CTF-Sinnical

[Sinnical]

It was my first UT map, so i made it my first UT2k3 map as well. Whether or not you like this map i think it is safe to say it is significantly better than the first Sinnical.

This is a very small map, playercount is set at 4-6 now. Cap runs can go as fast as 15 seconds without speed combo if you meet no resistance, which is pretty short even for a 2v2 3v3 map. Shortest flag run is through some lava, whereas the slightly longer ones are much safer in that respect.

Map is at 1.5 scale width (i think). this combined with very limited mesh use in narrow areas means even in the smaller halls you can keep some good side stepping up. (As a side note, a lot of the meshes do not collide actors. Some that don't collide in this version probably should.) limited mesh use of course means it isnt as pretty as it could be.

i wanted to make it so a lot of drops could be done between the beams (specifically from the upper base entrance to the health below, behind the flag base, upper center room down to health keg, and getting supershield). Drops can be used to pull you out of a tight situation, or buy you a second to find some cover. the four lifts in the center room were designed to launch players up to the shock and rockets, which they do. :)

Used a lot of bsp in the map too, probably more than i should have, the map has about 2700 polys. since i didnt knock myself out making it look pretty, frame rates shouldn't be a problem. It plays better for me than pretty much every map that came with the game.

Thx to a kickass Blitz tutorial, bots should be able to get everything, including those pesky super shields. Except i think the keg they cannot, bots dont want to cross the lava, but do like to drop down from the center room and take the lava back to their base from there. they also dont seem to want the sniper rifle, but im pretty sure i can fix that. Take a minute to chill in the lava room . . . see if you can see the skeletons that bubble up out of the lava now and then (blitz ty again for finding that trigger tutorial :) )

im sure there was more i wanted to say, maybe ill think of something else later. anyway screens:

first one is the outside part of the map. pretty much the only thing that isnt quite beta ready. you can't see the sky zone in the pic, as i had to zoom out VERY far to take the screen. despite the maps tiny size, the tower it is in is huge.

[Sinnical]

a picture of the base room. this is the entrance that comes from the lava pit room.

[Sinnical]

another pic of the base room, this one taken from the ledge of a stained glass window.

[Freakish]

Originally posted by Sinnical
a picture of the base room. this is the entrance that comes from the lava pit room.

omg it's q3!!! :p

[Sinnical]

the lava room, lower of the two center rooms.

[Sinnical]

All these screens are already slightly out of date. The difference between these and the final version should be almost indiscernable except in the base room, where a lot of lights were added.

the beta map should be up by tomarrow afternoon. :)

[Sinnical]

Originally posted by Freakish
omg it's q3!!! :p

damnit.

[Mozi]

Nice detail work man kinda of a face world thing going on there ??? :evil:

Can u link me to BLitz tutorial ???

[Spiffness]

Can u link me to BLitz tutorial ???







http://www.leveldesigner.com/index.php?action=doc&ID=172

[Sinnical]

Map is here (http://www.unrealplayground.com/download.php?mapid=2496 ) .

Already done a little testing myself and found that in the red base room, the low ceilings above the side exits next to the flag dont seem to be lighting at all. Seems fine in blue base. tell me if you get this problem. Also i found its pretty easy for a FC to defend him/herself easily camping under the lifts so nobody can get in from above, only from the one entrance from the base room. Im thinking of making the lifts kill u if u get stuck under them, maybe even just the four center lifts. oppinions on this would be helpful as they will change gameplay a decent amount

thanks to all who can find the time to help me out :)

[Blitz]

First, thanks for the kind words on the tutorial....who says flattery gets you nowhere?

A couple of notes on the map. The layout seems pretty good; I especially like the Z axis, which appears well thought out. I would have liked a little more connectivity along the higher routes, but it depends on how many players you're targetting.

The lava needs some visual work...check into fluid surfaces, but pay attention to framerates, as they can be somewhat costly. on this subject, you may want to read up on antiportals, as the framerates seemed a bit low considering the size of the map.

There were a couple of places where it is possible to jump to a spot, but a translocator is required to return. One example would be the supershield area, and another would be a platform I noticed outside. Flag carriers and insta-gibbers can get stuck in these spots. For the supershield, I would get rid of the shaft that lets players drop down inside, and just keep the translocator "window." (Just today, I updated my tutorial with details on how to get bots to translocate through such openings.)

