Baddcog

Well, I think everyone agrees that the Cradle is one of the best official Thief maps ever. And the mansion (without the load zones) was pretty damn sweet too. But also everyone else pretty much agrees, WHY? If you want to play it it's already made. Most authors want to create their own map. It can take a bit of time to make a map as it is, but recreating one takes longer (due to having to copy every detail precisely) and in the end it's not even your own. Of course there are legal issues which to this point we've pretty much avoided as we don't want all our hard work lost due to copy write infringement over a map, name, storyline... That still doesn't stop anyone from using that stuff, we (TDM Team) just can't support it. ------- Still, you should learn to DR, it's really not that hard. Following the A-Z tut, asking questions here. Within a few months you could have a nice small map of your own that we could all enjoy It may seem daunting at first, but once you've messed around with the editor for a few days it really is pretty simple. Small basic missions without any fancy scripting or custom voices, etc.. can be quite fun.

Biker will be able to give you good advice on that when he sees it, but map size really shouldn't effect performance at all if it is portalled well. Of course maps in later stages will be slower than earlier... ai thinking, more tris, shadows, etc... But size alone should really only effect load time. ------ Since you said 'I'm putting some finishing touches on my map so-far' I am going to guess you maybe sprung an internal leak. I just had one myself that took an hour to fix, basically a u shaped hallway. Each end of the U was leaking to the other under the floor. All the portals worked seems to work (they were green [aka showing up] and they would close. But the portals in the middle of the u would close off, the ones on the ends ...well anyway... Sometimes you spring a leak from one zone to another. If portals are sealed right they will still be green and work (close and turn red) but not always properly. It won't crash due to leak because it's all within the hull of the map. But it can diminish performance in areas because more of the map is open than you intended. ======= You said you deleted all the lights which really are the largest performance drain. Next would be AI probably. Did you add like 10 of them between testing periods? --- Obviously if performance was good, you changed some stuff and then it was bad, something happened along the way to effect it. The obvious guesses are >tons of lights, tons of ai, internal leaks...

the op say that that map just marks locations but images can be used instead. The map I'm working on is pretty small and really won't require an automap, but i think I'll use it anyway since Ob went through the trouble of figuring it out.

That mod looks pretty cool. ----- As said above the visportals are one of the main issues. Hammer on the other hand does it for you and you only have to do a manual placement once in awhile if you can improve performance. (I did quite a bit of TF2 mapping) They are as easy as dragging a brush to fill the gap, right clicking ortho and selecting : make visportal But the biggest challenge is mapping around them. -------- Biggest performance hit is shadows. Unlike Source (in which one dynamic light can completely tank FPS) Doom3 can handle them pretty well. Where they see a hit is if multiple lights cast shadows on the same tris. Typically up to 3 sees no real perf. hit. 4 starts to show and 6 pretty much tanks out FPS. But it's pretty easy to keep the number below 3 and still have good lighting. Of course moveable sources like torch guards and candles can increase the number anywhere so you need to leave a little buffer room for that. You can also make entities non shadow casting, etc...gotta pick and choose which ones won't be noticed, which will look horrible without shadows... -------- Static meshes work great for details. But we've only got a few, doorframes, some buildings... Would always be great to have more. We really need some good arches and whatnot but I've somewhat lost interest in it atm and have probably 3-4 arches sitting around unfinished. But we can change func_statics of brushes and patches to models in DR, or export them to models. While not quite as good as a modelling suite you can do some pretty decent work in DR to use for this stuff (or just save as prefabs to load at anytime, benefit of prefab is they can still be altered in DR by anyone at any time).

What can I say, it's been on my mind a lot lately.

Great Grayman, this has been on the back burner for years. But it needs someone who can understand it I don't think size of visportal should have any effect at all. If the sound_loss prop is working properly (to and from ai) then the author can use that to adjust for the size of a hole.

Alright, been working on the map a bit more. (new job at night, slowest liquor store in town, so I get to map and get payed for it ) Been working at home too though. Got the Pub in tonight, not finished but it's in and sealed.

exactly

It is a form of z fighting. Doom3 has some bug where at distances it has a hard time calculating the z factor (technical terms) Had the issue on my door models. The hinges and lockplate are alpha 'decals' basically, they were offset from the door face (not flush) but at a good distance they would disappear or z fight anyway (it was a pretty long way off). Something with the rounding errors or whatever. I think Sneaksie Dave fixed them (just made the offsets greater)and he knows what the issue is, haven't seen hi around for awhile though.

Hammer actually has a great decal set-up. Much like how ours works as far as applying them. But they have a setting of priority on decals. So you can layer 5 decals and number them 1-5. 1 being rendered with highest priority I think. And of course this is always on top of terrain. You can also set which brushes it is applied to. So say you have a doorway and want the decal to be splattered across the walls on either side of the door, you just select those brushes. If we did that in DR we'd have splatter across the door opening also. (so it takes two decals)

Moonbo, that's z fighting. You can increase the distance of the decal from the surface (set the grid to 0.125) and bump it once. Unfortunately up close you can see a line where the gap between decal and wall exist.

did you use caulk on a brush, then create a new brush but caulk is filtered?

Well, is it that sound is loss across portals (I don't think it does - that could make sound traveling straight down a hallway diminish just due to portals) or that having more of them routes the sound better. Which can in a lot of cases minimize heard sounds because it has to go around corners instead of straight through walls. So if you have a hallway with a hole in the wall to the outside it you get more sound loss if the sound travels down the hall, around a corner and out the window. If you removed the portal from the window the sound would go straight through the wall to the player. In many cases the sound going through the window could be 10-50 feet further away then if it went straight through a 2 foot wall. So naturally the sound diminishes over that distance. The real problem we have is the player being directly on the other side of a pane of glass. While the guard may not hear you as well through the sound location portal, the player still hears the guard as if the glass wasn't there. Making it harder to distinguish how well the guard actually can hear you. While you might be able to walk and not be heard it makes you feel you need to crouch or even creep. Placing two portals in a window (one aligned with each side of wall) might gain you a couple inches of sound travelling distance and it won't chop up the terrain into more tris it still seems like a bit of over kill for very little gain. adding a couple more portals won't redirect the sound anymore, but it'll give you quite a few more tris.

