I do think that Bashiok's responses of "a lot of work", and "a massive undertaking" were a little defensive on his part. In my opinion Bstapylton brought up a a very good point, and random generation of body parts on creatures would add another element of uniqueness to every players experience while playing diablo 3.
I agree with everything you said. The question, however, is not if it would make the game more immersive for the players, the question was if the additional meshes and 3D-modeling, programming, and algorithms for this to become a reality would be feasible for low-end computers, disc space, speed and efficiency, the Diablo team's time restraints, etc. Also, for monsters that die in the amount of time it takes you to click the mouse button once or twice, is it really worth it? There would have to be sacrafices of other parts of the game that we might have enjoyed more.
Now I'm not saying that every creature should have this. For instance I think that the cultists should look all the same, because they're in an f'ing cult. But monsters that are supposed to spawn from body parts of mass buirials is another story completely.
If it had to be done, I would go with your view of it there.
Yeah, true enough. That's just the way I reason it and what I think would be best, and I'll probably stick to it even if they change their minds later. Like you said, they still have a lot of work to do right now, so this is probably a minor thing in the grand scheme.
I say...Unless you can prove that it is not an undertaking that will put more stress on the engine then keep your mouth shut and don't judge the decision. Unless you can accurately break it down step by step the process of doing this, the stress it will cause and why, just don't call it bs. It's like saying "wtf get the troops out of Iraq today! it would be sooo easy!" when i have no idea what so ever as to what it would take to get all of those troops out of there.
So I say, let's leave the game making to the game makers ahy?
Rollback Post to RevisionRollBack
"I want to say something but I'll keep it to myself I guess and leave this useless post behind to make you aware that there WAS something... "
-Equinox
"We're like the downtown of the Diablo related internet lol"
-Winged
I personally think they do have better things to focus on right now, but it would be nice to see a few changing aspects later on, placement of spikes, faces, slight color variation.
I personally think they do have better things to focus on right now, but it would be nice to see a few changing aspects later on, placement of spikes, faces, slight color variation.
Agreed its trivial compared to the many other things the team has to worry about right now. they might go back and consider this when more important things are seen to tho
Giving every monster in the game a simple randomization of physical appearance would take an incredible amount of work, and time although it could be accomplished. The fact is that, will Blizzard really want to spend the energy to make each mob look slightly different rather than spend more time on making each type of monster more unique besides from the enormous amount of balancing which is involved in the late stages of game development. Also is the extra cost in developing a random monster generation worth it? Will more people buy the game if the incorporate this feature? I think not..
Rollback Post to RevisionRollBack
Dream as if you'll live forever, and live as if you'll die today. "James Dean"
Honestly, I think that the comment that it would take a lot of work is quite accurate. Think about it, anyone who plays the game for more than a few hours and pays attention will surely be able to notice 3 variations on monsters as easily as one. While it would be nice to have that I agree it would be difficult to "rearrange" this particular monster without creating an entirely newly designed mesh due to it's lack of symmetry and organic placement. As a artist leaving the generation of monsters to chance is a lot harder to swallow if you want perfection. The best thing to do from a design perspective would be to create multiple versions by hand. Obviously it would need to be a number higher than 3 to make a large difference to the average player. I'd say 10-15 variations, to take that concept and apply that to every character in the game (They are already creating multiple variations most likely, ie: Diablo was different for the different levels of Diablo II.) then you are definitely looking at multiplying the work load for the team by A LOT. I'd think that you'd want to have the game as far along as possible then attack it again and again and again to refine it. So what someone said about them possibly reconsidering doing something like this in the last polishing phases of production is the most reasonable if they wanted to do it. Thats just my opinion. I haven't posted on these forums yet, but I felt compelled to add my two cents on this topic. I think the game is shaping up nicely.
Most likely all the meshes in the game aside from MAYBE player avatars for armor attachment etc. are a single mesh, meaning that the parts aren't separated. The game engine would have to be designed from the start to handle different mesh parts and combine them in game for example like Quake, where the models are divided into head, torso, legs so that all the parts can play animations at once, ie: a walk/run cycle and a shooting animation. Diablo III does not look to be constructed this way therefore it is not a simple task to randomly place objects on characters bodies as some of you are suggesting. It just doesn't work that way.