Overall, nice work!

[Sinnical]

thanks for you time blitz. i agree that the lava needs something else. ill look into fluid surfaces and such. i really do need to be careful though, as that is already the area where fps is the worst (since u can see from one flag to the other almost). as for antiportals, i tried to set them up, but perhaps i am not using them correctly, or am using them inefficently or something. i might try to redo it entirely to see if i missed a key area.

as for the supershield, i intended it so a fc could drop in and hammer jump out. the shield's benefit wouldnt be entirely lost and they wouldnt have to drop the flag. i really didnt take instagib into account at all tho. hopefully ill get some more feedback on this.

yeah u can get stuck on some of the beams and supports outside the tower. the ones behind each base (the slanted ones) i think i will make steep enough that you just slide off. as for the flat ones on the sides, i could something similar but im not sure if it would look right. still something to consider tho.

thanks again for your time :)

[|PesT|Lexivore]

Lets see where to begin. Hmm.

I have located your frame rate problem. 874 hardware brushes. Antiportals will do you no good, in fact adding them will only slow things down further. Antiportals only cut out static meshes that are behind them, they don't help with BSP. Add some distance fog to lower the brushes that can be seen, that will take some of the load off of the processor. Rememebr BSP very processor intensive.

Also some things like the rocketsmoke generators, they each are set to 150 particals, multiply that by the four you have going and you have 600 particals being tracked at all times. A few less may releive a little of the pressure on the processor.

I was looking across the map and getting a frame rate of about 28. In a map that most everything is hidden by BSP that seems be a very low frame rate.

As for te lava. There are some in fire_engine textures pack. Nice wet looking lava.

I will take a more specific look at your BSP tomorrow. That many hardware brushes seems a bit excessive.

Also for static meshes rather than mess with antiportals give them a culldistance. With that and a matching distance fog the frame rate should come up nicely.

[Sinnical]

hehe
me likes bsp what can i say
there inlies my misunderstanding of anti portals. good to know.
i think i can cut down the amount of brushes. not sure how much i can cut them down but it will be an improvement.

ill cut down on the rocket smoke as well.

ty for the tips dude

[Blitz]

Originally posted by |PesT|Lexivore
Antiportals will do you no good, in fact adding them will only slow things down further. Antiportals only cut out static meshes that are behind them, they don't help with BSP. This is completely untrue. Antiportals do in fact occlude BSP, as long as the entire BSP node is behind the antiportal. The renderer will sometimes render nodes (meshes included) in spite of the fact that they are occluded by APV's, which might be what has led you to this erroneous conclusion. A simple test map showed me that I was able to occlude a bsp cube behind an antiportal. Use the fixedvisibility command at the console to see what geometetry is being occluded while walking your map.

[|PesT|Lexivore]

Hehe. I got to stop typing at two in the morning. Okay. Lets see here. Yes Blitz is correct antiportals can occlude BSP. What I meant but failed miserably to say is that in your map they are of no help at all. Let me explain this.

For an antiportal to work it must be touching some part of the map you can see (but not sticking out into it or it creates a HOM) and the rest can be embedded in bsp or outside of the map in general. Now what happens is that antiportal casues anything completely hidden by it to not be redered. This is a very useful thing when dealing with terrain maps becuase they do not occlude anything.

The reason they will not hepl with your map is because antiportals are nothing more than an unredered bsp object, they simply block your view. Your abundant bsp already does that. Which means your meshes and hidden bsp is occluded already. Thus adding an antiportal won't help because it will be trying to do what your bsp is already doing. Which is occlude geometry that is 100% behind it.

I noticed the antiportals that are in there are completely embedded in the bsp whichs causes them to be occluded by the bsp negating any effect they may have had.

In my expereince it is a waste of time working with antiportals in a bsp based structural map (anything but a terrain map) since the bsp already does the occluding. Adding antiportals does not occlude anything more or anything less. Using the fixedvisibility command I was able to confirm that.

