Sorry to hear that STiFU! There's a great shortage of TDS FMs, so nearly anything would be welcome, but I can understand what you're saying.
To STiFU, or to anyone around this "Thief III Editors' Guild", which of your messages would have the very best general guidance or at least, some good "over-view directions" to prevent this sort of FM designing-problem that STiFU apparently ran into?
Hey, it's nice to see that you, T3eders don't give up and keep up hard work. Mine is suspended right now, due to latest big projects at work - we've just finished translation and testing of Oblivion (polish version), and we've already started Guild Wars (as you probably know, a huge project). But that doesn't mean I gave up, T3ed and 3dsmax have safe place on my HDD
Ok, and I made one "time-consuming" mistake, paid for my Eve-online account renewal :P
Last edited by Judith; 30th Jun 2006 at 11:44.
The first and most important step is to start with a good (or at least rough) plan of what you want the level to look like...Originally posted by Tiger:
Which of your messages would have the very best general guidance or at least, some good "over-view directions" to prevent this sort of FM designing-problem that STiFU apparently ran into?
Since I've never succeeded in accomplishing the first step, I have no idea what to do next. I guess I'm not any help to you on this. I'm probably the worst organized level designer here. I just like clicking the pretty buttons in the editor.
Oh dear, Iíve fallen at the first hurdle then. Iíve changed the direction of my level at least 3 times now. I keep creating large open spaces and now Iím trying to rewrite it in order to eliminate them. Guess its because I didnít want to go with any of the ideas Iíve had as a first mission and thought it would be better to start with an enclosed underground space.Originally Posted by nomad of the pacific
Oh well canít be helped and i think ive landed at a final idea now.
Same advice as always - know what you want and what you can get away with.
A large part of the problem is that TDS isn't as capable as you'd expect from a UE2 game, so you run into a lot of problems you didn't anticipate.
The insistance on Max5.1, and the lack of in-editor material creation and other random niggles present something of a big hurdle to creating FMs.
There are big things in the works though - just remember that FMs on old engines are difficult and time consuming, and the more up-to-date the engine, the more effort (and art assets) required, so ours naturally take longer when it's levels done by individuals rather than professional groups.
I think the most important thing is just to know the editor perfectly in all aspects, BEFORE you start to make a map-overlay of your FM. I just started to build my FM with an openair area full of detail. When I loaded the map ingame it ran with 2 to 3 fps. At first I thought "Ok that might be because I am running it from the editor!" and just worked on, but after some time I found what zoneportals are for... Far too late!! I tried to seperate the openairareas with gates and stuff, but that was no good solution... Besides it didn't work properly in most of the times. Damn BSP errors...
So I think a concrete plan for the FM you produce is not necesary, but might help you a bit. The most important thing is that you have to know what the editor and the engine are capable of and how all that stuff works. So don't learn by doing!!!
I feel your pain, STiFU! That's exactly what happened to me. I've started and abandoned more levels than I care to think about, but that's part of the learning curve, I guess. It's a shame that starting with large open spaces makes the best looking levels. It's a real art to make strings of small volumes look large and open. Maybe someday I'll figure out how to do that.
I am checking the map step by step, when there is a slow down, it's zoning time, reducing flame lights that hits way to many walls.
Even like that is not guaranteed that you might hit some major bug in the future.
Yeah, you're right! But the problem is mostly : Open Air areas OR many details + lights . In fact it is the lightning which makes Flesh look good in my oppinion (and the details of course too). So If I'd ever start another map, I'd probably have only a small openair area, so that you can go through multiple entrypoints into a building. Hmn but it sucks that you can add very much and nice detail to house-exteriours too and this would just not take place in my fm then.
But anyway, I guess I will give TDM a try, when It's out... The Radiant is much cooler anyway! Though the doom3-engine's exterior parts I've seen so far look bad too, but we'll see about that.
Martin Karne, may I ask about T3 and its fogging ability in its Level Properties?
(I've heard several UnReal mappers talking about simple fog-parameters giving them a room-sized "moving view", from its ability to dynamically isolate the player's viewing-area, for some fairly "wide open" geography.)
Last edited by firstname.lastname@example.org; 1st Jul 2006 at 09:13.
Not a chance, from what I have read and worked back in 1999, Unreal engine accepts about 20,000 polygons, and it seems that fog is not good for detail at distance, you can see it, at least from what I have read now in this fleshworks.
I am not new to mapping, just keeping it to the most probable candidate, so far I have made one (wow) Unreal I MP map, not very detailed or well done anyway, a demo map pack for HL 1, and still working on another, and two RtCW maps, so I'm not very present in public, but in private I do most of the testing and stuff, and if there is the inspiration and the capability I try to make one good map.
