Slight Confusion on twist bones in Max
polycounter lvl 18
Ok, so with the bipeds twist option.. I weight the twist bones and ignore the main joint bone? I did a test and it seemed to work this way.
So if you need two twist bones to make a "twist", why does it offer the option of one? In those cases, do you use the main bone weighted as half of the twist?
Example: Upper Arm (weight 1/2)
Upper Arm Twist (weight)
Versus: Upper Arm (ignore)
Upper Arm Twist (weight)
Upper Arm Twist 1 (weight)
So if you need two twist bones to make a "twist", why does it offer the option of one? In those cases, do you use the main bone weighted as half of the twist?
Example: Upper Arm (weight 1/2)
Upper Arm Twist (weight)
Versus: Upper Arm (ignore)
Upper Arm Twist (weight)
Upper Arm Twist 1 (weight)
Replies
We use version one in GTA and it works (1 foretwist), though most of our twist bones are scripted normal max bones (since we started off on Max version 7).
If I only weight the main bone, the twisting motion does not show up in the joint. It will only show up if I unfreeze the twists, and envelope those instead of the main bone. I did a test biped with some twists on the forearm and used a tube. Shoulda took some pics, but as I state above, that was my results.
I use the skin modifier if thats maybe why??? Does the twisting "automatic" only work in physique?
Wouldn't you need more than 1 twist on the forearm Ruz? Right now I have 3.
My process of weighting is as follows..
Using envelopes, than manually going pack with pressure sensitivity on in paint weights to manually get a better gradient that matches joint/muscle movements, then for the final touch using vertex weights dialog on specific edges (like for creases).
EDIT: i don't know if this makes a difference, but this is with the twist option in the biped structure. I don't know if its any different than you manually constructed versions. I do know the tutorials and help in Max 8 don't give me enough information on its specifics in this regard.
Also, im sure max comes with a tutorial somewhere about twist bones.
Surely 2 bones are all that's needed for a forearm to twist ... the first bone being the one that joins to the elbow (and could have its rotation constrained to just one axis), then another one halfway down the length of the forearm, that is also constrained, but this time to the "twist" axis alone.
Then you just weight your forearm mesh smoothly along the length of those bones - the verts nearest the elbow only being influenced by the first bone, and the verts nearest the wrist only being influenced by the "twist" bone.
Same story for the upper arm, 2 bones is all you should need, one to connect to the shoulder, the other one to twist relative to that.
So in total if you have upper and forearm twist, you should have 4 bones in the arm chain between the shoulder and the wrist.
Edit: My description there is for standard max bones/skin, but I just tried it now with the Biped's automatic Twist bones and it works the same way.
You are on the right track with unfreezing the Twist bones, but you need to weight the twist bone AND the main bone - so for the forearm mesh, the main forearm bone would have NO weights nearest the wrist, where the forearm twist bone would have FULL skinned weights, and vice-versa so the twist bone has NO influence over the part of the forearm nearest the elbow.
Also make sure your twist bias (in the twist poses biped rollout) is set to 0.5, so the twist bone will rotate half the amount you rotate the hand bone (ie. rotating the hand bone to 90' will rotate the twist bone to 45' ... this probably looks best IMHO).
It's certainly not something you need Physique for though, skin works fine.
I can send you my really crappy 5-minute arm model and rig in max8 format if you want to see how i did it.
Mop.. thanks.. Ok, I will Xnay one of the links.
"need to weight the twist bone AND the main bone - so for the forearm mesh, the main forearm bone would have NO weights nearest the wrist, where the forearm twist bone would have FULL skinned weights, and vice-versa so the twist bone has NO influence over the part of the forearm nearest the elbow."
So "bony joints" in other words would be connected to the main bone versus the twists? So parts of the arm that the ulna shows up?
Maybe part of my confusion is the idea of twists themselves. The way I heard it hear before and what max describes, it again helps with making the mesh deform better and having more helps make the transitions smoother.
Or is a twist simply to mimick a particular bone? If a twist is to mimick a radius, couldn't I just use one twist and increase its rollout to 90 degrees or so? Or is more like the first idea and one bone twist cannot give enough of a smooth transition on the surface?
Hopefully this should explain it better.
If you look in the Skin modifier and select the bones you will see how I have weighted it (note which verts are weighted to the twist bones, and which are weighted to the "regular" bones).
Also try playing around with the Biped's Twist Bias in the Twist Poses rollout with the biped selected. The lower you set it, the less effect the twist bone will have - if you set it to 1.0 then the forearm twist bone will rotate the same amount as the hand for example. I have it on 0.5 since it was the default. 0.75 might make more sense from an anatomical point of view.
I would probably just forget about the ulna and the radius and all that stuff, and instead just focus on what you have to work with - the twist bones are just a tool for helping get the appearance of a more natural deformation, they don't really represent the "real" bone structure in any meaningful way.
Thigh (influences knee joint/patella - front side)
Twist (influences upper part of thigh)
Twist 1 (influences lower part of thigh)
Calve (influences tibia more than fibia at ankle)
Twist (upper part of calve/back of knee)
Twist 1 (lower part of calve and ankle)
Upper Arm (influences mostly around tuberosity except clavice/acromioun area
Twist (rest of upper arm)!!twist percentage!!?
Lower Arm (Influences around Ulna and ulna at elbow)
Twist (upper lower arm)
Twist 1 (rest of arm)
EDIT: Just read your post Mop. Ok that helps.. Any changes to my "plan" above you can see? I cant look at your model at this point as I am at work right now.
EDIT, EDIT: To better explain my reasoning, that the regular bone wont show the twist influence? So use it on parts of the joint that normally are static. In example, the ulna never changes position. the radius and carpals revolve around it.
EDIT, EDIT: To better explain my reasoning, that the regular bone wont show the twist influence? So use it on parts of the joint that normally are static. In example, the ulna never changes position. the radius and carpals revolve around it.
[/ QUOTE ]
Unless you are doing a medicinal model for which you need a more complex rig or muscle-simulation, your approach is way to complicated and prone to fail. There is simply too much moving tissue and muscles involved than could be simulated with help of just the twistbones. Especially having moving tissue next to bone parts like you suggest. In addition, for example the axis of rotation in the lower arm is between the little head of the ulna and the one of the radius and not the one of the twistbones. Simply skin them like Mop suggested. Physique has the ability of twistable links btw. so you wouldn't need twistbones there. (They are a real pain to setup there anyway, since a lot of overlapping links will run down the arms. )