[Blitz]

Originally posted by |PesT|Lexivore
For an antiportal to work it must be touching some part of the map you can see (but not sticking out into it or it creates a HOM) and the rest can be embedded in bsp or outside of the map in general. Actually, you can have it sticking out into your map. Your best bet is to align it one pixel in front of your wall. As long as it does not occlude the entire wall brush, and as long as a player cannot stand inside the APV, there is no HOM effect.
The reason they will not hepl with your map is because antiportals are nothing more than an unredered bsp object, they simply block your view. Your abundant bsp already does that. Which means your meshes and hidden bsp is occluded already. Thus adding an antiportal won't help because it will be trying to do what your bsp is already doing. Which is occlude geometry that is 100% behind it. This is incorrect. In 2k3 BSP does NOT occlude (as it does in UT).
In my expereince it is a waste of time working with antiportals in a bsp based structural map (anything but a terrain map) since the bsp already does the occluding. Adding antiportals does not occlude anything more or anything less. Using the fixedvisibility command I was able to confirm that. I have attached an example map with two rooms that are joined by a small corridor in the center. The left wall is occluded with an APV, the right wall is not. There is a mesh and a BSP cube behind each one. To see the behavior I'm talking about, spawn in the map and type fixedvisibility at the console. Then walk into the far room and you'll see a cube and a mesh on the right, but not on the left. I hope this clears things up.

[|PesT|Lexivore]

Hey Blitz that map is perfect.

Stand to the right and type in fixedvisibility. Walk around and the bsp wall has in fact occluded the mesh and bsp box. Stand in the middle and they are too close to your line of sight and do not get occluded.

Makes for an interesting case study though. In the case of the antiportal the collision is also gone, in the case of the bsp wall the objects are gone but the collision remains. Which makes the antiportal a more efficient form of occlusion.

But back to Sinnical's tower. There are no walls that provide an effective blocking of view like the two in the example. Which is the frame rate problem since just about everything in the internal part of the tower can been seen at once. I still don't think adding antiportals is going to solve the problem.

[Blitz]

Originally posted by |PesT|Lexivore
But back to Sinnical's tower.... Yes we got a bit side-tracked. In addition to cutting down on the BSP (as |PesT|Lexivore originally suggested), you could think about changing the layout to better zone off the map. Also, when considering where to place antiportals, don't forget about the floors and ceilings.

To be perfectly honest, though I don't think the framerates are a major problem...just a bit lower than you might expect.

As for the example map, perhaps we ought to take that discussion to another thread. ;)

[Sinnical]

i redid the antiportals, the newest version is building as we speak so i havent tried it yet, but i think its occluding a lot more than the beta did, so thats a good sign.

i cut down all the emitters to 100 particles and it still looks pretty good, so im glad that was brought to my attention.

as for other changes i changed the sounds on the lifts to be a bit more dramatic. planning on adding emitters in the skyzone (if they work in sky zones), particularly flame emitters on top of the towers in the distance, which was much like the UT version had. havent cut down on brushes yet but probably will tonight. im thinking something else should go in the rocky walled lift areas too, as they look a bit plain.

thanks again dudes for the help :cool:

EDIT: Well fps did improve, but not significantly. some of the lowest areas would get as low as 28-31 for me if i stood still there long enough. i managed to get these areas up to about 37.

[Sinnical]

as im reducing the number of brushes im getting a ton of bsp holes. which leads me to question:
does it wreck the brush if you use deintersecting to make a 6-8 sided brush, because thats what i did and got half the wall in the center room missing now. its not that im intersecting so the new points are not on the grid, because they are, but just seems to be making things worse.

[|PesT|Lexivore]

I am now looking at some o fthe BSP of your map. I would recommend that you take certain poarts that are several parts and makes them one static mesh, replace the mutiple parts with the static mesh, turn off the collision and use a blocking volume. Then you will be cutting down on the bsp, an da good deal of collision that needs to be calculated.
For example this picture shows a bit that could use what I just pointed out and simplify that part a lot. Not tomention since that is used at the entry to both bases, it will be only a small poartion of what the processor has to deal with right now.

[|PesT|Lexivore]

One more thing. Build any static meshes in a seperate file so that there is no chance of any BSP errors in the mesh.

Rememebr that meshes are much easier on the processor and since this bit of geometry is used twice, the second instance will cost an almost negligable amount of processor power. And Meshes are not BSP dependeant and will not cause BSP holes. Wonderful.

