• Ryu-Gi's 3D modeling thread
    207 replies, posted
actually tony I think I lost track of the message you were trying to convey perhaps it got lost in the river of bullshit you were spewing
[QUOTE=ILY;36423218]This argument would probably have gone better and been over with quicker if people weren't calling one another dumb in every post.[/QUOTE] I can't recall ever spouting ad hominem. I don't consider saying "you are not understand me" to be an insult, just a statement. Unlike the insults directed at myself from others.
[QUOTE=Maloof?;36423232]I don't recall any dumb-calling[/QUOTE] They aren't flat out saying it, but stuff like this is clearly an insult: "You've got no clue what you're talking about."
[QUOTE=Recurracy;36423238]actually tony I think I lost track of the message you were trying to convey perhaps it got lost in the river of bullshit you were spewing[/QUOTE] I really don't care [I]what[/I] you think. It doesn't really [I]matter[/I]. That does not change the fact that for highly realistic and convincing facial animation you simply [I]can not[/I] skimp out on polycount and expect great results if the face is going to move in any way unless it's a cartoony character.
[QUOTE=TonyP;36423225]Yes, those games have developers who go to the extra mile to have wonderfully detailed characters. So why cop out and not take advantage of what is capable by computers these days? The weak (by modern standards) PS3 and Xbox can have the fabulously detailed and complex sincere facial animation, and there's many PC games which do too, so there's no reason to not do it other than lack of skill or for some sort of retro art style or mobile gaming developement (PS Vita not included). Or outdated game rendering engines.[/QUOTE] I don't know, Half-life 2 seems to have pretty cool (put in your words, realistic and convincing) facial animations for just a few polygons. I really don't understand why the characters need more polygons in your eyes. Mind you that Half-life 2 was created in 2007 or so. The HWM faces in TF2 use blended normal maps on extreme expressions for foreheads to create that extra mile of 'convincingness' you desire, and those models aren't detailed as fuck.
[QUOTE=TonyP;36423225]Yes, those games have developers who go to the extra mile to have wonderfully detailed characters. So why cop out and not take advantage of what is capable by computers these days? The weak (by modern standards) PS3 and Xbox can have the fabulously detailed and complex sincere facial animation, and there's many PC games which do too, so there's no reason to not do it other than lack of skill or for some sort of retro art style or mobile gaming developement (PS Vita not included). Or outdated game rendering engines.[/QUOTE] So you want him to spend days/weeks making incredibly detailed facial animations comparable to ones made by a high-budget company with large teams of people with a full studio recording facial expression references from every possible angle of models who look exactly like the 3D character [i]why?[/i]
Because it's extremely simplistic by current standards and isn't nearly detailed enough to look realistic? Because I like progress? Do you like progress of animation technology? Or do you think we should stay at current quality forever and not try refine any further?
[QUOTE=TonyP;36423225]Yes, those games have developers who go to the extra mile to have wonderfully detailed characters. So why cop out and not take advantage of what is capable by computers these days? The weak (by modern standards) PS3 and Xbox can have the fabulously detailed and complex sincere facial animation, and there's many PC games which do too, so there's no reason to not do it other than lack of skill or for some sort of retro art style or mobile gaming developement (PS Vita not included). Or outdated game rendering engines.[/QUOTE] I don't get why you would model an in-game character with that many polys when you're viewing them from the ass-side for 99% of the game, using up processing power anyway because it's in-frame. Power that could be used to render the environment and you're actually going to be examining in-depth for game play purposes. Yes, it's impressive that they would do such a thing. Seems like a silly waste of time (and processing power) to get somebody to go that hard core over modelling when they could have animated the bare minimum expressions required for cut scenes and used a lower poly, bumped face for in-game.
