Now we come to what I feel is the most exciting part of the book—and what is perhaps the most exciting development in PC 3D graphics since hardware-accelerated Transform and Lighting. The 3D pipeline that OpenGL uses is pretty much the standard pipeline for creating real-time 3D graphics, regardless of which API you are using. However, graphics techniques have evolved to the point that the design of the pipeline has reached its full potential. True photo-realistic rendering requires a great deal more processing applied per vertex, or even to the fragments rendered between vertices.

To increase realism and extend graphics beyond what is possible with the standard pipeline, vendors have designed their hardware to be more flexible and allow not only the processing of a fixed set of commands and state variables, but also the actual execution of graphics code on their now renamed Graphics Processing Units (GPUs). The programmable pipeline is perhaps the single biggest development in commodity graphics that we will see in our careers. One vendor calls it “Cinematic Computing,” and that description is not far off the mark. We can now approach in real-time the photo-realistic effects that used to take hours or days to create. OpenGL, too, has evolved to meet this new era.

The following chapters take you through some of the more recent innovations in 3D graphics: from finer grained control of graphics processing memory to extended buffer capabilities and finally to a full-featured graphics programming language executed by the graphics hardware.