Ok, well theres alot of stuff going on in a character studio biped to get it to have both IK and FK. with regular bone hierarchies you dont get FK and IK for free unless you want to swap between them by animating the IK solver on and off. but then you lose the IK handle when you move the limbs away from it when animating in FK, matt and I have been discussing this stuff in another thread, its tricky to set it up for easy usage. I could go into an explanation of how biped works underneath but basically using biped in this manner is going to be a bitch to set up and even more frustrating to animate. you could maybe try and use the arms as front legs and add extra arm limbs with bones and animate them with FK seeing as you probably wont need ik too much for some arm flailing type motion. Personally I would save yourself from learning all the ins and outs of character studio and spend the time learning how to rig manually. set up IK limbs for all his legs, FK arms and spine and you will have a workable rig without too many headaches. alternatively you could grab CAT and start with a tiger preset adding some arms in a similar way. CAT is far and beyond a better choice for automatic rigs in max. believe me.

---

thanks tyson. the reason i chose to add the extra bones as the front legs, and not arms, is because the whole head-spine-clavicle bunch will be negated if i use the biped's arms as front legs. i'd probably have to delete the head+neck+clavicle bones, and i thought this would be an easier thing to do, cos anyway i'd need a head bone, neck bone, etc...
i read the max tutorial on adding extra limbs, and more or less followed the process, but as you say, the biped's bones are much better.

where can i learn how to rig manually? the entire rig, i mean, including spine etc... max has only character studio biped tutes...
i would design the rig similar to the rig i posted above...

---

Cogito, Ergo Sum

Ok dude. I've uploaded a series of four rigging tutorial movies I made a few years ago. They are a bit old now but I think they could still be usefull for some of the basic processes of bones rigging. They cover rigging the legs, feet, pelvis and spine. I never got upto the arms but feel free to ask any questions about that when you get up to it. They use the techsmith codec which I have included. I tried to get them up onto youtube in HD but I didnt have much success as yet. let me know if they are of use to you.

[Link to rapidshare.com]

---

wow thanks man am downloading it will keep you posted! i'm sure it'll be an amazing help

i've left CAT as a last resort i think i'm over-estimating CAT, i'm under the impression that it will be like "click here for a centaur-type rig"
lol

---

Cogito, Ergo Sum

here's an attack pose shot of the rig. can you tell by the pic if it sucks, tyson?
as you can see in the screenshot, your tutorial is still only some 30% downloaded, so these are just my premature attempts at trying to fix things.

Edit: what i did --> snapshot of the biped hind legs, then made link hierarchy of thigh-->calf-->ankle-->foot, (linked thigh to spine), then IK HI from thigh to calf, another from calf to toe. then made the calf-toe IK the parent of the first IK.

---

---

Cogito, Ergo Sum

The pose looks pretty cool actually. If it all works like a tiger with a man ontop you can probably move on to the joys of skin weighting
maybe attempt a walk cycle with the rig before you move into skinning. The thing about biped compared with regular max rigging is that you dont really have the same level of smooth control over everything. To animate a walk cycle you need to use planted keys whenever the foot is on the ground and free keys whenever they are in the air. then you can't really adjust the besier/euler curves the way you should be able to. With max objects they all have a besier position and euler rotation controller but in biped they have querternion rotation with an optional euler curve ontop which gives some bizarre results. Bones are a far better option in the end. but you do have to account for everything in the rig. I'm not sure how complete these vids are I havent sat through them and they were intended as a introduction to rigging nothing more. Ive learnt alot since then too

---

You can adjust your curves in biped using TCB values. You'll achieve the same results but yeh it is a little bit more complex.

---

[Link to www.3dprevis.com]

just a note on the pose if your interested. strong solid poses are very important. I would def do something with the hands and fingers as they can be extremely expressive. Maybe clench his fists and have his arms point straight our behind him almost like he's thrusting his chest out or something.

---

and it's a story that might bore you, but you don't have to listen, because I always knew it was going to be like that.