We are talking about random parameters concerning attributes of already designed monsters (like the spikes on the unburied's back), and you guys honestly think it would take "months"? My god. There's as much "code" involved as in giving monsters a random hue.
Except most programs already have a list of nearly infinite hues and shades which are already pre-programmed into it. The variations in the design of a monster would have to be created, thus taking much more time..... Dunno how you haven't grasped this concept yet.. Sorry Doppel I usually agree with you.
Rollback Post to RevisionRollBack
Dream as if you'll live forever, and live as if you'll die today. "James Dean"
Well, I have to agree with the people stating that this is easier than what Bashiok makes it.
You could have a single simple overlaying mesh, that could be applied to 80% of the monsters, and then just rotate it, flip it, turn it upside down, etc. Heck, they could have 3 different ones. They wouldnt take more than a few hours to make in PS for a skilled artist, and I don't expect the coding would take more than a few hours either.
For each monster tho that would take quite some time considering how many monters are gonna be in the game....well thats assuming there is but well D2 had a lot of diff monters so we can assume D3 will to
I think you guys are thinking that this is something that can be solved solely by a programmer. While programmers are without a doubt very important to game development, it's not just a matter of a guy sitting at a keyboard and typing some code to get what you want. Give some credit to the hundreds of artists who design and create by hand every model in the game from scratch. A lot of the discussion here is based purely on imagination. Can what you guys are talking about be done? Surely. Can it be done easily to a game that is already in progress and wasn't planned from the start of the modeling process to function that way, thats a tad bit harder than you are suggesting.
All the characters are a SINGLE mesh, you don't put them together like Spore animals. The creation of the models in the game and how they are arranged is not a code intensive process in and of itself. Even the randomly generated environments have walls that are pre-modeled and just arranged in different configurations by the game engine. Having all the spikes as different meshes could work, but it on the most basic level would take more polygons and thats just a waste when you are trying to make an efficient model unless it's something that HAS to be done.
Randomized creatures are not a lot of work, it's very simple. It's also very heavy on the CPU/VC. Project Offset did it and people consider it a big achievement.
90% of Video Game stuff are not hard to implement, they're hard to optimize.
Bashiok is certainly overdoing it. In fact, so far from what I read of Bashiok's stuff I think he behaves like an idiot teenager.
Quote from "Erebus;393784 »
All the characters are a SINGLE mesh, you don't put them together like Spore animals.
Characters are a bunch of poly's... what are you trying to say here? By your logic, a skeleton with a sword is a single mesh, and we can't give him an axe instead.
Quote from "Devil_Dark"" »
the monsters are a veritable DUMP of code that has been put together, every single SPIKE probably holds att LEAST 10-20 rows of code.
Did you follow some course where triangles are generated via vertex methods? Because in game industry, that stuff is usually done by programs automatically, nothing is manually programmed, it would take ages.
No, i think Erebus means something else. Weapons and shields are indeed separate meshes, they are rigid AND NOT related to the monster mesh in any way. Equinox, a mesh is a bunch of polys, but the polys are connected together. in spore, the different body parts are separate meshes (legs,head,arms etc), and that's why they look so blobby (because the meshes are not connected at the joints). Erebus' logic is flawless here.
and of course, a mesh's code is written automatically. you model it in 3ds max or maya or similar, and then you export in the right format and import it into the engine. done. i don't know what you mean by your vertex methods here, equinox. that's completely irrelevant as far as this discussion is concerned (Devil_Dark's argument does not hold either).
as for the spikes: as i look at them, i honestly can't see whether they are in one mesh or not. the spikes on the arms seem to be, because the spike's protrusion is aided by the texture on the flesh (i mean more red, you can see how the spike ripped through the flesh and left scars - but that's just the story behind it). the spikes on the back seem to be primitively attached, though.
so, to keep looking it that good (with the marks on the body), not only is the position of the spike relevant, but also the texture on the flesh, because both spike and body are related to eachother. if you put the spikes in randomly, you won't be able to put the flesh texture accordingly to it (=you lose visual quality). why? that would require additional matrix transformations and projections for each spike, and you still wouldn't get the quality that you get when you do it by hand. UV-stretching and UV-seams are guaranteed here.
and there is another big problem with animations. usually, after modeling and texturing is done, the character gets rigged and then animated. and, in the case of the unburied, the animator exactly knows where the spikes are and sees immideately where the limitations of his movements are (e.g. you couldn't put the arm close to the body because the spikes would pierce the body). when you randomize the position of the spikes however, you won't be able to control such things. unless you utilize procedural animation. and that's out of question.
so, characters (monsters, heroes) are separate meshes, and equipment are separate meshes. and that's good this way. but the spikes are a different topic, because they relate to the body visually. equipment doesn't. you can attach it anywhere (in the hand, at the hip, on the back) and it'll be okay. but the spikes have grown out of the body, so there need to be some marks on it that represent it. and it's the case right now. if it were automated, it would look much uglier.
