havent used any func_viscluster ....didnt even notice an entity by that name tbh. I replaced all my water textures with the same ones from cyclopean. did a decompile of cyclo and copied it exactly as I could see it was done there. Still no change... EDIT: ok got one problem under control. water is compiling fine now, Im not even gonna says what I was doing wrong cause it's so simple....anyways... :p Just need to figure out how to change the view distance/fog distance underwater caue right now it's so short you cant see shit underwater. Any ideas?
You control water fog in the VMT for the water shader. Let's say you're using water_ajacks01. Open up water_ajacks01_beneath.vmt from /Empires/materials/water/ in a text editor, such as Notepad or Notepad++. You will look for the following variables and their values: Code: "$fogenable" 1 "$fogcolor" "{33 127 158}" "$fogstart" 200.00 "$fogend" 1600.00 Change the values to whatever you desire - you will likely need to experiment and recompile with VIS (you don't need to do RAD for water) and test what it looks like underneath water. Keep in mind that if you are going to make a change to an existing shader, you should copy all files associated with that shader (in this case, all 3 of the water_ajacks01 VMTs) and save them with a new name, change what you wish, set that shader to be used for your map via Hammer, and you will need to include the shaders either in your .res file or pakrat them into your map's final BSP prior to release, so that others will be able to have those shaders and view your water appropriately. If you have any additional questions, you can ask here or reach me in IRC or on Steam.
Thanks for that john. I'll take a look into that. I'm running into a new problem whilst compiling now however. This map is fully optimised and used to fully compile in roughly 8mins. Now since switching to OB it's taking forever to run vis portalflow. There is about 10 times as many portals being made by vbsp. Beta3: 2 threads reading e:\games\empiresmaps\emp_uaxactun_b3.bsp reading e:\games\empiresmaps\emp_uaxactun_b3.prt 182 portalclusters 648 numportals 0...1...2...3...4...5...6...7...8...9...100...1...2...3...4...5...6...7...8...9...10Optimized: 32 visible clusters (0.00%) Total clusters visible: 29630 Average clusters visible: 162 Building PAS... Average clusters audible: 182 visdatasize:9784 compressed from 8736 writing e:\games\empiresmaps\emp_uaxactun_b3.bsp 19 seconds elapsed OrangeBox Beta4: 2 threads reading e:\games\empiresmaps\emp_uaxactun_obb4.bsp reading e:\games\empiresmaps\emp_uaxactun_obb4.prt 1351 portalclusters 2615 numportals 0...1...2...3...4...5...6...7...8...9...100...1...2...3...4...5...6...7...8...9...10Optimized: 132 visible clusters (0.00%) Total clusters visible: 1132101 Average clusters visible: 837 Building PAS... Average clusters audible: 1342 visdatasize:439650 compressed from 475552 writing e:\games\empiresmaps\emp_uaxactun_obb4.bsp 10 minutes, 31 seconds elapsed Visible Clusters has gone from just under 30k to over 1 million!!! I only got this one to compile through a few giant func_viscluster entities. What's different in OB that's causing this? Are hint tools being ignored or what? Sorry I should really go read up more on whats new in OB, I kinda jumped in the deep end without looking at whats been changed with OB... MEh, it's late, I'm probably not seeing something properly here...
Nothing has changed as far as how tool textures are handled. Hints & skips work just fine in OB, and most current Empires maps have them (all of mine & Keef's do). I'd have to take a look at your VMF to see what's going on. You can PM me or catch me on Steam if you're interested. One thing you might double-check is all of your tool textures in your map. You may have noticed in the past how if you type "hint" into the texture browser, two different hint textures come up. They're actually the same texture, but one is located in /tools/ and the other is located in /common/tools/. For Empires, you have to make sure you're using the ones located in /common/tools/, because a long time ago the Empires devs did something different (I can't remember the details of the situation off the top of my head). For any other game/mod, you'd just use the ones from the /tools/ directory. If you're using the ones from /common/tools/ already, then I'd imagine you were using them in the pre-OB version of your map, and in that case I'm not sure what could be happening, and would need to see your VMF most likely.
It seems I'm already using the common/tools ones unfortunately.here's the VMF along with the portal file and the b3 ones as well for you to compare with. Notice the b3 .prt file is approx 4KB whilst the b4 one is 846KB. Crazy. Anyway if you can see something I don't I'd really appreciate the help. thanks Same goes for anyone else who'd like to DL this and take a look. cheers http://www.megaupload.com/?d=NZ8EWMQ4 PASSWORD = riot EDIT: I've discovered that running the map WITHOUT hints or func_viscluster triggers my .prt is 1/4 the size it was. So the hints are in fact increasing the number of visleafs....I'm going to try remaking them and see if that helps
pass=riot update: I removed all the hint brushes and compiled with only func_visclusters and it worked. still about 5 times as many visleafs as there should be but it compiled fully, no crashes and water is working properly with some nice reflections. I may make this version the next playtest beta and I can work on reducing visleafs in the meantime.
tl;dr after you compile the first bit of vis, you can open the portal file under the map tab. it'll show you how the vis is being cut up. If you have stairs or somthing like that that aren't func_detail ed, then it'll fuck up your vis.
