Seams between brushes


(Grand Nagus) #1

Hey guys. Gotta question…

I’ve been working on a map for some time where I’m using triangle brushes similar to what GenSurf would generate, where I’d cut up my terrain into triangles and then pull the top vertices up to where I want them. None of the triangles meet ‘halfway’, each triangle meets up with another triangle’s tip.

Now for my question… If I use the original q3map that was included with an old version of Radiant, the map compiles fine. Unfortunately, it takes absolutely forever and I can’t take advantage of any of the cool shader goodies that are provided for with Q3Map2. Now if I compile the same map with Q3Map2, I get a few areas where triangles don’t quite meet together… It’s a small gap that runs the length of the seam between two triangles where you can see a hall-of-mirrors effect. So I curse for a little while, and then I select and move the entire map to another grid location in Radiant, recompile, and search for where the same problem has happened at another location of the map (the problem always changes location when the entire map’s moved in Radiant).

Here’s my ‘map’ bat compiler setting:

“E:\Applications\Games\Raven\STVEF\Tools\NewCompiler\q3map_2515\q3map2.exe” -fs_basepath “E:\Applications\Games\Raven\STVEF” -game ef -custinfoparms -meta -patchmeta %1

and this is where I see the occasional seam in my map. Oh, this is for Star Trek Elite Force. Radiant tools don’t come up with any problems with the map, every vertex for every brush on the map has been ‘snapped to grid’, and I’m not scaling the map in any way. The problem is very random, where it happens once at one location in the map, and then the problem is at another location of the map when the map’s moved to a different location in Radiant. If I repeatedly compile the map without making changes to the map, the location of the error always stays in the same place.

I think it’s some sort of precision error, because the map compiles OK with the old compiler (but it takes about two hours). So I think there’s probably a command switch that I’m missing… Is there any way I can get Q3Map2 to compile this map?

Thanks in advance for the help!


(WolfWings) #2

I believe “-mv 999” in the bsp phase (what you’re worrying about right now) will help correct that problem with ‘floating triangles’ in tri-souped brushwork, but I may be remembering some other problem that command-line switch helped with.


(Grand Nagus) #3

Well, I think that worked… I had to move the map in Radiant again because I was getting a funky shadow in the middle of an open area where there was nothing to cast a shadow. Strange stuff; I’m beginning to think my map’s possessed.

But I implemented your idea and now I only get a small sparkle at one of the seams. Which isn’t bad, because that problem is approximately 99 percent more permissible than that friggin’ HOM gap that was there before. I haven’t tried it without the command line switch, because I don’t want to mess up the looks of the map from where the bsp’s at now.

You’ll have to forgive my ramblin’, 'cause I haven’t yet had enough coffee this morning. …need more coffee…

Thanks for the idea!


(Mr_Tickles) #4

Ummm, copy the current BSP into another folder then re-compile… you’ll still have the BSP you like to revert to if you’re unsatisfied.

I’ve found the problem with vertex editing a hell of a lot, and yes, I do mean ‘Hell’. The way I have found of getting it to work, even when it looks fine in Radiant is to make sure I’ve moved the vertex of every triangle at that point so they all snap to the same grid points. If you just do one and move it to where the other triangles point’s appear to be at, there may be a small discrepancy not visible on first, or even second inspection. I’m currently doing this a lot with the terrain I made in Easygen. Also I’m trying to create a large cavern using this method… complete nightmare! As for your map being compiled differently with the different compilers, I’m not sure why that would happen, but if something was done badly in q3map and ydnar fixed it, it’s probably for the best.


(ydnar) #5

Can you post a screenshot of this? Particularly with r_showtris 1 and r_clear 1?

If you’ve vertex-edited a lot of brushes, chances are you’ve corrupted some and the geometry is borked. Try using Brush Cleanup on the map file and recompiling.

y


(ratty redemption) #6

I use the same technique that Mr Tickles mentioned, as regularly run brush cleanup, and considering I map almost nothing but trisouping and vertex editing, I haven`t had any sparklies for ages, so these techniques must be working, without the need for me to move the whole map or trisouped area.

good luck with this though, as what your going through would annoy the hell out of me if it kept happening to my maps.


(Mr_Tickles) #7

Glad to hear someone else is doing the same Ratty :wink: I nearly went to Bath Uni… shame, could’ve exchanged tips, lol. Although not on hairstyles :wink: :smiley:
Yes, brush cleanup is most definately a divine gift, use it frequently and it will be easier to correct the problems that it un-covers.


(ratty redemption) #8

hee hee and cool :wink:


(Grand Nagus) #9

Thanks for the replies, guys. However, I thought I had the problem licked and then THIS happened again…

In this picture, I’m staring at a hill made of two triangular brushes. The verts were stretched to make it the shape you see. It looks fine, for now. But if I step toward the hillside…

The hillside disappears and I see a blasted HOM effect. All of my Radiant plugins turn up no errors or bad brushes, and q3map2 returns no errors during compile. I tried moving the map around in Radiant, and I tried rebuilding all of the brushes that make up the hillside as well as all of the surrounding terrain brushes, but no dice. The problem still occurs, and this has happened to me before on this map (even after I completely rebuilt the map from scratch).

Anyways, this problem happens when bsp is run (no vis or light is compiled into this map). My bsp batfile compile listing is the same as what I posted above, and I tried the ‘-mv 999’ switch (caused errors when I compiled another map that had an undulating flag shader in it). Anybody gots any ideas? My head’s gettin’ sore from banging it on my monitor.


(Grand Nagus) #10

No idea, huh? Oh well, back to the drawing board.