer graphics has a profound progress, especially in the past several years, with the promotion of GPU technology and 3d applications development, real-time computer graphics has developed much more. And the feature of real-time has also become one of the most important features for 3d applications.Two most important features of real-time computer graphics are good real-time rendering efficiency and good reality of rendering result. Although nowadays GPU can render tens of millions of polygons per second, but within the situation of complex scene, if no acceleration algorithm is used, the requirements of real-time rendering cannot be fulfilled.This paper will have a deep discuss and research on 3d real-time rendering techniques, and by combine with the 3D world editor developing project in which the author practices, methodology and result about how to use the essential outline will be given. Based on how to fulfill the two features of real-time rendering, we will analysis the characters of 3D software/hardware and rendering pipeline, and a systematic solution will be given also.High efficiency method to manage a complex scene is the base of using visibility algorithms. BSP tree and Octree is two most widely-used methodologies. These two methodologies are used for visibility culling and collection detection traditionally. Actually the characteristics determine that they can be used for many other purposes, for instance, shadow creation and ray-tracing methods based on BSP, volume rendering based on Octree.Render efficiency and render quality are inconsistent in common sense, desiring for good efficiency and good render quality separately cannot have a satisfied rendering result. Image based rendering methodology (IBR) can give a good balance between them. In this paper, we will have a sample about how to use IBR to rendering a large-scaled forest scene.Finding the bottleneck of a system is the precondition of optimizing. We propose a good way to find bottleneck in GPU. All kinds of techniques should be combined with each other, and used in different situations of graphics application. Folks have involved in development in computer graphics domain for a long time, so algorithms should have been good researched, so the key to develop a high efficiency computer graphics system is developing and selecting the strategy.In one word, if you learn this paper carefully, you will know how to develop a real-time computer graphics application and know how to make the application run efficiently.