everything that should be already has been func_detail ed. Ive taken a look at the portal file inside hammer. its crazy. it seems like it's cutting every single block into a vis leaf (I think the default is every 1024 square units) this never happened before and I've no idea why it's suddenly doing it now. It's not cutting up visleafs according to any bad geometry, its simply cutting the whole level into small cubes. download the VMF and take a look for yourself. The old and new portal files are there too. Like I said the hint brushes seem to be screwing things up now since I loaded it in the OB config
Alright, here's what's up. with empires, you need a tall skybox for the commander to view stuff. What you do is take the area that you aren't using for gameplay in the skybox, and you create a viscluster. also, I peaked at your vmf, you need to fill the map under the displacements with nodraws. not a huge issue, but there are a shit ton of vis leafs under the map being cut up by the nodraws that don't go all the way to the bottom. (this is b3 btw, not ob one)
so, I added 1 viscluster to the top of the map just above where your hint/skip is on the top, and then 1 nodraw brush to the bottom of the map that lines up with the bottom of your water. Valve Software - vvis.exe (May 19 2009) fastvis = true 4 threads reading d:\valve map files\vmf's\emp_uaxactun_b4.bsp reading d:\valve map files\vmf's\emp_uaxactun_b4.prt 1141 portalclusters 3350 numportals BasePortalVis: 0...1...2...3...4...5...6...7...8...9...10 (4) Optimized: 378 visible clusters (0.00%) Total clusters visible: 1177493 Average clusters visible: 1031 Building PAS... Average clusters audible: 1139 visdatasize32075 compressed from 328608 writing d:\valve map files\vmf's\emp_uaxactun_b4.bsp 4 seconds elapsed ** Executing... ** Command: Copy File *edit* Other thing you can do is reduce your fog to around 12000 units (the farthest visible distance for buildings iirc) and then leave the base of your map in your 3d skybox. This will create the effect of the whole map being visible, even when it's out of render distance, and it allows you to extend your fog a bit, but atm the map, I think, will stop rendering well before it hits the 15000 units for fog end distance in the env_fog_controller. **EDIT** Valve Software - vvis.exe (May 19 2009) fastvis = true 4 threads reading d:\valve map files\vmf's\emp_uaxactun_b4.bsp reading d:\valve map files\vmf's\emp_uaxactun_b4.prt 988 portalclusters 2903 numportals BasePortalVis: 0...1...2...3...4...5...6...7...8...9...10 (3) Optimized: 385 visible clusters (0.00%) Total clusters visible: 895319 Average clusters visible: 906 Building PAS... Average clusters audible: 986 visdatasize:250572 compressed from 252928 writing d:\valve map files\vmf's\emp_uaxactun_b4.bsp 3 seconds elapsed ** Executing... ** Command: Copy File ** Parameters: "D:\Valve Map Files\VMF's\emp_uaxactun_b4.bsp" "D:\Valve Map Files\Maps\emp_uaxactun_b4.bsp" that's what happened when I removed your hints and skips. Hammer automatically divides up the vis leafs based on the 1024x1024 grid lines, and yours were not on those grid lines, making excessive cutting up of the grid lines. This is simply by removing the hint/skips that are vertical and then adding the viscluster to the top of the map and adding a nodraw under the map that lines up with the bottom of the water entities. You could probably do a ton better if you better filled in the map's void under the displacements.
hey, thanks for taking a look and the input. yeh I've already added a load more viscluster brushes under and on top of the map to speed up compile time. Ive done a bit more since I posted the vmf. added in visclusters in place of the hint brushes and it speeds things up quite a lot. It still doesnt explain why the hint brushes arent working when they were working perfectly well before. Also you're still getting over 1 million visleafs with your method there and youre only running fast vis. By adding visclusters I've reduced that to about 200k but its still 5 times as much as it used to be. I'll post the new vmf and see what you make of it. It looks like I'll just have to re optimize this whole thing from scratch with new hints better placed on the grid lines eh also the extra fog distance I allowed was because the skybox was actually getting fogged out too early and you could hardly see it except for outlines most of the time. I think it looks ok now. I was previously ocmpiling with fog of 8000-12000. Not really an issue. new VMF is here. pass = riot http://www.megaupload.com/?d=1DH7V5Z5
the second compile was under 1 million. It was nearly half of what I was the first compile . also, placing hints on the grid lines is pointless, as the vis leafs are already broken up on the grid lines by default. The only thing you can really do to improve there is to add vis leaf caps in areas that are not visible to the other areas of the map. At that point, it won't help the map compile faster, but it will help the in game fps. Total vis leafs isn't really the issue, it's properly placing the vis leafs so that the performance in game is optimal. certain things can't be helped, like the fact that the tall sky for the commander will eat the hell out of your vis. If it's compile time that bothers you, hop on #empiresmod on gamesurge on IRC and there will probably be someone with a high powered CPU that would be willing to run it for ya.
It's not so much the fact that it's taking so long, it's trying to figure out why it takes so long since I switched to OB. It used to compile in under 10mins full. And the number of visleafs its creating has been blown out of proprtion. I'm jsut trying to track down the "why" of the problem. Right now I have a workable compile that runs with a decent framerate of 50-80fps on my machine. It's jsut really frustrating that a problem has arisen where there was none previously. Anyways I appreciate the help. Did you have a look at the newer vmf? Its a bit more optimized with visclusters. I'll probably remove the 4 blocks in the sky to just one big viscluster like it was originally for the comm view. Here's what that last vmf spits out 2 threads reading e:\games\empiresmaps\emp_uaxactun_obb4.bsp reading e:\games\empiresmaps\emp_uaxactun_obb4.prt 404 portalclusters 2111 numportals 0...1...2...3...4...5...6...7...8...9...100...1...2...3...4...5...6...7...8...9...10Optimized: 4 visible clusters (0.00%) Total clusters visible: 147679 Average clusters visible: 365 Building PAS... Average clusters audible: 402 visdatasize:44237 compressed from 45248 writing e:\games\empiresmaps\emp_uaxactun_obb4.bsp 20 minutes, 33 seconds elapsed