[QUOTE=TonyP;36423287]Because it's extremely simplistic by current standards and isn't nearly detailed enough to look realistic? Because I like progress? Do you like progress of animation technology? Or do you think we should stay at current quality forever and not try refine any further?[/QUOTE] If he went to work for a big company worth millions of dollars where facial animation is incredibly important (e.g. L.A. Noire), then sure. But it's a stupid waste of time and a waste of resources to do what you're suggesting for this model. [editline].[/editline] Especially since, as Maloof pointed out, you're going to be seeing the back side of her for most of the game. Only an idiot would think a game is bad because [i]the eyes lids didn't bulge[/i] in a cutscene.
[QUOTE=Dr. Evilcop;36423280]So you want him to spend days/weeks making incredibly detailed facial animations comparable to ones made by a high-budget company with large teams of people with a full studio recording facial expression references from every possible angle of models who look exactly like the 3D character [i]why?[/i][/QUOTE] That's a massive exaggeration. He just needs two things: proper topology, and a decent face rig. The actual animating of the face rig shouldn't take much longer if it's properly rigged. He doesn't need to have realistic facial motion for the folding and bulging of the face to occur. [editline]21st June 2012[/editline] [QUOTE=Maloof?;36423288]I don't get why you would model an in-game character with that many polys when you're viewing them from the ass-side for 99% of the game, using up processing power anyway because it's in-frame. Power that could be used to render the environment and you're actually going to be examining in-depth for game play purposes. Yes, it's impressive that they would do such a thing. Seems like a silly waste of time (and processing power) to get somebody to go that hard core over modelling when they could have animated the bare minimum expressions required for cut scenes and used a lower poly, bumped face for in-game.[/QUOTE] LOD says hello.
[QUOTE=TonyP;36423287]Because it's extremely simplistic by current standards and isn't nearly detailed enough to look realistic? Because I like progress? Do you like progress of animation technology? Or do you think we should stay at current quality forever and not try refine any further?[/QUOTE] May I need to remind you that most players won't even get to see the wireframe models, they don't care how it's made as long as it's convincing enough. L.A. Noire got a very positive reception for facial animations and those models aren't extremely detailed. I like progress too, but why fix it if it ain't broke? LA noire's system looks just fine.
[QUOTE=Recurracy;36423269] Mind you that Half-life 2 was created in 2007 or so.[/QUOTE] Try < 2004. Crysis came out in 2007
[QUOTE=TonyP;36423287]Because it's extremely simplistic by current standards and isn't nearly detailed enough to look realistic? Because I like progress? Do you like progress of animation technology? Or do you think we should stay at current quality forever and not try refine any further?[/QUOTE] No, only a [I]few[/I] very financially well-off companies are doing it. Companies who likely have multiple people working on each character. People who likely have many years of experience in modelling. These are MAIN CHARACTERS we are talking about - they don't get newbies to model main characters [editline]21st June 2012[/editline] [QUOTE=TonyP;36423325]That's a massive exaggeration. He just needs two things: proper topology, and a decent face rig. The actual animating of the face rig shouldn't take much longer if it's properly rigged. He doesn't need to have realistic facial motion for the folding and bulging of the face to occur. [editline]21st June 2012[/editline] LOD says hello.[/QUOTE] Drake in Uncharted 2 had no LOD [editline]21st June 2012[/editline] [QUOTE=Recurracy;36423342]May I need to remind you that most players won't even get to see the wireframe models, they don't care how it's made as long as it's convincing enough. L.A. Noire got a very positive reception for facial animations and those models aren't extremely detailed. I like progress too, but why fix it if it ain't broke? LA noire's system looks just fine.[/QUOTE] Also exactly this. Video games (like films) are about smoke and mirrors. If you're spending more than you have to in terms of time, money and end-user processing power, you're doing it very wrong.
I just do not approve of him saying he's going to use modern standards and then present models of quality that wouldn't be out of place in 2006. (the textures are another story)
[QUOTE=TonyP;36423389]I just do not approve of him saying he's going to use modern standards and then present models of quality that wouldn't be out of place in 2006. (the textures are another story)[/QUOTE] Nor would they be out of place in today's games. Boo-hoo.
