Various characters seem to have their pros and cons, strengths and weakness'

I have seen on TV how forensic sculptors recreate a human face starting with the skull

and then building up muscle and various tissue to end up with a completed sculpt of the person.

Could this same approach be applied to character creation?

By this I mean instead of modeling the figure from the outside in, it would be modeled from the inside out. There is nothing new about what I am suggesting here, Ryan Kingslein (ZBrush) talks about the importance of understanding anatomy all the time.

These principles have been used since Greek and roman times to create beautiful sculpture.

So my question is this, would it be possible through rigging and perhaps weight mapping

so that posing and character movement were focused on the underlying bone, muscle, and tendons. Rig or weight map the figure in such a way that when posed the appropriate muscle and tendon groups flexed or extended and the skin surface reacted to the underlying structure in a natural and realistic manner. Would this approach help eliminate unnatural bends etc.

Not in Poser or Daz Studio, not very easily anyway (without tricks I can't think of), the rigging is too rudimentary for such complex mechanics.

Maya and Houdini and possibly few other high end apps can do this, I haven't done it myself past watching a few videos about how it's done. 

Nope. A lot of the issue with bending comes from user error and people not carrying a bend or rotation through the logical body parts involved. A perfect example is raising and lowering the arms. People will select the shoulder, rotate that and call it a day. Ignoring the fact that the collar is also involved in raising and lowering the arms and that at certain angles you not only rotate up or down but you also twist. Same goes for characters that have a buttocks actor. People would just select the thigh, rotate and then complain that the bend is bad because they just didn't bother to rotate the buttocks as well as the thigh.

So no, I don't think that'd solve the problem. You'd have to educate your end user on correctly using the figure also.  Believe it or not, with our figures, Darrell attempts to simulate what the muscles do when bending.  Through helper bones, magnets and morphs he tried to simulate this stuff with Miki 4 and Rex and Roxie and most folks complained that rigging clothing for them is too complicated because of that. So there's also the issue of a lack of willingness to accept the very changes that would help achieve the better bending you speak of.   I mean, he may not have gotten it 100% right - nothing is ever perfect - but he tried to do exactly what you're suggesting and met opposition because of how complex that made both working with the figure and creating content for it.  That's why Rex and Roxie are a little less complex than Miki4 in the rigging.

So yeah, until a way can be found to get over the X factor of the User, you're not going to see much of a change there in terms of how well things bend. Folks'll just keep complaining, not ever looking at themselves as well as the character's rig.  At least, that's the experience I've had in creating for the masses.

In theory you could handle those collar and pelvis moves with ERC, but that would only work for the case where we turn the clavicle to extend the range-of-motion of the humerus.  

In real life most of the clavicle's motion tends to be opposite to the humerus, in order to keep the hand in the same global position.  For instance, when we walk carrying a coffeepot, everything moves neatly to keep the coffeepot at the same Y and X, sliding smoothly forward on Z.   You might be able to handle this with a very smart IK system, but I sort of doubt it.

I think it's possible, even in poser, but it would never fly as a standard figure. The rig would be so complicated, the average poser user would go cross-eyed.

As Teyon points out, the average end-user doesn't want complexity, they just want the end results. Not possible to have both, but also not possible to get them to see that either. Most people won't even use the cloth room cause it's "too complicated" but then complain that conforming clothing doesn't look natural. 

As it is, figure creation and rigging for poser is a lot more involved than it is in other higher end software. When you're building characters for games and film, each figure has its own rig that's specific to that character's shape. Other than facial expressions, which are also specific to that character, the rig doesn't have to be as versatile as a poser rig does, in order to accomodate all the different characters that the base figure and its rig are meant to accomodate. 

So in order to have such a rig, and expect it to always work properly, you'd have to forego the versatility of the morphing that gets you from Aiko 3 to the Freak on the same mesh, and instead start using more specifc rigs and their designated characters. Even most facial morphs would pretty much be out the window if they change the shape of the face too much. Even V4's expressions get wacked out if her face has been morphed too much away from its base shape, and it doesn't even have to be all that different. Most full facial morphs should include adjustments to expression morphs too, and most don't bother doing that.

As I see it, I think there is a market for advanced poser figures, for users who want a more realistic figure to work with. But with that comes the necessity for the user to actually learn how to use the figure properly. Some will. Most will complain they don't get the results they see in the promo pics after clicking 2 or 3 buttons and spinning a dial.

Not trying to bash poser users, cause I know there are a lot out there who actually are willing to dig in and learn new tricks, but as a whole, a figure has to be built to appeal to the masses who expect all or most all of the characters they work with to be based on the same base figure. In order to have that, the rig will always be limited in what it can do.

~Shane

"smart IK system"

That was my thought - coming at this from a COMPLETELY naieve standpoint. If user puts body part in a certain position,, the "solver" (don't they actually call them that?), can go back through and adjust the appropriate parts to produce the anatomically correct solution. I'm sure there may be some variation, i.e. more than one solution that would end up with Vickie's hand on her bum, but the solver should be able to pick the most "natural" or "comfortable" one, maybe by minimizing the stress in the system. An instructor teaching a model to pose or a parent getting their child to 'sit up straight' does hands on adjustments. Those are refinements to normal human positioning. Here, I would think that first and foremost, you're talking about just eliminating the unnatural/impossible. If you got that down, you could add a provision to widen the envelope for things like age, fitness, weight etc. 