Do you do this stuff for a job or something? seems liuke you know yer shit
The weapons and shields of creatures and players are in most cases separate objects in order to aid in collision detection. Also, for the specific case of the skeleton's they are symmetrical and randomly selecting different objects for them to spawn with on attachment points is a lot easier than rearranging a characters base structure. In theory the skeletons can be set up in the exact same manner as the player characters and have any armor/weapon placed on them in the same area as the avatar characters, although they most likely have a selection of only a few items designated for them.
Weapons and shields are indeed separate meshes, they are rigid AND NOT related to the monster mesh in any way.
Which means you can just as easily create a lower priority monster mesh and attach extra monster meshes to it in a random pattern. This has been done before... it can be done again. There is a system for it by now. You may not know it, nor do I know it, but it's there, and you need not assume all the complexities if they may not even be there anymore.
Quote from "CapKosmaty" »
Equinox, a mesh is a bunch of polys, but the polys are connected together.
And they can be disconnected. Meshes can be toyed with.
Quote from "CapKosmaty" »
in spore
Spore is a video game. I'm talking in terms of programs that create 3D models, not a game that is specifically adapted for people to toy with it and stuff. Please stop bringing that game up, I never mentioned it at all.
Quote from "CapKosmaty" »
i don't know what you mean by your vertex methods here, equinox. that's completely irrelevant as far as this discussion is concerned (Devil_Dark's argument does not hold either).
I'm sorry that you don't know anything about vertex methods. If you don't know something, I suggest you don't mess with it. I commented on a guy's post. What do you want here?
And, vertex methods: http://msdn.microsoft.com/en-us/magazine/cc164112.aspx
You can generate vertexes, polygons, etc. on a very very low level of programming.
Obviously, nobody uses that, but to give you an idea, since you totally had to butt into our convo.
Quote from "CapKosmaty" »
but the spikes are a different topic, because they relate to the body visually. equipment doesn't. you can attach it anywhere (in the hand, at the hip, on the back) and it'll be okay. but the spikes have grown out of the body, so there need to be some marks on it that represent it. and it's the case right now. if it were automated, it would look much uglier.
I don't think anyone is arguing that "It can't be done.", by going back and reading all the previous posts you can see clear examples of why the process of RANDOMLY generating spikes on this creature would be a task that is more indepth than simply clicking a few buttons or typing a few lines of code.
Yes indeed you can theoretically script something that would randomly generate spikes in designated locations on a mesh, but why would you do that? Why wouldn't you just edit the mesh by hand, it would end up taking less effort. The thing is though that you'd still have to take the time to retexture the new meshes anyway. The whole work flow you are suggesting would take more time than actually just making a new variation on the character. The reason Spore was brought up was because people are assuming that it's easy to drag one part of a model and place it in another position. You CAN detach part of mesh and then weld the points to another part of a mesh or another mesh all together, but from an edgeflow point of view and with polycount in mind it would make more sense to make a new character from scratch if you wanted pieces of a character in different positions. We've already brought up the topic of bones etc. You'd have to re-weight the new mesh to make sure it deforms correctly regardless of how it's created. You can't leave these things up to pure computer generation. Why take the time to have the computer make something then go back and test it and tweak it to make sure it's right when you can just make it right the first time yourself?
I don't think we need to accuse each other of not knowing what we are talking about, but I think a reasonable person will conclude that it will take a large enough degree of effort to warrant the original response by Bashiok that we are discussing.
I agree with everything you said. The question, however, is not if it would make the game more immersive for the players, the question was if the additional meshes and 3D-modeling, programming, and algorithms for this to become a reality would be feasible for low-end computers, disc space, speed and efficiency, the Diablo team's time restraints, etc. Also, for monsters that die in the amount of time it takes you to click the mouse button once or twice, is it really worth it? There would have to be sacrafices of other parts of the game that we might have enjoyed more.
