UniKorn Posted May 31, 2002 Share Posted May 31, 2002 I'm getting weird lighting erros when I build my jk2 map. When I compile it for Quake III Arena I get no errors, but once I compile it for jk2 I get blue lights in areas where there is no blue light at all. I checked it with r_showtris, the engine doesn't even draw the areas where there are blue lights located. Any ideas on this? Link to comment Share on other sites More sharing options...
Xcom Posted May 31, 2002 Share Posted May 31, 2002 Does this blue light affects your weapons and entities only or you also see it on normal brushes? Something like this happended to me long ago. (Not sure if this is the same problem). The colored light was projected on guns etc. where no light source was present. Apparently, it's some bug in lightgrid caculation they haven't fixed yet. I solved it by placing a caulked brush directly in the world's origin (x:0, y:0, z:0). Also, radomly placing several semi-big caulked brushes in the void (below and above your actual map) may solve this. Link to comment Share on other sites More sharing options...
Expandable Posted May 31, 2002 Share Posted May 31, 2002 I also have a problem with the light. In the blue base, surfaces have a red light map, though there is most definitely NO red light "seeing" the surface. Here's a screenshot, the red light on the textures should be blue... http://jk2-world.ravencommunity.net/images/others/light-problem.jpg Link to comment Share on other sites More sharing options...
Expandable Posted May 31, 2002 Share Posted May 31, 2002 I wrote an email to Raven Software about this error. Kenn (Hoekstra) replied: That's a vis error that's happened to us as well. It's probably because the architecture is too complex in that area. Have you tried doing a bsp with light extra so it takes more time to do light calculations? Maybe that'll help you. Although I don't understand how this error is possible. Usually, the compiler creates a line from every lightmap point to every light entity (right?). Depending on the length of that line, the lightmap - which is black at the beginning - is brightend up. However, if there is a structual, vis-blocking brush somewhere in the line, the lightmap is not brightened up at this point. So how is that error possible? There must be something totally wrong. As far as I know, an extra light compile creates 5 lines (thus 5 times the compile time). But does that really help? I doubt it, but nonetheless, I try. Thanks for your help, Kenn! Update: If you thought you were doing something wrong or if you even doubted your mapping skills (who did?) Kenn has some motivating words for you ;o) We got the same errors, so it's nothing you did wrong. And as you see, Raven got rid of them. So you can, too. Let's try to find out, how! Link to comment Share on other sites More sharing options...
Jesterspaz Posted May 31, 2002 Share Posted May 31, 2002 Yes, i am getting those same anomlies as well. If i were to do a --extra flag on the light compile, my map would be like 30mb, and it would take like 2 days of compiling. Link to comment Share on other sites More sharing options...
UniKorn Posted May 31, 2002 Author Share Posted May 31, 2002 Fullvis doesn't take me that long, my map isn't finished yet. Xcom, it is not on entities, since there are none of that in the map. Something went wrong with the conversion from sof2map to jk2, rickmus of Raven Soft didn't have a clue what I was talking about, so I don't think it is something that came from sof2map. Link to comment Share on other sites More sharing options...
Expandable Posted May 31, 2002 Share Posted May 31, 2002 Fast Light: http://jk2-world.ravencommunity.net/images/others/light-problem.jpg Extra Light: http://jk2-world.ravencommunity.net/images/others/light-problem2.jpg As you can see, the lighting improved quite a lot - but not enough. I hope I'll find away to work around that as well. Jesterspaz: No, you're wrong. Though the compile time increased about 1 hour (previously 30 mins, now 90 mins), the size of the map file hasn't changed! It's still 10.6 MByte. And that seems logical as the extra flag only affects the compile prozess (which is more precise) but not the end result... Link to comment Share on other sites More sharing options...
RichDiesal Posted May 31, 2002 Share Posted May 31, 2002 I have a feeling the problem is occuring from the engine getting confused between where to combine additive light. Are you using light emitting shaders or light entities? If entities, have you tried moving the entities just a grid unit or two up or down to see how that affects the lighting? Link to comment Share on other sites More sharing options...
UniKorn Posted June 1, 2002 Author Share Posted June 1, 2002 I'm gonna try with a smalled samplesize as suggested in the sof2 forums. Link to comment Share on other sites More sharing options...
Expandable Posted June 1, 2002 Share Posted June 1, 2002 *adds to his list: *Try moving light entities (by RichDiesal) *Try hint brushes (by Kenn Hoekstra) *try lower samplesize (8?) (by UniKorn)* Link to comment Share on other sites More sharing options...
RichDiesal Posted June 1, 2002 Share Posted June 1, 2002 Well, that wasn't a technique to fix it... actually, that will just help diagnose the problem... because if you have light entities, move them around, and then the miscolored area moves or changes shape, then you know it's something about those entities... you could then try deleteing and replacing those entities to see if it still does it... or simply move the lights around to a different pattern... I'm all about diagnosing problems, but it's up to the mapper to try stuff out. Link to comment Share on other sites More sharing options...
Jesterspaz Posted June 1, 2002 Share Posted June 1, 2002 i was aware if you lowered the sample size, the bsp will get larger. lower sample size the better looking, and less lagged shadows. i will try to compile my map on -extra, it will take a while i think. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.