Upon further review here is another thought. You use a lot of slats and they could all be the same plank type of static mesh and then retextured. It would make the BSP much freindlier and speed things up. :)

The more I look the more I see things that could be converted to Static Meshes to save on processes. Anything complex that is used more than once could be converted to a static mesh and save processor power. Since I have little else to do tonight I may do this just to see how much of a performance boost it may provide the map.

[Cobra]

Great map. CTF-[Sinnical] but a mineur thing. Sometime the bots can't go up. And when they do keep circling and falling down. look at the pics. great job, great map. Lot's of fun in there.

[Cobra]

2 pic.

[Cobra]

I have open ED, there is the cause of the first one. Never put a pathnode there. Look at the red dot. There is the place to put them. So you whon't ever have a bots problem. And where are the X, erase those pathnode, they are no good.

[Cobra]

And why he was keeping faling down. Where the red dots are put your pathing there. And where are the X erase them. Hope it help.

[|PesT|Lexivore]

Cobra some good advice but sadly that is the UT map, he has another one he is creating for UT2003.
hhtp://www.unreaplayground.com/download.php?mapid=2496

[Blitz]

lol...Right map, wrong engine. Though I'm sure you are correct as usual, Cobra. ;) You can get the 2k3 version of the map here (http://www.unrealplayground.com/download.php?mapid=2496).

Ah, WidowMaker beat me to it. :D

[Cobra]

Originally posted by |PesT|Lexivore
Cobra some good advice but sadly that is the UT map, he has another one he is creating for UT2003.
hhtp://www.unreaplayground.com/download.php?mapid=2496

Thanks I didnt no that. But since a love the UT version. I thought that if he whant's to change it so the bots will perform perfect. But thanks PesT Lexivore. I will check they other one. Man I love that map. Super fun in it. Thanks a lot Blitz. It doens't beat CTF-Warzone, but I realy like that map.

[Cobra]

WOW!!! SUPER Sinnical. Man A realy love that map. But the bots didn't try to get the key of healt. Know I have fix that. And some minor bots pathing. Super map Sinnical YES!!!!.

[|PesT|Lexivore]

Okay Sinnical here is what I am doing for you at the moment. I am rebuilding the major ofenders brushes (those that have lots of brushes but would be better as a single mesh). When I am done I will post the map with the static meshes. At the moment I have cut the map down to 750 brushes with many more to go as I am moving towards the center area. I am hoping to bring the brush count down to 500 or so. That should get the frame rate back up and take care of any potential BSP problems for future revisions. And the best part is it will look the same.

You will learn the way of the mesh young padawan.

[Cobra]

Find stuking bots. Only happend once. I have arange all the pathing. know they take all route. Take the key of healt, and so on. If you whant it just tell me.

[Cobra]

:)

[|PesT|Lexivore]

http://www.unrealplayground.com/download.php?mapid=2526

There is the modified version of your map. I have taken out something in the range of 350 brushes or so. Included is an external static mesh package.

While I was doing this I did find the source of some BSP problems. In the center area you had eight arches that had a bad node on each of them. Replaced with a mesh now so that is one potential BSP problem solved. The total brush count is now down to about 550 or so.

Also note that there are still bits that could be converted to static meshes. I did most of the major ones, but a few minor ones still remain.

Let me know when you have this and I will yank it down.

[Sinnical]

arg screenies of the original. somebody hide those!
j/k

got the d/l.
wow thanks for the help widowmaker.
i was really hoping to avoid custom meshes to reduce download time during online play, but i will have to seriosly consider otherwise.

as for bots, i know they have a few problems, fixing them is in the works. i may yet take you up on the pathing help if im unable to figure it out tho :)

[Sinnical]

i gave the new static mesh version a run thru. i have to say i'd really prefer to keep the bsp, the engine doesnt seem to be able to get the lighting on the meshes the way i'd prefer it to be, i'd rather not have an extra file, and i dont think the fps is in such dire need that it will be a real problem. i managed to bring the base room up to a minimun of 40, and the center room up to about 37. And thats on my rig, which isnt very top of the line. thanks a ton for making the pack though, i appreacite it, and may end up using some of the meshes yet.