[QUOTE=TonyP;36423325]That's a massive exaggeration. He just needs two things: proper topology, and a decent face rig. The actual [B]animating of the face rig shouldn't take much longer if it's properly rigged[/B]. He doesn't need to have realistic facial motion for the folding and bulging of the face to occur. [/QUOTE] In order to animate a face so that individual wrinkles and shit move about, you have to rig those individual wrinkles. That is going to take a shitload of time. I'm not talking an extra half hour; I'm talking many times longer than it would normally take.
[quote]Drake in Uncharted 2 had no LOD[/quote] While he didn't have a traditional LOD I believe he has a dynamic degenerative system applied to his model but I may be wrong. But the reason they tried to maintain high quality in gameplay is because you often get closeups of his face ingame and his face has many animations that respond to situations.
[QUOTE=TonyP;36423389]I just do not approve of him saying he's going to use modern standards and then present models of quality that wouldn't be out of place in 2006. (the textures are another story)[/QUOTE] [URL="http://www.zbrushcentral.com/attachment.php?attachmentid=167535"]Modern standards say hello[/URL]
[QUOTE=Dr. Evilcop;36423401]Nor would they be out of place in today's games. Boo-hoo.[/QUOTE] Maybe not, but not to the standards set by current games. [editline]21st June 2012[/editline] [QUOTE=Maloof?;36423410][URL="http://www.zbrushcentral.com/attachment.php?attachmentid=167535"]Modern standards say hello[/URL][/QUOTE] That isn't up to modern standards, set by other games, but either way the face and body animations are simplistic. Half Life 2 has more complex looking facial animation than the AC games. [editline]21st June 2012[/editline] [QUOTE=Maloof?;36423402]In order to animate a face so that individual wrinkles and shit move about, you have to rig those individual wrinkles. That is going to take a shitload of time. I'm not talking an extra half hour; I'm talking many times longer than it would normally take.[/QUOTE] Actually, If the face is modeled correctly and the rigging is set up right, then sliding the 'smile' slider should automatically result in the crease forming. You don't put seperate bones for the valleys and peaks of the wrinkles. That's absurd and simply untrue.
[QUOTE=TonyP;36423416]Maybe not, but not to the standards set by current games. [editline]21st June 2012[/editline] That isn't up to modern standards, set by other games, but either way the face and body animations are simplistic. Half Life 2 has more complex looking facial animation than the AC games. [editline]21st June 2012[/editline] Actually, If the face is modeled correctly and the rigging is set up right, then sliding the 'smile' slider should automatically result in the crease forming. You don't put seperate bones for the valleys and peaks of the wrinkles. That's absurd and simply untrue.[/QUOTE] Yet regardless it's still a lot more effort, and he can get perfectly reasonable looking facial animations from the current mesh.
[QUOTE=TonyP;36423416]Maybe not, but not to the standards set by current games. [editline]21st June 2012[/editline] That isn't up to modern standards, set by other games, but either way the face and body animations are simplistic. Half Life 2 has more complex looking facial animation than the AC games. [editline]21st June 2012[/editline] Actually, If the face is modeled correctly and the rigging is set up right, then sliding the 'smile' slider should automatically result in the crease forming. You don't put seperate bones for the valleys and peaks of the wrinkles. That's absurd and simply untrue.[/QUOTE] If you're not individually rigging all of the wrinkles... wouldn't you get the exact same effect by using a displacement map to apply extra sculpted detail to the mesh at render time? I mean, the EXACT same effect - that of a series of wrinkles squashing or stretching based on the movement of a simple rigged plane?
You still do not understand do you? Look in your face in the mirror and raise your eyebrows or smile. Creases form and skin bulges out. You don't have a static face of creases and bulges that simply move with your face. Unless the displacement map is animated it's not going to look right at all.
