Search the Community

I got these 2 errors today when the model preview window opened. These errors just kept repeating and I had to ctrl+alt+del to exit dr.

There is a female townsfolk model, she just isn't in-game yet.

I'd say it will give you the same results as making normal map with Nvidia plugin. Unless you're actually modelling hipoly model, there's no point in complicating your workflow. I would use the method, which mikebart uses for his textures - make several models of stones (you can put the photo in the background and trace the shapes), then build the whole wall out of them, give some rough plane as mortar where needed and make a low poly texture out of that.

In Blender you can use the displace tool to create the irregular rocks; I posted a link to the video demonstration in the Documentation forum. As others have mentioned, you would need to use this as a normal map if you want to use it in game; the full model with individually-modelled rocks would be far too hi-poly.

Yeah, Oddity would beat me to it. (edit man, 2 more people are faster than me) I agree with him, make a normal of a high res model and use that on a low res/optimized one. Here's a quick one I made from the first pic. in Max make a plane, increase the segments, use the displace modifier. Then make a normal and optimize your mesh.

Well I've now implemented Display Lists for model rendering, and it certainly does feel faster to me. Especially in the Bonehoard entrance where all of those rocks are -- there is no obvious slowdown even when all of the rocks are visible. I haven't actually done any benchmarks though.

Sounds like every point is being welded to every other point, so every vertex ends up at the same location (effectively making the model "disappear"). In Maya there's a "threshold" value in the weld tool's settings that enables you to only weld points that are X units apart. I would expect Lightwave to have something similar... find it and set it to 0.0001 or so and see if that helps.

I'm using 8.0. Unwelding everything allowed me to move it, thanks. But then when I select all the points and weld them, the whole model disappears? I don't know anything about welding so I'm not sure what it actually is supposed to do, but I assume not that.

Obviously I'm not an artist, but I can still hazard a guess: making an L-system is the easiest, making low-poly trees is easy, and making high-poly trees is the hardest. However, I don't think it's quite as simple as that. High-poly trees are (relatively) easy to make look realistic. Leaves are essentially 2d; everything else is already modeled in 3d. Obviously talent is essential to make it look realistic, but the process is about as straightforward as it can get. The only (real) limiting factor is the modeler's ability. Almost any modeler that can fully model a tree can make it look realistic; it's building it in the first place that's hard. Low-poly trees are incredibly hard to make look realistic. Generally speaking, poly counts are increasing slower than players' expectations of realism, particularly for models that are going to appear in large numbers on-screen, like trees. The artist is required to balance poly limits and still maximize realism. Almost any modeler can make a low-poly tree ("Look! I put a green sphere on a brown cylinder and it looks like a tree!"), but it takes a lot of talent to make such a tree look real. In fact, the difficulty can increase exponentially unless some serious concessions are made, such as limiting the realism or using lodding to make the process of balancing easier. As a programmer, I'm biased, but I think L-systems are the hardest to make realistic (at least for games). The first problem is that it puts it entirely under programmer control, and believe me, nobody wants programmer art. Second, it's even harder to make a low-iteration L-system look good rather than just having an artist do it (the advantage is that you can populate a forest with unique trees in a matter of seconds--or less). Third, it's hard to make L-systems that can run in real-time, so the main focus with this technique is generally making it easier on the artists. And finally, the entire implementation has to be lod-friendly, and with minimal loss in noticeable detail. How accurate was my assessment, oDDity? I'm probably missing something, but I have no idea what.

oDDity, I have to disagree that musical ambience in games is much different from a film score. I would say that viewing a good movie is also an activity for the audience (at the very least for their brains), although games obviously provide a greater level of involvement. In both cases, musical score a tool to enhance a mood or emotion, used to overcome the distance or separation created by each mediums interface. The same is done visually with lighting or positioning of the camera in movies, and sometimes in games, too. I'm not a modeler, but I would imagine that you use similar tools when you're setting up a scene for a model shot. I however fully agree that it should be up to the player if he or she wishes to have ambient music, and think a seperate slider for it should be implemented, like Springheel said. On the topic of how intrusive ambient music should be, I'll quote Brian Eno: I'll be the first to admit that I've not followed this rule with a lot of tracks I submitted here - that's mainly because I'm not good with melodies, so I constantly try and try again. And to get back on the original topic of this thread: I'm sorry for the loss of momentum, but recently work leaves me pretty much wrecked for the rest of the day. However, I have my tools laid out and got SVN for the sound folder working (thanks greebo) - next thing I'll do is to give comp-music's submissions a proper rehearsal and then commit them. I've also extracted Doom 3's sound files and briefly browsed through them, there are quite a few sounds we can make use of. I'll sort the good ones out and then complete the list of requested/required sounds. Yes. Of course, different sets for each would be a lot better, but there are other sounds that need to be done as well. We can not put in all the little extras until we have the core set of SFX done. You have to bake the cake first before you can put the cherries on top of it. Edit: This is very subjective, but for me the clash of the breakbeat and synths to the setting mirrored the clash of medieval zeitgeist and victorian/steampunk design and technology. Except if you should happen to have a different taste than oDDity, of course.

