Quake3World

What are the q3map2 settings you use for compiling -light for you map when you're ready to release it? Is there some sort of "optimum" configuration?

I guess you'll get very different answers to this question. I use

-light -dark -dirty -patchshadows -filter -bounce 3 -gamma 1.4

Leaving out -fast takes hours and hours more to compile, but it's really worth it. You've read about it here?

-dark creates darkened edges and creates more visual depth this way
-dirty darkens small gaps and cracks
-patchshadows is used to make patches cast shadows
-filter will smooth your lightmap. I think the outcome of -filter ist much much better than the one of -samples, even if most people will disagree with this. -samples sharpens your lightmap and can cause problems because of this. You can end up with too sharp edges on your lightmap instead of smooth transitions from color to color. -samples also elongates compile time and increases BSP size
-bounce 3 three bounce stages are enough when not using fast, because there is much more light data to compile and your map gets bright enough with three stages
-gamma 1.4 that's a value you have to figure out by yourself. I use 1.4 for some time now, it's not too bright but bright enough to make your lightmap look uniform and not like a dark cellar

I use

-cpma?

Afaik, that's enhanced vertexlighting.

I assume you need a newer version of q3map2 for that? I've never seen that switch before. -dirtdepth and -dirtscale are also new to me.

I currently use Q3Map2 v2.5.16

Those switches are all available with Q3Map2 2.5.16 (well, I've never tried CPMA).

-bsp -meta -patchmeta
-light -lightmapsize 1024 -bounce 12 -bouncegrid -export -lomem


everyday I remind my computer "it's a hard life"

lightmapsize 1024 eh? are you crazy?

Solitude used this I think:

-bsp -meta -samplesize 16 -v
-vis -v -saveprt
-light -patchshadows -fast -dirty -dirtscale 1 -dirtdepth 16 -samples 6 -v -bounce 100 -threads 2

-samplesize 16 on -bsp?

It affects surface meta triangles. I think larger values result in fewer meta triangles.

Oh, I thought that was a switch to be used in combination with -light? Or does it exist for both?

Eraser wrote:lightmapsize 1024 eh? are you crazy?

for the purposes of my current employment, no.

-samplesize 16 on the bsp phase sets the q3map_samplesize shader parameter to 16 on all surfaces. It needs to be entered into the lightmap phase too afaik in order to work.

Hipshot wrote:-bounce 100

It's a good thing Q3Map2 culls light traces, unless you are also using q3map_bounceScale 1000 or something.

Sorry to bring that old topic back up, I'm in the process of trying different settings for a final compile and I was wondering what would be your thoughts on lightmap and vertex lightning, the new settings in particular, the -cpma switch. Is it really much better than the previous switch of q3map? From what I understand, correct me if I'm wrong, I can compile a map with both settings, lightmap and cpma and only one of these 2 method will be loaded on the client side, based on the client q3 settings.