We have plenty of people here who can eyeball it. There should be data sets somewhere detailing the limits for the average human. There should be algorithms available in general - shouldn't have to go with something like neural networks or advanced 'AI.' The question I suppose would be is it a limitation of the rigging system (duh idiot, that's what they've been saying) in terms of not enough information or flexibility. Or, is it a limitation of the way Poser is dealing with it, some combination or something else altogether.

Quote - Nope. A lot of the issue with bending comes from user error and people not carrying a bend or rotation through the logical body parts involved. A perfect example is raising and lowering the arms. People will select the shoulder, rotate that and call it a day. Ignoring the fact that the collar is also involved in raising and lowering the arms and that at certain angles you not only rotate up or down but you also twist. Same goes for characters that have a buttocks actor. People would just select the thigh, rotate and then complain that the bend is bad because they just didn't bother to rotate the buttocks as well as the thigh.

 

So no, I don't think that'd solve the problem. You'd have to educate your end user on correctly using the figure also.  Believe it or not, with our figures, Darrell attempts to simulate what the muscles do when bending.  Through helper bones, magnets and morphs he tried to simulate this stuff with Miki 4 and Rex and Roxie and most folks complained that rigging clothing for them is too complicated because of that. So there's also the issue of a lack of willingness to accept the very changes that would help achieve the better bending you speak of.   I mean, he may not have gotten it 100% right - nothing is ever perfect - but he tried to do exactly what you're suggesting and met opposition because of how complex that made both working with the figure and creating content for it.  That's why Rex and Roxie are a little less complex than Miki4 in the rigging.

 

So yeah, until a way can be found to get over the X factor of the User, you're not going to see much of a change there in terms of how well things bend. Folks'll just keep complaining, not ever looking at themselves as well as the character's rig.  At least, that's the experience I've had in creating for the masses.

This.  100%.  

We REALLY need something like "like" or "agree" button here L

Quote - We REALLY need something like "like" or "agree" button here L

Nod nod nod

Quote - Various characters seem to have their pros and cons, strengths and weakness'

I have seen on TV how forensic sculptors recreate a human face starting with the skull

and then building up muscle and various tissue to end up with a completed sculpt of the person.

Could this same approach be applied to character creation?

By this I mean instead of modeling the figure from the outside in, it would be modeled from the inside out. There is nothing new about what I am suggesting here, Ryan Kingslein (ZBrush) talks about the importance of understanding anatomy all the time.

These principles have been used since Greek and roman times to create beautiful sculpture.

So my question is this, would it be possible through rigging and perhaps weight mapping

so that posing and character movement were focused on the underlying bone, muscle, and tendons. Rig or weight map the figure in such a way that when posed the appropriate muscle and tendon groups flexed or extended and the skin surface reacted to the underlying structure in a natural and realistic manner. Would this approach help eliminate unnatural bends etc.

motion capture ,MoCap for short will Simulate real human movements.

Quote - motion capture ,MoCap for short will Simulate real human movements.

 

Even motion capture has to be tweaked by an animator. 

And there's no current way of getting mocap data into poser. Unless I missed something somewhere.

~Shane

Quote - I think it's possible, even in poser, but it would never fly as a standard figure. The rig would be so complicated, the average poser user would go cross-eyed.

As Teyon points out, the average end-user doesn't want complexity, they just want the end results. Not possible to have both, but also not possible to get them to see that either. Most people won't even use the cloth room cause it's "too complicated" but then complain that conforming clothing doesn't look natural. 

As it is, figure creation and rigging for poser is a lot more involved than it is in other higher end software. When you're building characters for games and film, each figure has its own rig that's specific to that character's shape. Other than facial expressions, which are also specific to that character, the rig doesn't have to be as versatile as a poser rig does, in order to accomodate all the different characters that the base figure and its rig are meant to accomodate. 

So in order to have such a rig, and expect it to always work properly, you'd have to forego the versatility of the morphing that gets you from Aiko 3 to the Freak on the same mesh, and instead start using more specifc rigs and their designated characters. Even most facial morphs would pretty much be out the window if they change the shape of the face too much. Even V4's expressions get wacked out if her face has been morphed too much away from its base shape, and it doesn't even have to be all that different. Most full facial morphs should include adjustments to expression morphs too, and most don't bother doing that.

As I see it, I think there is a market for advanced poser figures, for users who want a more realistic figure to work with. But with that comes the necessity for the user to actually learn how to use the figure properly. Some will. Most will complain they don't get the results they see in the promo pics after clicking 2 or 3 buttons and spinning a dial.

Not trying to bash poser users, cause I know there are a lot out there who actually are willing to dig in and learn new tricks, but as a whole, a figure has to be built to appeal to the masses who expect all or most all of the characters they work with to be based on the same base figure. In order to have that, the rig will always be limited in what it can do.

 

~Shane

:thumbupboth:    100%agree