I don't know if I will be able to complete it, but I am trying.
I could make a demo that would prove your statement wrong When you use fog and zones (doubled in pairs) it reduces e.g. the number of static meshes displayed in player's view. It's not perfect, but it works, and enables you to make fairly large landscapes, at least in comparison to what we've seen in OM's.
So you have to split the area into multiple zones, and then the fog is taken into account when zones are being considered for visibility? Is that how you're doing it? If so, that's a very nice method, good work.
Then someone must tell them to correct that "typo".
I could try some fog, I don't know how it will come out.
Yep Crispy, it works that way. Let's say you shape 3 road segments surrounded by hills in 3dsmax (e.g. 2048 uu long), put some trees on the hills all the way, then divide those segments with 2 pairs of zones and add fog (in my FM the distance between them is about 16 or 32 uu). Some of the trees should disappear when you're at either "end" of this road.
I wish I had some more spare time, because my SM's still look very simple, it would be great to find a method to "texture paint" landscape. For now my "road and hills" have only one texture - grass..
That I might have to try - I know fog by itself doesn't reduce visability (it does in UE2), but maybe with superfluous portals it does...
Do I understand you wrong, or do you use veeeery big zoneportals to split your area into multiple parts? because it was said that big zoneportals would hit performance badly. More detail about that plz judith. (Perhaps you should even go further and write a tiny tutorial for the wiki. It is always handy to have such information for later FM-authors)
Hmm.. these zoneportals are indeed big, 640 x 1536 for example. I thought you might want to check it out yourselves, so I made a kind of a demo:
1) unpack the archive into your t3/content folder (it shouldn't overwrite any files)
2) run the editor
3) open map forest_demo.unr
4) hit the "play map" icon
Ok, head straight out the cave, and take a look around - I hope the performance is fine. If you want to check this concept turn right and go to the end of this road. You should see a torch, and blue feet. Stand there looking towards the west and start strafing south. You should notice that as you move, more trees appear in the gap between the hills.
Good news: it runs fine, which is surprising since my graphics card usually chokes on plantlife in this.
15fps is fine (on this computer, I say 10 an average and 8 an absolute, and rare, minimum)
Bad news: the fog's having no effect on the visibility - that's just your portals doing their thing quite normally. You also seem to have used two portals in each place, which is totally unnecessary and can (I don't think it is here, but it *can*) hurt performance more than it helps.
And yes large portals are allowed, but aren't recommended - as a little hint though, you can kind of cheat: use a BF textured brush to cover most of the space and put smaller portals in where you can see enough stuff for it to be worth it. The portal will draw the BSP and any meshes "visible" through/behind the BF brush. Mostly only of much use with large Smeshes, but it's still potentially useful to know. And if you have a smesh floor, you can step out of the world through the BF too, for the ultimate in visibility control (no portal, not even any world between A and B, just smeshes visible in both A and B, so they still get drawn inbetween!).
Ok, that's really important piece of advice for me, thank you So the fog just hides some visiblility "errors" which might have been noticed otherwise - I mean seeing through 2 BF textured substractive brushes, divided by zone portals. I'll get rid of those unnecessary portals, they don't seem to hurt performance on my GF 6600 GT, but other cards' users might suffer from it.
Oh and if you had a spare time and could "illustrate" your "cheat idea", I'd be grateful. I think I get it, but I'm not sure
I don't get it at all, so some further explanation would be great.
Ok I added a bigger area and removed the doubled zoneportals. I don't see much difference, but you might want to check it out. Use the ghost mode as the area is quite big and a bit repetitive.
I don't know about very big portals, but like I should have said, logic area occlusion thru the sum of added brushes and zone portals to hide the underlying area, in this case a small yard that was formed helped to hide some performance hit, since I have got a GF6200, I can feel the normal user worries about hitting a slideshow once in a while, so I am just preventing on the spot bad design.
By logic and experience thru all those years of amateur in house mapping (1996 Quake I [I've never published any of those and there were not many tutorials and my english was even more crappy] to the present) I have some degree of prediction of what is going to work for the quake series of engines, and since this heavily modified engine is slowing down with big open areas I feel like I am doing a beffed up quake I/HL I map.
PS I have recently started working on a DeusEx FM too, who knows if I can do some maps here and there.
I use backspace to get the developer panel, form there I can get FPS, several menus of useful and useless stuff, inventory and cheats.
How is the FPS and the milliseconds? As long one is big/medium and another is low/medium you should be having an OK performance hit, since even on the original maps my new card can crawl a bit, hey I missed the FX5200 slideshow anyway, I swear!!!