BlackThief

yes it's possible and it has been done before - ask oddity for more information about the tools he used to convert everything.

yep that looks a lot better - though I'd still make the wood darker. here's an example what I mean:

don't know if it's the case in sparhawk's area, but here in this area quite a lot open air baths are run by the local city council or whatever it's called in english. you still have to pay of course, but they're not a privat company.

looking good, however I'd dirty the roof up quite a bit. it looks almost brand new at the moment

doesn't look bad to me - it's quite normal that the normalmap is shaded like that. just test the model ingame now and use a completely grey tga as a diffusemap - that's the best way to test if the normalmaps looks good.

that doesn't matter at all - I think he used "textures" here because he was generating a normalmap for a brush texture and not for a model. but this doesn't matter, as it's just the path of the **_local.tga and the name of the material. I think it would make things easier if you could upload your lowpoly and highpoly + your mtr-file here so that I can take a look at them myself. otherwise I'd have to guess what's wrong

1)you have to place the well.mtr into doom3/darkmod/materials 2) no because when you launch the doom3 editor your shortcut should look like this: D:\Games\Doom3\doom3.exe +disconnect +set r_gamma 1 +set r_brightness 1 +set r_mode 6+set r_fullscreen 0 +set r_multiSamples 0 +set fs_game DarkMod +editor because of the "+set fs_game DarkMod" doom3 is already "working" in the directory "darkmod" and thus you don't have to write "darkmod/models.."

1) yes - your lowpoly model needs to be uvmapped and needs to have a material name applied to it. I've got a seperate renderbump folder in my doom3/darkmod/models directory, there I place all lowpoly and highpoly models for the renderbump process. then you have to write a material shader - that looks like this: models/renderbump/bed { renderbump -size 1024 1024 models/renderbump/bed_local.tga models/renderbump/bed_high.lwo } models/renderbump/bed is the name of the material, and thus the material name of your lowpoly object. 2)the actual name of your lowpoly object isn't important at all - you specify that later when you enter the renderbump command into the console 3) you have to specify the name and location of the highpoly model in your material shader "models/renderbump/bed_high.lwo" oh and btw don't forget to triangulate both model before renderbump - otherwise d3 will either crash or you won't get a normalmap. the command is pretty simple: renderbump models/yourfolder/yourmodel.ase or yourmodel.lwo since I'm working with lightwave anyway I prefer lwo and I think it'd be best to keep everything in lwo - of course you can work with ASE since it's max native format, but I think it would be better to convert them to lwo later so that we have uniform file formats. hope that helped

1) actually I think it would be most efficient to use a tileable brick texture for the whole lower part - and I think you could do that brick texture a little more rough and messy than on the highpoly model from atti (which is not bad, but would look better with not so straight and clean bricks) I also think you could go with a tileable texture for the roof - this way we don't have to go so highres with the actual texture resolution and we're probably able to reuse the texture for some other stuff 2) it depends on wether you're going to mirror stuff - when you mirror stuff it's better to map the whole thing first and then generate the normalmap, otherwise you'll notice a seam on the normalmap ingame (a little seam will always remain though). in others cases it really depends - there's no rule you'll have to try what looks best in this case. 3) unfortunately I can't really help you with that, since I'm not using max, but is it possible that you can somehow set the size of the uvmap in max and maybe this is set on 256x512 or something similar? 4) you don't have to make a new thread for each object - it's enough if you create one for each kind of category or so - like Pinkdot's light objects or Pinkdot's architecture props etc - just so that we can keep track of, who's doing what and later it'll be easier to find an old WIP thread

try this: http://208.49.149.118/TheDarkMod/Files/DeepEx.rar

well then you're living in the wrong area - the atmosphere is really great here. I was at the Schlossplatz in Stuttgart during the match against Costa Rica and it was really awesome - almost 40000 fans. the whole city was one WM party. tonight people hooted (?) for more than half an hour on the streets - you can see a flag on almost every house here.

you didn't see the match brazil vs croatia, did you? It's really a shame that the Croats didn't win the game, they were more than twice as good as those fat football millionaires, who think they don't have to run anymore, just because a single Brazilian player has got more money than the whole Croatian team. looking forward to the match tomorrow though - I'm curious, if they were able to improve german defense - if that issue is fixed, they'll definitely have a great chance to win the tournament!

