Wednesday, January 26th 2000
Discussion: Lightmaps Or Real-time?
Here's one for you guys. Since roughly the time of Quake, most of the big-time game engines have been using static lightmaps to present (the illusion of) "realistic lighting". Static lightmaps, which can be calculated as a pre-processing step, allow for very expensive lighting using techniques such as radiosity, complete with shadows, but can also take up an incredible amount of memory for large data sets. Its also generally slower to render with lightmaps, assuming multiple passes.
On the other hand, we have a lot of people hyping their new real-time lighting ideas. Some as simple as gouraud shading with dense meshes (to be accelerated with the new T&L cards), while others boast of real-time filtered shadows, highlights, and snazzy things of this kind.
So, what do you think? Do you believe one will eliminate the need for the other? Will lightmaps co-exist with snazzy real-time lighting "on top"? Will faster/efficient multi-texturing "save" lightmaps? Do you have another idea about lighting/shadows all together? Share your opinions...
To participate in the daily discussion, simply click the "# Comments" orange link to the right of this news header.