If it had to be done, I would go with your view of it there.
Right now they just have more important stuff to do rather than worrying about monsters that will be killed in a matter of seconds, like he said.
Ideally though, I wouldn't mind them to have a little variation.
So I say, let's leave the game making to the game makers ahy?
-Equinox
"We're like the downtown of the Diablo related internet lol"
-Winged
Opinions may vary I guess.
I think this thread would have been better titled "Mobs that WON'T look slightly different by Bashiok" though.
Agreed its trivial compared to the many other things the team has to worry about right now. they might go back and consider this when more important things are seen to tho
>{Dynasty DiabloFans Thread}<
Artist Portfolio: http://www.JonathanSabella.com
>{Dynasty DiabloFans Thread}<
Artist Portfolio: http://www.JonathanSabella.com
Except most programs already have a list of nearly infinite hues and shades which are already pre-programmed into it. The variations in the design of a monster would have to be created, thus taking much more time..... Dunno how you haven't grasped this concept yet.. Sorry Doppel I usually agree with you.
For each monster tho that would take quite some time considering how many monters are gonna be in the game....well thats assuming there is but well D2 had a lot of diff monters so we can assume D3 will to
All the characters are a SINGLE mesh, you don't put them together like Spore animals. The creation of the models in the game and how they are arranged is not a code intensive process in and of itself. Even the randomly generated environments have walls that are pre-modeled and just arranged in different configurations by the game engine. Having all the spikes as different meshes could work, but it on the most basic level would take more polygons and thats just a waste when you are trying to make an efficient model unless it's something that HAS to be done.
>{Dynasty DiabloFans Thread}<
Artist Portfolio: http://www.JonathanSabella.com
90% of Video Game stuff are not hard to implement, they're hard to optimize.
Bashiok is certainly overdoing it. In fact, so far from what I read of Bashiok's stuff I think he behaves like an idiot teenager. Characters are a bunch of poly's... what are you trying to say here? By your logic, a skeleton with a sword is a single mesh, and we can't give him an axe instead. Did you follow some course where triangles are generated via vertex methods? Because in game industry, that stuff is usually done by programs automatically, nothing is manually programmed, it would take ages.
Do you do this stuff for a job or something? seems liuke you know yer shit
>{Dynasty DiabloFans Thread}<
Artist Portfolio: http://www.JonathanSabella.com
And they can be disconnected. Meshes can be toyed with.
Spore is a video game. I'm talking in terms of programs that create 3D models, not a game that is specifically adapted for people to toy with it and stuff. Please stop bringing that game up, I never mentioned it at all.
I'm sorry that you don't know anything about vertex methods. If you don't know something, I suggest you don't mess with it. I commented on a guy's post. What do you want here?
And, vertex methods: http://msdn.microsoft.com/en-us/magazine/cc164112.aspx
You can generate vertexes, polygons, etc. on a very very low level of programming.
Obviously, nobody uses that, but to give you an idea, since you totally had to butt into our convo.
See Project Offset. Crowd test demo.
Yes indeed you can theoretically script something that would randomly generate spikes in designated locations on a mesh, but why would you do that? Why wouldn't you just edit the mesh by hand, it would end up taking less effort. The thing is though that you'd still have to take the time to retexture the new meshes anyway. The whole work flow you are suggesting would take more time than actually just making a new variation on the character. The reason Spore was brought up was because people are assuming that it's easy to drag one part of a model and place it in another position. You CAN detach part of mesh and then weld the points to another part of a mesh or another mesh all together, but from an edgeflow point of view and with polycount in mind it would make more sense to make a new character from scratch if you wanted pieces of a character in different positions. We've already brought up the topic of bones etc. You'd have to re-weight the new mesh to make sure it deforms correctly regardless of how it's created. You can't leave these things up to pure computer generation. Why take the time to have the computer make something then go back and test it and tweak it to make sure it's right when you can just make it right the first time yourself?
I don't think we need to accuse each other of not knowing what we are talking about, but I think a reasonable person will conclude that it will take a large enough degree of effort to warrant the original response by Bashiok that we are discussing.
>{Dynasty DiabloFans Thread}<
Artist Portfolio: http://www.JonathanSabella.com