"so I can earn more for one model that you earn in a week. What do you do?" Presumtive there! slavatrumpevitch might be a top surgeon, or businessman!

I've already had that experience here Dom, and I wasn't even being paid for it. I'm currently modeling professionally, doing high res digital sculpts, so I can earn more for one model that you earn in a week. What do you do? You don't need people skills when you have talent, and in fact, if I could just quote the response from my employers about a recent horse sculpt I did - 'Overall, as before your work is stunning. I think it's safe to say that these are probably the best digital horses ever created' So you see, I don't need to tell them jokes, or become friends with them.

Maybe the polgyon normals are incorrect? I don't know how you'd fix that. In one of the games we made at my office recently, if there was an error in the model conversion process, the normals for the polygons would be messed up which caused shading problems.

I've noticed that our model website hasn't been updated recently. Specifically the new player equipment and bonehoard models aren't there yet. I know it's kind of tedious, but can modellers please upload a screenshot to that website when they upload new models to SVN? Just makes it easier to see what we have and where to find it. Here's the login info again: Login: www.tjoff.com/jens/darkmod/gallery/admin/login.php User: darkmod Pass: unclegay

I made a feather for the hats, and then noticed we didn't have an ink & quill model yet, so...presto.

I have a few of those castle molds, from Hirst Arts. My wife has been casting a zillion different peices from plaster of Paris, and sorting them all into little containers (bricks, windows, arch pieces, floor tiles, detail pieces, etc), but we haven't actually made anything yet. You can use stronger resins and other materials in the molds, but they're more expensive and take more expertise. We got a couple sets of basic bricks and floor tiles, plus some gothic dungeon/accessory sets, and an all in one set that has everything for a "wizard's tower" design. What we actually bought are the molds, green slabs of silicone junk with all the little bricks and pieces carved in relief (molded themselves from the guys master molds). We are planning to make some custom dungeons/castles used for D&D. They are extremely high quality and detailed, pretty much the best thing available apart from just sculpting your own models. It looks like the link above is someone who has molded all the pieces for you ahead of time, but you still do the rest yourself. It's a real hobby type thing, and you can add details from the local hobby shop where they have the model train section, stuff like moss, shrubbery, vines, etc.

Oh, I know what you mean. I don't want to go too far with magic but for minor injuries, for example, a magic balm that accelerates the healing process would be acceptable to me. An arrow in the lung is another thing altogether, as is a "shield". Location damage with realistically modeled damage effects. Arrow to the leg or arm = limited/no use of that limb dependent on the actual part hit. Head=dead. Chest=dead if heart/major blood vessels, the slow death of respiratory failure if hit in the lung. Abdomen=slow exsanguination. You could survive a lung/abdomen hit if you manage to escape quickly enough to get help. I'm all for stitching of minor injuries and setting your own bones +/- balms for accelerated healing/analgesia and potions for exhaustion. This is only loosely based on reality, but I'd extend the same model to the AI.

It's nice to hear that the system could actually e implemented in full with little work. However, any shot to the body or leg with an arrow should be final, I don't see the reasoning behind making only a head shot fatal. The body or leg shot would not be fatal, but it would be incapacitating, which is essentially the same thing as far as the game goes. The thing to remember is that the player is wearing no armour, and broadheads are designed to do a lot of damage to flesh. The bottom lin is that no health meter is required, you're either hit and incapacitated, or you take a minior wound, in hich case you bleed, and it's that which determines your survival. It could become quite complex of course, you could take into consideration the position of bones and major veins and arteries, and factor those into the damage model for the player avatar I would love to be able call up a detailed account of my injury with a hotkey, rather than just have some red line on the HUD go down by 25%.