@Tyson: just went through the vids man they were realllllllly good, and extremely helpful. just what i was looking for, in fact. is your IK setup for the legs the way to go? i have a few questions if you don't mind.
1) when you created the IK chain for the leg, you did one for thigh to ankle, 1 for ankle to foot, and one for foot to toe. any reason in particular?
2) you created a toe helper but aligned it to the ball goal. you then linked the ankle goal to the ball helper, and linked both the ball and toe helpers to the ankle helper. i didn't understand your reasons for these few steps, i saw the results though and it was a very good, natural leg. but could you explain your reasons for this sort of cross-linkages?
3) you created the head bone, but you intentionally refrained from starting at the neck nub. later on, you linked this head bone to the neck bone, but not the neck nub. why did you choose to link this way, and not start at the neck nub for the head bone?
4) the spline you created for the tail, does that mean the tail can bend only at the vertices of the spline?

thanks a ton for those tutes man.

@Tyson, Gerard : i'm really lost with all that you guys said about euler and quaternion rotations etc, and TCB values. will read those up in the Max help

@ Matt: yeah man this was just a brief demonstration based on a concept image i have of the uuka's attack pose. was going to do the clenched fists, and the right arm is behind like he's thrusting his chest out

---

Cogito, Ergo Sum

thats cool, I'm glad they were usefull and clear enough to understand.

ok.

1. The ik solvers through the leg and foot are necessary for the foot rig I used later on.

2. There are many ways to rig a foot but this one seemed the easiest option while still giving you the control you need to animate the foot properly. ie: you need the ability to rotate the foot around different points throughout a walk cycle. the relationships between the helpers and the ik solvers gives you the right rotations. for example you would use the foot control to move the feet up, back and forth and to rotate at the heel. then use the ball control to make the foot roll up on the ball of the foot. the first contact pose would have the front foot with the heel on the ground and the back foot with the heel off the ground and the ball of the foot rotated flat on the ground. then as the front foot rotates down flat the back foot comes up and you use the toe helper to make the toes drag behind the motion.
Its a bit hard to explain. maybe I will make walk cycle video some time

3. I'll have to look at what I did again on that one but the head and neck are pretty simple, just have a neck bone and a head bone and link them to the top spine bone. it will all be FK anyway.

4. Its not totally necessary to use spline Ik but it does make for a nice simple way of controlling the tail. once you apply the spline IK you will have control objects on each vertex to manipulate the spine via the besier curve of the spline. I added a spring controller to the tail to get automatic jiggle too.

All that stuff about different kinds of rotations is pretty complex maths stuff. I don't fully understand it myself but basically the Euler rotation controller gives you the 3 axis of rotation as seperate besier tracks which is far better for character animation. besier is the controller that gives you curves with handles in the trackview. the primary way of manipulating animation in CG.

---

thanks man yeah i have a decent idea of what bezier curves are about...i dealt with them in splines, photoshop and trackview to a small extent, but ok, now i get the gist of what you said about euler rotations

thanks for clarifying those points. so the general idea is to have some degree of control over the way the foot bends, at certain points, and not just relying solely on IK-solvers to work things out for you. i don't think you'll need to make a walk cycle vid man it was pretty clear what you were saying so you basically use the IK solvers to get the whole leg in position, and then fine-tune it using your foot rig? and for the rest of the rig, you use FK.
i will try and get that level of control over the foot by experimenting with different foot rigs, and links, and if none of them work out, i will use your technique

---

Cogito, Ergo Sum

cool, good luck with the rig. look forward to seeing him animated! which reminds me I should really do some work on my caveboy walk cycle.

