10 seconds loading is a long time if you can prerender t and include the image in the distribution.
I love this kind of raytracing since it makes everything soooo very easy. don't care about normals etc. simply darken a pixel. Unfortunately this is dead slow. It may be used to create nonlinear renders, but for games design it's just too slow. I even made a test and replaced the linepick with a renderworld using a 2nd camera with a viewport of 1*1 pixels and believe it or not, it was almost as fast as the linepick (about 90%). You also may use some soft linepick replacements, I guess there is something in the archives, it may or may not be faster.
There are some possibilities to optimize your code (eg. no pivot required, just use a linepick from the picked coords to the lightsource), but even then it's still way too slow.
There are some other possibilities we have figured out over time, and believe me, a lot of people used to hack for hours to find something fast.
The thing that seems to be the fastest is to position a camera at the lights position, color all shadowcasters grey, hide the shadow receiver, take a render and use this render as a lightmap.
In this example:
http://www.blitzbasic.com/codearcs/codearcs.php?code=332
I get about 200 FPS with my crappy machine. It may not be useful for multiple light sources and/or large scenes, but I guess it would look very nice with some kind of wrestling or tekken style game.
I love this kind of raytracing since it makes everything soooo very easy. don't care about normals etc. simply darken a pixel. Unfortunately this is dead slow. It may be used to create nonlinear renders, but for games design it's just too slow. I even made a test and replaced the linepick with a renderworld using a 2nd camera with a viewport of 1*1 pixels and believe it or not, it was almost as fast as the linepick (about 90%). You also may use some soft linepick replacements, I guess there is something in the archives, it may or may not be faster.
There are some possibilities to optimize your code (eg. no pivot required, just use a linepick from the picked coords to the lightsource), but even then it's still way too slow.
There are some other possibilities we have figured out over time, and believe me, a lot of people used to hack for hours to find something fast.
The thing that seems to be the fastest is to position a camera at the lights position, color all shadowcasters grey, hide the shadow receiver, take a render and use this render as a lightmap.
In this example:
http://www.blitzbasic.com/codearcs/codearcs.php?code=332
I get about 200 FPS with my crappy machine. It may not be useful for multiple light sources and/or large scenes, but I guess it would look very nice with some kind of wrestling or tekken style game.