So after working a bit on Irrlicht last weekend, I stopped and sat down to write something funny. In the beginning I just wanted to write a nice raycaster
. Maybe you've noticed already, I really love raycasters, maybe because it's the first technology which brought me into contact with realtime 3D graphics. I already wrote raycasters in lots of languages - VisualBasic, Java, Delphi, C++, TurboPascal, C.
But then I thought 'how about trying to write a light mapper instead'? So I started a lightmapper. And if you think about it, it's nothing more than an inverted ray caster. :) In short: This was the result after some hours of work:
The results are surprisingly good: The lightmapper has a very high quality (nearly no artifacts) and is astonishly fast. For the scene in the picture, it needs about half a second to calculate everything with shadows enabled.
I wrote other light mappers several times ago, which had different quality and speed, and I didn't like most of them. But this time I took a different approach and I really like the result. The thing I wrote now is just a toy of course and lots of features are missing to make it useful for users (saving lightmaps and meshes, optimizations, compression etc). The question now is: Should I invest some time to finish this up? Would people use this? Or is it a complete waste of time?