every entity made out of brushes in your map is a unique model and counts towards the total. This even includes triggers since they have a brush.

On top of that, every unique external model file used by the level (*.mdl, *.spr, and *.bsp for the ammo boses) is counted once. So 10 ogres that use ogre.mdl, will only count once. (But then there is the ogre head too, so that's two unique models.) 

for func_wall or func_illusionary etc bmodels, you can have multiple brushes inside the single entity. So for example if you have say 20 brushes in a room that you want to be func_illusionary, you can select them all and create a single func_illusionary that will count as 1 bmodel. Instead of 20 seperate func_illusionary that will count as 20 bmodels. 

Thanks metlslime and RickyT23.

I was confused because bspinfo was telling me 193 and I thought I was safe.

I just now added up all the brush based entities, including illusionaries and one worldspawn (as reported in Netradiant) and I get a total of 193. The difference is obviously the monster/ammo/item models. I would assume the torches count also (one per type?).

And thanks Daz, that's a good tip. I should be able to trim the count a little bit by doing that.

I know it's not a real problem to exceed this particular limit, but I'm trying to avoid it if I can. 

When adding multiple brushes to a single func entity, visibility has to be taken into account. For sometimes, if a good portion of such a func is in a vis-blocked area, it's possible that the whole entity disappears when looking at some of its brushes from a certain angle (outside the room). At other times, it may just render everything regardless of the surroundings.

Multiple-brush illusionaries can easily cause "too many efrags" warnings, so you may want to make sure they are not too far apart. 

"When adding multiple brushes to a single func entity, visibility has to be taken into account."

Yep, I ran into this problem trying to group some DM spawn pads into one func_wall. In some spots they would disappear even though they were in plain sight.

Oh well. I was still able to eliminate a lot of models just due to re-evaluating how I was doing things and combining some stuff.

Also, last night I was testing with Reaperbots and got a Movetype_push error. I did a little checking and found a bunch (like 5 or 6) brush entities with no brushes. I'm not sure where they came from but I don't think I put them there. I suspect some bug or quirk in Netradiant or possibly I'm doing something wrong.

Anyway, that freed up a few more models, so now I'm down to 239 total which will hopefully leave enough to finish this map. 

Yes, sometimes Radiant doesn't delete the entity correctly when deleting the brush. No idea when then happens, or if a certain action has to precede it. There's an "ungroup" option. Such rogue entities can be removed by going through the entity list (L) and looking for brush-based entities without a ">" icon next to it.

Bengt Jardrup's engines and possibly Fitzquake show a warning in developer mode if such an issue is present. 

Sometimes. IT allows you to select the problematic entities via the 'check for problems' dialogue, then I just delete them with the delete key :D 

It happens when you add a brush entity to another (combining them). Doesn't happen when you just add brushes to an entity. 

Modern engine should fix the flicker problem with large func_s. It was disussed and solved at i3d, I think, long ago. 

Is it possible to have a monster clip or any way to contain monsters to a specified area? In Quoth or with engine mods maybe? 

A trigger monsterjump will do it, with low values.

There are a few other, possibly cleaner hacks that can do it - for example cutting a 16 unit trench in the floor, then filling it with a func_illusionary will stop monsters from walking over it, since it interrupts their navigation. 

My skies end up solid, i.e., rockets explode! What could be wrong? I use tools from http://user.tninet.se/~xir870k/ and GTKRadiant 1.5.0. 

Oddly knights, hellknights and dogs walk over even a 32 width x 32 depth gap, even though they don't walk down a 32 high ledge.

Trigger_monsterjump 10\10 does the job, but they get stuck for a bit. 

I was only remembering, now we use progs for that sort of stuff.

Want an (unpaid) job?

http://icculus.org/remakequake/forum/index.php

If not, then your only recourse is an altered progs - no doubt Preach knows a decent way of doing things in id1. I managed it once under Quoth, but as I remember it was very limited and only with melee enemies. 

I think that's an engine issue. 

Nonsolid sky: Txqbsp does it automatically, Treeqbsp needs the -transsky switch for it to work.

'Monsterclip': Low trigger_monsterjump works just fine, but 10/10 sounds a tad low. Experiment with the values - so that the monsters don't get stuck, but don't visibly jump, either. It's not a perfect solution, but it does the job. 

I guess I should have tested it in another client as it seems that DarkPlaces always have solid skies :(

Btw whats the difference between Txqbsp and Treeqbsp and when is one preferred over the other? 

first, i'd use neither, unless you are talking about aguirre's versions of those compilers. second, txqbsp is better iirc, since aguirre spent more time on it. i've noticed it seems to handle complex brushwork better but that's nothing concrete. 

Make sure you have all the faces of a sky brush covered in a sky texture. 

Find LordHavoc on irc.anynet.org and tell him :-P

Tx vs Tree: Tree always worked well for me, but I agreed to switch to Tx when it came time to choose a compiler for BSP2, simply because everyone else seems to use Tx and I wanted to avoid an unnecessary conflict. Greg lewis did good work with Tree but MH said the code was pretty messy when compared to Tx.

Tx compiles all my maps fine, and the BSP2 version compiles all RMQ maps just fine and supports proper bmodel rotation as well.

I guess the BSP2 tools should be in the RMQ Demo 3 as well as the engine, but we should make it clear that these tools / that format is experimental. 

But they are things that the community has been wanting for a long time - "no limits" mapping with colored light support, proper rotation (qc dependent) multi-threading as standard and a few other features.

See no reason not to include them, even though they are beta. 

mh mentions on his blog that high resolution lightmaps are becoming a possibility in Quake due to some new tech he has written.

Just wanted to say FUCK YES PLEASE to that, the sooner the better :D 

Forgot about that one. 

I appreciate the dry humour there. 

Is there any way to stop Quoth monsters infighting? I know knights of any kind wont fight each other, would like to have base monsters behave the same way and dont agro at their enforcer brothers.