---

erm just one more thing man....
in your 2nd video, (part 2), you set up all the linkages between IK chains and dummy helpers while the bones are orthographic (as in you haven't rotated any of them at that point - in the Left viewport you can only see straight lines representing the bones which means they are perfectly orthographic)... and so you align your dummy helpers with the IK goals. so when you demonstrate the rotation of the ball helper (at 3:38 - 3:45 in the video), the heel rotates well. but later on when you're fitting the bones to your mesh, you rotate some of the foot bones, and thereby even the IK chains get rotated (5:42 - 5:51).....doesn't this put the dummy helpers out of place? and doesn't this mean you'll have to re-align the dummy helpers with the IK goals, and then from that point on, they're no longer orthographic . you'll have to switch to a Local co-ordinate system in order to rotate the ball helper, because if you just rotate it from a Left viewport with View co-ordinate system, the toe bone starts rotating weirdly out of place.
i hope i've been able to explain this properly.....

---

Cogito, Ergo Sum

yeah its a flaw. I should have aligned the helper objects to the joints before linking everything up. It should work if you move all the bones into place then align the helpers link it all up and freeze the transforms on the control objects, then animate everything in local space just as you would a hand or finger joints. Sorry about that. I'll try and review that procedure tomorrow and just make sure its all sound.

---

no problem man it's just a small change. your revised procedure sounds fine

---

Cogito, Ergo Sum

i'm encountering this problem where, after rotating and positioning the bones into place, the IK goals still remain orientated (?) in their earlier orthographic space...i even tried deleting all IK chains and helpers and re-doing it, but the IK's seem to not get the swivelling plane right.
will try and work out a procedure from the earlier orthographic state to the re-oriented, re-positioned state....

---

Cogito, Ergo Sum

It doesnt really matter what the local rotations are for the IK goals. they only ever move and subsequently rotate the bones. whats important is that the orientation of the bones in max. The pivot point of the bone is the key factor. the bone itself is just a visual guide to the orientation of the pivot point. If your control objects are aligned to the bones rotation when you link it all up everything should transform and rotate as it should. Another option would be to make all the linkages when its straight and only use the control objects to rotate everything into place, or, Yet another option to avoid this on the foot controls would be to create the foot bones straight and only move the knee into place, as in make your model have his feet pointing straight ahead.

---

i didn't know that the orientation of the IK goals doesn't matter. that's good news! i actually tried using only the control objects to rotate things in place, while the IKs were active, but it was a bit of an iffy process.
i managed to solve it the realllly long way round, by getting my perspective view as parallel as possible to the leg's orientation, creating a camera from it, drawing tape helpers as guides/axes, and manually rotating each IK's goal to fit the orientation of the leg as best i could. i then aligned the dummy helpers and now everything's good
but in future will just align my controls to the bones like you said!

---

Cogito, Ergo Sum

If you did want to align the rotation of the IK goals to the joints you could just use the align tool and check all the rotation axis. But yeah It doesnt really make any difference. If you rotate the goals they have no effect. The only other thing to mention about the IK is that you may want to create a control for the swivel angle of the heel IK. so that you can swivel the knee. Normally I would use wire parameters to connect the swivel angle on the IK goal to a control object's rotation. wiring is really easy. just select an object>right click>wire parameters>find the parameter you want to connect>click then click the other object> find the parameter you want to connect>then the parameter wiring box opens and you can make a one way or two way connection.. you could then align the control object to the knee and use a position constraint to stick it to the knee joint. then its easy to be animating the feet and then just pick the knee and rotate the control object to rotate the solver plane of the IK. if that makes sense.

---

when you said "wire parameters to connect swivel angle of IK goal to the control object's rotation", which IK goal did you mean? the one on the first IK chain (thigh to ankle) or the second (ankle to heel) ? i'm guessing you meant the first, cos it's closer to the knee joint ?
will try that out now, have never wired parameters together, so it should be fun

---

Cogito, Ergo Sum

ok swivel angle is a single value, whereas the control object's rotation has x, y and z rotation parameters. which of these 3 do i wire to the swivel angle?
edit: i'm guessing whichever rotation axis is parallel to the ground should be wired?

---

Cogito, Ergo Sum

yeah, basically its whatever makes most sense to you as you animate. It could be wired to a slider or something (a bit trickier to set up) I just use the z rotation of a dummy so its like rotating the knee the same way as you would turn the head. then position constrain the dummy to the knee joint. (or shin bone) its not orient constrained or parented so its z axis will always point up as it moves with the knee.

