Thread: Compile leaves scorch marks

Yeah, definitely. I seem to remember mentioning lightmapdensity in there as well, but this needs to be added as evidence of it.

Btw, I guess those shots are of a single brush floor covering several rooms? This is one of the reasons I break up the work, creating a separate floor for each room, so the light map isn't extended along the floor past the walls . Sometimes lightmapdensity alone doesn't fix it entirely, but I'm glad it has in this instance!

Actually no, not this time. But they are still big brushes.
Yeah I understand what you mean and often break them down to smaller brushes to interrupt the light seeping through.

Yeah, for it to stop seeping on the floor without changing lightmapdensity, you need a separate floor brush that fits snugly around each individual room, leaving a void where the bottom of the walls are. Either way, it has worked! When I found out about these 'hacks' over the years it made mapping so much more satisfying.

Latest compile scorch marks look like burnouts in the car park. lol

shot0021.jpg

The compile before that they were inside the building.

shot0015.jpg

The same crap on both floors.

Both compiles were "-final" as the map is complete.

The main two lines would be closely in-line with the tunnel. But the rest??
Not forgetting the interior walls.
Yes all brushes drawn in caulk before seen face(s) were textured. Also checked after you pointed out some accidental texturing during density editing.

No, it was taking forever with "-bounce 0 > -final" after two attempts. Both times it was crawling so slow through the vis phase, almost as if it had hung. Like only 29% vis after an hour.
I terminated those compiles but never ran the bsp file. Just fired up the compiler again.
There were no errors either, just running slower than molasses on a winter's morning.

So I just went "-final" for a few compiles then I went "-bounce 0 > -final" again and she worked well.
The scorch screenshots (below) were from the two "-final" set compiles.

BSP: -v
VIS: -fast
LIGHT: -bounce 0 > -final

I have been writing the batch file using Map Mate.
The actual "compiler" is the "q3map.exe" and "MOHlight.exe" that comes with Radiant.
So I don't see any difference in MBuilder, MOHPile or Map Mate writing the batch file.
I use the "greater than" sign between multiple compile commands as that's what I saw years ago in a Quake 3 compiling tutorial / forum.
It seems to work.
Sometimes when I have not used it the compile output (Windows command console) drops that "-final" part. In fact I've seen it also drop the "0" from "-bounce 0" when I've failed to add the ">".

Yes it appears to be the radiosity causing the scorch marks as "-bounce 0" has made the world of difference in most compiles since you advised me to use that instead.
Nope, I have always had Vis on "-fast" throughout most maps and only ever change the Light phase of the compile.

My student mapper wrote this morning telling me that he also has just got his first scorch marks. I hope I haven't infected his map by fixing it on my PC.