[QUOTE=Rusty100;36423360]Try < 2004. Crysis came out in 2007[/QUOTE] Oh, my bad.
[QUOTE=TonyP;36423490]You still do not understand do you? Look in your face in the mirror and raise your eyebrows or smile. Creases form and skin bulges out. You don't have a static face of creases and bulges that simply move with your face. Unless the displacement map is animated it's not going to look right at all.[/QUOTE] But if you're not rigging the wrinkles how will you control their amplitude?
[QUOTE=Maloof?;36423511]But if you're not rigging the wrinkles how will you control their amplitude?[/QUOTE] Because the bones each have an effect on only some faces, when they move they'll form the creases automatically. It's simple. You can check youtube for some videos of how it looks if you like.
I guess the overarching point is that just because a couple of AAAA billionaire game developers do it, doesn't make it a standard. It's more of a niche at the moment, it seems [editline]22nd June 2012[/editline] [QUOTE=TonyP;36423529]Because the bones each have an effect on only some faces, when they move they'll form the creases automatically. It's simple. You can check youtube for some videos of how it looks if you like.[/QUOTE] So you still have to configure the wrinkles to the rig, which does mean significantly more time setting the rig system up and troubleshooting it
[QUOTE=Maloof?;36423539]I guess the overarching point is that just because a couple of AAAA billionaire game developers do it, doesn't make it a standard. It's more of a niche at the moment, it seems [editline]22nd June 2012[/editline] So you still have to configure the wrinkles to the rig, which does mean significantly more time setting the rig system up and troubleshooting it[/QUOTE] No, it's no different from rigging a low poly, but correctly topolgized face.
[QUOTE=TonyP;36423490]You still do not understand do you? Look in your face in the mirror and raise your eyebrows or smile. Creases form and skin bulges out. You don't have a static face of creases and bulges that simply move with your face. Unless the displacement map is animated it's not going to look right at all.[/QUOTE] Again, normal maps
Animated normal maps are far more cumbersome and hardware taxing than a few extra thousand polygons. Here is a massively oversimplified example of how a crease would be formed without having to have three bones per wrinkle [img]http://img7.imageshack.us/img7/204/simplification.jpg[/img] With a bone controlling the blue section, and a bone controlling the pink section, with a falloff of course to blend the bones effect on the model with nearby bones. Of course, even in a low polygon face, it's not just a single bone in the cheek, but for the sake of simplicity just imagine it to be. When the mouth is stretched, a smile line would be formed if the model has sufficient polygons. If it doesn't, a smile would be formed but it wouldn't have a noticeable smile line. This is how the very old games look, no smile lines, in most games older than half life 2. And even some poorly made modern games made by very lazy dev teams. This concept is applied to all of the large creases. At most, it would only require a few more bones in the forehead and eyes over a lower detail rig setup. Note that this face is not an example of a high detail face at all. It's at best a small step up over Half Life 2, which does have rudimentary and simplistic smile and brow furrowing.
[QUOTE=Ryu-Gi;36351053]All right, fair points. I'll post a wireframe. Not all T-poses look like this, though, in some games the character's arms will be straight out. However once I get this character rigged I will put it into the angled arm T-pose, since I'd like to port the model to Source eventually.[/QUOTE] Generally speaking when you want a skin pose you want to have a pose where at least the majority of joints are in between extremes. And as hands are rarely lifted over the the head, having them level can already be considered one of the extremes. Those level t-poses are more of a legacy thing these days. [QUOTE=TonyP;36423049]I sincerely hope this video is an example of non animated folding, because there's none in this video. All folds in this video are static. Although Half Life 2 does have some basic fold animating of the nasolabial folds and brows. Still very simplistic compared to current standards.[/QUOTE] That's because they don't use normal maps attached to the flexes, as is the case for L4d for instance. Also generally speaking, a shapekey based animation gives you a little bit more control than a bone based one though it takes harder to set up.
Sorry, you need to Log In to post a reply to this thread.