1) you definitely need to get deep exploration - it's a nice little prog, that can view 3d models of almost every format and also convert them. eventhough Doom3 is able to import ASE we stick to lwo (lightwave) 2)this is true, however heightmaps used by id were actually just simple greyscale bumpmaps for fine details on the characters - you don't have to create a heightmap anymore, since we have the nvidia photoshop plugin for normalmaps nowadays, that does exactly the same. work flow: create highpoly object (go as high as you want/your computer is able to handle) create lowpoly object (just like in any other engine - stay as low as you can, as high as you need - if it's a pretty generic prop you probably shouldn't go over 1000-1500 tris, but if you have to and it's an object with a lot of round stuff it's not too bad to go over 1500 - characters in TDM have around 3000-5000 tris) uvmap lowpoly object run doom3 renderbump --> here you get the best normalmaps for d3 in tga format - here's a short explanation by id software create a highdetailed heightmap in photoshop and convert it into a normalmap using the nvidia plugin merge both normal maps in photoshop - not sure what the layer-setting is called in the english version - I've got the german PS-version and it's the 5th layer-option from top of the scroll down menu - called "Ineinanderkopieren" = copy together if you have any further questions - just ask me it's a lot more complicated and more work to do, when working with a new engine I know....but the result is worth it

hi pinkdot welcome aboard, sorry to let you stand out there in the rain without a task, but I couldn't spend much time on the computer lately - there's probably only one soccer worldcup taking place in your country during your life you know the waterpump is ok for a start, but actually it's not a very important model, and I just noticed while checking the ASE file that it's quite highpoly for a simple prop (more than 1000tris) and the ASE file is missing uv-data so you would have to unwrap it yourself. so if you like to i'd suggest to start with a city theme of models - the city-well, streetlamps, electronic stuff etc, just stuff like you find all around the city mission in the old thief games (everything needs to be modelled except thewell). I'll post some references for your here... streetlamp - concept by Springheel - inspiration for detailed top - other lamppost - yet another one walllantern - concept of a victorian lamp more will follow...

yes but those stroggs have actually quite usefull tools for a thief: http://doom3.planet-multiplayer.de/images/...2/etqwe34_g.jpg ..and there's SNOW http://doom3.planet-multiplayer.de/images/...2/etqwe35_g.jpg

yes they all have a head-bone, so head-turning should work.

yes splashdamage confirmed that they'll release an sdk with quake wars. (and I guess they'll be quite modder-friendly, since the whole team consists of ex-quake2-3 modders)

certainly not better than quake war's modified d3 engine that makes use of the megatexture technology. quake4's engine is actually pretty much the same as d3 except some improvements in the multiplayer code and I think I've heard that they improved ambient light as well, but I'm not sure on that one. @spar - it's quite common that parachutes have holes in them http://www.foreverflying.com/parachute-striped2.jpg

HL2 isn't really THAT good for outdoor scenes, especially for nature scenes - they don't look any better than unreal2 outdoor stuff. I guess you're not aware of Quake Wars - it's using a modified D3 engine as well and is able to render huge outdoor scenes a lot better than most other engines. compare - valves showcase level: http://www.neuro-tech.net/images/lc5.jpg quake wars forest: http://doom3.planet-multiplayer.de/images/...shotsec_010.jpg http://doom3.planet-multiplayer.de/images/...shotsec_001.jpg

and that's why it has to be done by code and not by an animation otherwise we'd have to do 180 different head animations for every possible angle...actually head turning should be almost ready out of the box, since doom3 NPCs are already turning their heads slightly towards to player.

well it's not that complicated once you've got the right maya version + export dll. best way would be to read through this tutorial: http://people.freenet.de/darkmod/exportModelsD3.pdf and then get maya 6 or 7 (btw you don't really have to read all the shit - important stuff comes at the end, like writing the def file etc. ) edit - missed that: I guess that would be all, yes. Head turning and stuff while in search-mode should be done by the programmers.

I know what spar envisioned, but to implement everything exactly as described in his script ingame is way too ambitious - we'd have to make several specific animations for the guard and for the thief + a camera animation and implement that as a cutscene. Apart from the fact that no one here actually knows how to do an ingame cutscene it would require probably weeks to get the animations done properly. However we could do almost the same thing (with less fancy effects and dramatic scenes) ingame as soon as searching works. We can easily make an ingame vid of the player knocking over a chair or something and then we could do the "zoom to guard" using noclip, nohud and a completely blind guard.

don't think we need a specific animation for the vid - that would be too much effort for that - when the search anim is coded in it should work without any pre-rendered stuff imo.

dom, would be nice if you could do it for me. I'm busy with the factory map and thus I don't want to start learning maya animation now just to place some fingers ; )

the anim is ingame - there's only the hand issue left. I tried to pose the fingers in maya, but they flip back, when playing the animation.