Arm Joint position

oXYnary

Does this look right to you? I thought it is already a little bit too broad as it is. I had it further in initially. Or do I need to go further out on the arm? (I hope not, I will have to re-position all child joints and re tweak the weights again.)

I realize part of the problem is the way I modeled the sleeve, but I can live with that/reweight if it is the cause of what makes it look like the arm is pressing against the chest.

animatr

i generally put the upper arm joint right where the ball and socket would be. you would need to bring yours in a bit. i also tend to put the joint smack in the middle of the "meat"
but yeah, bring it in to the next loop for better results.

oXYnary

I meant go out, not in. If I go in it looks like he is squishing his arms in more.

Hold on a sec...

Ok, here is the older version.

aesir

I think thats more of a weigting problem you're dealing with than a positioning problem.

Rob Galanakis

Because the loops are positioned incorrectly, you're never going to get proper joint placement.

If you think of vertex weighting as the vertices as children of the bone, think about how the bone rotates... the vertices you're concerned about for the shoulder come at the socket and down the arm/shoulder on the top of the shoulder. These are what need to bend down. However, the loops are mostly behind the socket, which means they can't be part of the deformation (and serve no purpose, really). If you re-angled the loops so they pointed diagonally up and out instead of up and in, you'd have this problem mostly fixed. Try positioning the joint where it 'should' be proportionally/anatomically, then positioning the geometry/loops to where it will provide deformations correctly.

Thinking about weighting as a parent/child relationship will help you understand where to position your joints, since you can visualize their rotation before you actually weight them.

oXYnary

Im lost. Up and out seems to be what they are now from how I am interpreting you. Can you show a visual by chance?

After correcting these loops, which bone placement should I use.

Thanks,

Rick Stirling

Even in that small rotation you need a bit of clavicle action. Try it by repeating the animation backwards on your own body - have your arm hanging by your side, then lift it. You will see your clavicle kick in once you get past about 45 degrees. It's not much, but by time your arm is horizontal the clavicle will have lifted the entire joint about an inch.

Take your top off.

Shoulders can be nasty, but helps to think of them as three joints.

The first joint goes through 90 degrees like youa re doining (then it continues up as you lift your arm, but thats very much clavicle based.

The second joint is with the arm hanging, where it swings. You can see how a twist bone would help here, pushing some of that swinging motion onto the clavicle meat.

The third joint is with the arm horizontal where it swings forwards. Try a rowing exercise. Y0u'll notice once again that the actual arm only has about 45 degrees of rotation before the clavicles kick in again to do the heavy moving.

oXYnary

Rick, see your PM.

Rick Stirling

shoulda just put it here.

oXYnary

Ok.. I guess it can help others.

(me)
[ QUOTE ]

Understood I think. That was impressed into me for the classes I took, but I guess Im not making the correlation to 3d as well with it?

Because from what I understand, the clavicle and humerus/upper arm bone use the same pivot in a 3d model, while in a human body, they are offset from one another (humerus attached to wing of scapula which floats below end of clavicle). Thus I have to make the pivot based on the position of the humerus versus having the correct angle for the clavicle for the arms outstretched in its figure mode.

Or am I missing something?

Unless I can deattach them from one another and offset? Or would I need a dead/space bone between (Biped wont allow will it?).

[/ QUOTE ]


Rick
[ QUOTE ]

Shoulders in 3d are a fucking horrible thing.

You are using things called bones, and people assume they are organic bones, but they are not. They are rigid objects trying to emaulte rigid bones, tendows and muscles. On top of that you have layers of sliding fat, skin and usually clothing.

it's all a big fudge.

The way I work in in a 3d model - the clav pivots from the center of the chest, the arm from the shoulder ball socket. Don't worry about an offset, just use two bones per "bone" as twists to carry the rotation.

[/ QUOTE ]

Yea, its still messes me up. I try to think of skeletal bones pivots as being my closest guide, but even then as you point out..

Thanks for the input so far Rick, it helps with this very confusing aspect.

So I'm trying to understand the "2 bone per twists" idea in that above, how would this translate? Which of the two 3d bone placement above would be closer to what you state and what when you say two bone per twist, are you referring to the vertex weights or other?

Rick Stirling

I'm just saying that you should use twist bones. You have two bones in the forearm for a twist, you have two in the upperam, and you can do the same in the clavicle. The clavicle twist is a clone of the clavicle, it inherits its position from the clavicle, but takes some X roll from the arm X rotation.

oXYnary

Oh ok, so instead of that helper bone (green one) or in addition to? I assume I attach it as a child of the clavicle?

oXYnary

Can anyone give me a visual of what Professor is meaning above with the loop angles/placement? Possibly by 2d'ing some lines in the correct position/angles over the images above? I'm still confused on that.

I would like to start again today after work since my bios is back up and running. I figure the replacement of the edge loops would be a good place to start. (Then the clavicle extras via Rick.