Welcome. Very nice shots there, looks like you've got a good handle on space, textures, etc..

Looks pretty sweet Is it just me or does anyone else think the square wood trim in that last last has some really funked up normals?

It really is one of the last mapping issues imo. Even if the propogation to ai works properly (which is a good thing) the sound to the player breaks their perception that it works properly. (If I can hear them they can hear me, and where is that guard actually at?)

I wouldn't say that they play it as unintended. Obviously BJing and killing was a game design choice and a possible way to play. It's more like a game flaw that you can clear the map of ai and have a free for all. Most games don't have this issue. Say Halflife, very linear game with spawning ai and one way points. Once you reach a certain point you can't turn back and revisit an 'empty map'. They are always throwing you into a new area with newly spawning ai so the action never ends. We could have spawning ai but how many guards would be in any given map? It could get weird. Mappers could use time limits which makes the player rush and even if the bj everyone they won't have time to run around an empty map. But players hate time limits and they do like to run around empty maps. I used to bj everyone, now I find it much better to only bj a few if I have to. So it might be a game flaw, it might be unintended. But it really does suit a large range of players. If they aren't limited by Objectives (no kill, no bj, etc...) they can play TDM in any fashion that pleases them. That's what makes this game so great, it's got a high level and sandbox action.

I loaded up some arrows in my map to play around. Honestly I think the idea of pulling the bow back and having it wander for 2-3 seconds before 'locking into place' is a great idea. It doesn't make aiming harder it just makes you not be able to reload 3 times in 3 seconds. You can fire arrows off in really quick succession. Of course guards need to work properly, but no matter what we do there will still be the issue that the player can get a lot of arrows off accurately and fast! Having a bit of warm up period is realistic (as realistic as a guard taking an arrow to the head and still charging at least) Other wise what other option do we have to fix semi-automatic bow? Make AI invincible?

Well, it depends how the portals leaving are done and how much action is going on in there. My suggestions would be : stretch the building heights. The yellow one on left looks great, but there's not much room between the bottom and top floors (is there an inside yet?). I always like really tall rooms in the bottom of large buildings and more regular sized rooms in higher floors. you can always squeeze the buildings a bit closer too. Maybe a horse/cart stall in there, a water fountain and some benches. And those window textures, you can use the bottom 2/3rds to make the windows really tall (looks better then super wide) a cheesy proposal:

I know unalert headshots were supposed to be insta kill while alert ones were not. Though a head shot should still be a kill realistically. At the very least it should stop them in their tracks. And it should make them scream. If we can do that I think that it's the best choice. It effectively turns them from an immediate charging threat to an alarm which alerts other nearby AI. They'd still be a threat for the time being. It also disallows the player from taking down 5 charging ai in a row. Might still be able to drop them, but they won't be completely out of the picture, would take more arrows to finish the jump. Also think non-reusable arrows on ai kill is probably a good idea. Broadheads are cheap and abundant. If you use them to take down an ai they should be spent.

Alert an ai, run into the dark. they don't loose you just because you gem goes dark. you have to hide when they can't see you.

Problem with ai changes on a map to map basis are the same they were with t2... Nobody expects it. You get used to how the game plays and it should always be this way. There are things mappers can do to make their maps more challenging without changing core gameplay. ------- I think people have grown used to playing on expert because they like the higher loot, all objectives, etc.. typically that's how difficulty levels change. (and no kill, no ko, etc..) There's also ego... "I've played for too long NOT to play on expert', but I don't like forced ghost so I avoid it and have to drop down to an easier diff... The author can add more ai, more lights, lights that can't be put out, fewer tools, blocked routes, etc... it always sucked playing a map in T2 where a guard is lighting quick and gets you by surprise and you can't get away. TDM ai CAN see you in the dark if they are alerted, but were purposely programmed to be able to loose sight of you. Run around a corner, find somewhere to hole up, if they can't see you then they stop and search. It would suck if this mechanism didn't exist. Then you'd HAVE to fight them. And if you're playing on expert there's a good chance you'll run around the corner into another AI anyway. ------ IMO we need to encourage good mapping practices to make difficulties harder rather than encouraging changing core game play. this also goes with objectives and making them more fun for the player (not the author). forcing ghost or no kills might be how YOU like to play but what about everyone else. An optional objective gives the best of both worlds, it lets you know that there is a challenge to try, and it lets you know if you fail, but it doesn't force you to play in the authors style. Forced ghosting makes a map harder, it also makes reloads more frequent (and or forced). it also makes players play on easier settings (and miss other difficulty settings/objectives) because they are avoiding playing a forced ghost. --Yeah, that map was fairly difficult to sneak around in. Ambient doesn't effect AI does it? I thought it was soley so the player could see, no effect on light gem.

depending on what it is, might be best to have the player step into a trigger anyway. A lot of times someone might start up the map but not be ready to play instantly. If something triggers they might miss it... If they have to actually move into a trigger then you know they are present.

As far as that list, it is a bit outdated. Guess whoever was doing it isn't anymore Actually devs/mappers spend more time devving and mapping than updating wiki pages which is a good thing.

thanks, yeah the street/canal will extend to the left past the guard, just trying to get the current area pretty well and finished before I move on (I always get to spread out and waste time jumping back and forth between areas. Hopefully a little more focus this time will help me finish and release )