Well I agree with Oddity for the most part. If you get hit by an arrow properly anywhere you are dead without immediate medical attention. However, in a world with magic one can always have healing potions, balms and charms. I think a realistic damage model in TDM should allow for a limb to be taken out by arrow or sword, but healable with magic. The possibility of bleeding to death doesn't really make sense from a gameplay because realistically speaking any injury that you will bleed to death from will be a serious one and you really wouldn't patch yourself up mid-mission, you'd get OUT and seek help. There are very few injuries that you will bleed to death from that are not that serious as to quickly incapacitate you.

The only new thing I learned from all these discussions is that it's possible to model a mind using independent clusters which compete with each other. I have for years wondered about the similarity between a collection of neurons and specialized clusters becoming concious and a collection of individuals and groups forming a society, which can perhaps be called concious (not yet certainly). But how will we know when a society is concious? Only by obvservation because we're capable of noticing because we're concious? Because neurons are not aware of the brain. The next question that I always thought of is how do we organize people in such a way as to promote a concious society? Do we increase the amount of interconnections, and grouping of people, and pre-defined filtration and routing of information?

What Oddiy always seem to forget it is, that even if we would model a very realistic combat or damage system, that you still lack one big part, namely your body. As long as computergames are controlled via keyboard and mouse, and no real feedback on whats happen, it will always stay a crutch. It's very different if you can use your whole body, with arms, feets and heads independently, while doing the same with only your fingers and no ffedback on the various parts supposedly involved.

oDDity if you have a system that complex with melee I doubt it would end up functionally different to a random chance, as with the confines of non-body awareness 2D FPS view you won't know what got hit or how. Random chance makes more sense from a work point of view, and most players couldn't tell the difference. Which is why STALKER's A-life isn't great, if it was randomly done in a simple way I couldn't tell. It might be in fact. And maybe other minor blows could give a temporary effect (COD2) then you recover, to simulate the initial sharp pain of bruising or soft tissue damage, which fades. (ignoring the dull ache after) Bleeding... meh. If you are bleeding bad enough to cause shock how much are you going to be able to do anyway? Survival isn't guarenteed. No sterilisation, tourniquets often lead to necrosis and amputation as most are done badly, no anaesthetic so stitching is going to be hard... You need morphine antiseptic sterilisation and blood transfusions for guarenteed survival. Not worth it. You'd fail the mission anyway through incapacitation or being slow or stationary for ages, or just die later from infection, better to avoid that side altogether. KISS. Don't bother simulating everything in massive detail for negligible effect on the player. With your logic it'd be better to model each footstep and clip each foot in FPS games...

Whats truly scary is the degree of mushiness in terms of superstitious beliefs amongst some of the most highly educated technically advanced people over here. I dont know how it is in Europe, but over here we have a significant portion of our scientific community talking about the need to respect faith as a form of knowledge, about how you can have both a scientific model of the world and still believe in hocus pocus shit. I used to work with a guy, degrees in astrophysics and (even more unbelievably) philosophy who was a Christian. We got along pretty well but were careful to avoid *that* particular conversation. He could brow beat our colleagues about his faith but he knew better than to fuck with me. To my thinking, this individual is little more than a technician, not a scientist or even a student of science. If you have any understanding of the scientific method, you can instantly see how it dissolves belief systems based on faith, let alone disintegrating the silly particular beliefs involved. How do these people get so far without these contradictions exploding in their faces?

In Doom 3 you don't need to have a visible body in order to see the player shadow. You can have either or, however you want it. However, we don't support body awareness in the sense that you wont see your body in first person. You can switch on the player shadow though, but that's not officially supported. Since we still have to provide a player model, for a variety of reasons, the shadow would still work, but it might not be correct in all cases. And we also can not make our AI detect that shadow. Once D3 is released under GPL, this might be feasable, though.

The major problem with believable AI is that they would need to have a realistic modelling of their responses and beheaviour. But to implement something like this, you would need a selflearning system, because with any predetermined behaviour, you will always run into limits. But self learning systems are much slower. After all, it took you at about eight to ten years to learn a comprehensive set of rules how to behave and form models that give you a sufficient variety, and not even then are you finished. A ten year old human is still not really considered to be fully developed though. Of course, since an AI is tailored for a specific task, it doesn't need all that knowledge completely, but if you try to model a realistic human, then you need much more knowledge what it is about to be a human, then just telling it how to perform guard duties. And that's your complaint. We can not give them proper knowledge, because we neither have the time nor the machine capacity. Heck, we can't even properly program models that don't have clipping problems in a hundred percent of cases, because this already exceeds the performance that we need to achieve. Much less complex decisions in unknown situations.