Actually the figure pair project is only tangential to the Neziņa movie, wasn’t originally in my plans, but seems a good way to put the rig in other’s hands and also give them a head-start on the modeling and skinning to the rig. Plus if it helps finance the movie, all the better!
Thanks, Brachi! No ETA as yet but since I’ll be using the generic models as the basis for all my other movie characters, I’m now giving it my full attention. I’ve already found a few places to improve the model topology and I have to test the rig modifications and skinning tweaks thoroughly, so it will take some time.
Also, since this is being produced in 2.5x, still in beta, I want to feel confident that future changes to Blender won’t break my work. I’ve had that happen in earlier versions and it’s a bit traumatic, much better to wait 'til the app is blessed as full release-level, or at least damn close.
Here’s a preview of the modeling done for the male figure so far:
The body topology could easily be in a textbook. Faceloops look excellent too, but I wouldn’t expect anything else.
I know it’s WIP, and it may be premature to say this, but I think he is too perfect. I would push the top of his head up some, and flatten his cheekbones, possibly even pronounce and sharpen his jaw. All just subtle changes of course,…Ahh, I just fully read the post, this guy is supposed to be generic. I had automatically assumed he was either Nezina’s partner or kin to her of some kind. The similiarities in their faces are undeniable. Guess that tells you what a fantastic job you are doing chip!
That’s interesting to hear – Neziņa will have two brothers, and her father will also be portrayed, so having a “family resemblance” is a good thing. But I’ll also be using the generic male model to create other non-family characters, so I guess I’ll have to push those characters’ features to greater extremes to keep them fully distinct.
Penny Pickup Demo Password = chipsvids (due to nudity)
First animated sequence done with the generic male model I’m working on, meant to be stepped through frame-by-frame for analysis rather than played at speed (much like the female model sequence above). Deformations are looking good, though the more defined male musculature is making this a lot more complex in terms of vertex weighting.
I’m still in the process of refining the rig and its controls, but this shows good progress so far I think. Since I’m adapting the Sintel rig, I have to work out how much I should keep and how much can be left out or modified substantially. It’s a very complicated rig, but has a lot of outstanding features that make it worth learning about in depth.
Works fine now. Thanks. Beautiful! Movement looks great, impressive how everything maintains it’s volume so well. I assume you are using a mesh deform modifier? If so, I would love to see the deform cage.
Thanks, 3dementia! No mesh deform mod’r, just bones. Bones, bones and more bones, and lots of Transform Constraints. Volume is maintained by scale transforms of various helper bones. This is in keeping with the approach I took with the Othello face rig, using bone transforms to emulate muscle action.
I saw you mention this in an earlier post… I am curious… why do you use a bone scaling instead of using a stretch bone constraint? does scaling give you more uniform deformation?
I’m not familiar with how the Stretch To Constraint works in practice, but it seems that its scaling capabilities are uniform along both axes enabled for scaling, based on the numeric value assigned to scaling. I can see how this can be very useful but also limiting in some ways. Since I set the Transform Constraints to Influence < 1.0 in most cases, it means I can manually adjust scaling of a bone in either axis differentially as needed to get the best “tweak” of the deformation. I can’t see how the Stretch To will allow that kind of versatility, though using partial Influence may allow it as it does with the Transform Constraint.
The mechanics of the Stretch To also require setting up appropriate Target objects, another layer of complexity within the armature system. Using Transform the bone already in place for general movement (such as the humerus/upper arm bone, or the tibia/shin bone) is the constraint Target. No need for an additional Target object in the rig.
In the case of the deltoid helper bone, a Stretch To might be useful since it combines both tracking & scaling (I use a Damped Track plus a Transform in my rig), but I’m familiar with how to set up the scaling method so I used that instead. In most other case there’s no need for a specifically-targeted stretch (Y-axis scale), just a generalized transform.
Transform constraints also allow easy re-orientation of the bone via rotation & translation, necessary for the manual adjustment options I like to maintain. I’m not sure how flexible Stretch To would be in that regard, since orientation depends on specific constraint settings that aren’t adjustable via animation (I don’t think they are, anyway).
All of which is not a crit of the Stretch To constraint, it seems a very useful tool for certain automated deformations, but maybe not as amenable to manual tweaking as the process I’m using.
The matter of “uniform deformation” is really more dependent on vertex weighting than on what constraint is being used. Constraints modulate bones (in the case of armatures, that is); it’s the bones that directly modulate the mesh surface via the vertex weighting mechanism, so evenness of deformation (when desirable) is a function of vertex weight distribution, regardless of how the bone is being transformed.
I am no expert in stretch-to, either, but I found myself using them because in Messiah we had something called “Muscle bones” that behaved similar. The thing is that you get a bit of a “squash and stretch” action that may or may not be desired.
That is very true. In fact, I guess you could scale it on three axis if you needed to which is an interesting point.
While working on the generic male figure, I noticed how the untextured figure looked much like some kind of statuary. One of my male refs is Michelangelo’s “David,” so I borrowed that pose to work on the deformation/weight painting details. When it came time to do a test animation, seemed natural to have him using a sling, so I dug up a few video refs and put together DaveTheSlinger. Nothing spectacular but it shows well how the figure deforms and gave me chance to work with the Cloth sim in 2.56 as well.
So you are concerned with deformations of body, but you have good animations here. I think deformation is important but this is fairly unnoticed when you have a good animation, natural blur can fade these deformations.
Fine deformations, you’ve worked a lot I can see.
But the pectoral of muscles of your model are quite inactive in that …sling action, while they have a major part in providing the force that throws that stone, or projectile, or whatever it is. You have many bones there, make them strain a little.
Test this movement on real models to see other possible issues.
The rig has FK/IK switching for the arms & legs, so I did this test animation to get this feature squared away & see how the custom corrective systems work with IK-constrained motions: CleanAndJerk
In general it was successful but much more problematical than using FK, mainly because some of the custom bones act differently in an IK-governed arrangement.
IK was used to “pin” the hand and the barbells together, with the switch for each hand coming when it first makes contact with the bar. That worked smoothly, but the deformations of the arms and shoulders took a lot of finessing to get looking right.
If you inspect the action frame-by-frame you’ll find a few places where things aren’t optimal in terms of deformation at the shoulders, but the action is so swift that these are unnoticeable. Adding motion blur would help “smudge” these frames a bit as well as make the action look smoother, but that adds a lot of time to the render, and for these tests I’m more interested in analysis than visual perfection.
I like the sling animation. Everything is looking excellent. There’s something about the arms… I’m not sure, but they look a little bit feminine to me, compared to the well defined musculature in the chest and abs… What do you think?