---

am still working on the wire parameters man...anyway thought i'd post an updated rig
i haven't figured out where to link the ribcage box yet, so it's just hanging, but other than that it looks pretty cool

---

---

Cogito, Ergo Sum

looking good for sure. i think this rig will be much easier to animate than the biped one. I have found with the more rigs I have built. (not all that many really but.) generally, each bone's shape is quite unimportant. Having very thin bones can actually make the whole process easier when it comes time for skinning, while it is true that the size of the bone effects the initial size of the skin envelope you can always change the envelope size and in most cases you will probably need to turn them off and weight the vertices directly. In the case of the ribcage you would probably be better off just leaving it out and skinning to the spine bones. The thing is that because bones are really just a guide to visually represent a hierachy of joints it can be quite misleading to even think about tham as bones. This is another thing I've noticed in maya. In maya you have joints, there are non renderable shapes between linked (parented) joints but the shapes are only present to show visually how the heirachy is structured.

In fact you could rig a character entirely with point helpers (or in maya "locators" ) any object with a pivot point can be linked into a hierachy and used to drive the mesh. its just about how the mesh's vertices are weighted to each pivot point. With all that said though the point of max's bone fins is to help you see the rotation of the joints but strictly speaking you should really only have control objects and the mesh visible while animating. So I guess what I'm trying to say in a very long winded way is that the main thing you have to worry about is the position of the pivot points and the rotational axis of the bones not so much the volume of the bones themselves.

Ok, sorry theres a sizable rant. you can tell i like the sound of my own keyboard clicking can't you ....

---

haha yes i could tell even from your videos you hammer that keyboard!
but yeah man i think this rig will be a lot easier than the biped, and it's awesome to have so much control over the leg and foot. i think the only thing max is lacking as of now is that IK/FK blend that's there in the biped. why don't they just put that as a parameter for an IK chain? but really i don't even think it's all that necessary, it would be a good option to have around....
i made the bones nice and big to help skin it as close as possible....from the few half-hearted attempts at skinning that i've done (when i skinned the uuka over Bethesda's dog rig) i really didn't like it much, dragging those damn envelopes around, and squinting to find those control vertices and rings.... i was just wishing i knew more of painting weights, etc... just i feel that's just a much more natural way, except that you might unintentionally leave some vertices unpainted or unweighted (?)....
so yeah i'll probably get rid of all the fins, and make the bones thinner, and get rid of the rib box, and just hand paint the whole thing (after i learn how)
lol and don't reduce your response size man, your long-winded rants have always been really useful. now i know that point helpers are called locators in maya

---

Cogito, Ergo Sum

Yeah skinning is probably the least fun thing in the whole program next to uvw unwrapping.
The paint weights tools in maya make it sooo much more intuitive. In max its really a case of being very methodical at first. selecting large areas of vertices and hand weighting them to 1.0 for the closest bone using the weight tool (the little wrench icon) thats why its often good to have your bones very thin so the envelopes have no initial effect. either that or go though and make all your envelopes have zero influence (radius) then check the vertices button at the top and for example select the whole head, open the weight tool select the head bone and hit the preset value 1.0. then you can either select the first loop of the neck and make them .75 then the next loop at .5, the next .25.etc.. or work through all the bones as a first pass weighting the vertices in chunks down the arms legs and body. then you can get out the paint weights tool or use the blend button in the weight tool on selections of vertices to smooth the weihgts out between joints. Also, its usefull to animate the mesh into some poses and effect the weights when the mesh is deformed so you can see whats going on. and remember too that you can use the mirror weights function in skin to copy all the weights from one side to the other so only spend time weighting on one side of the mesh.

---

is there any easy way to make all envelopes influence 0? i'm beginning to hate the envelopes and their dratted influence... am going to start skinning all over again, i just messed up the envelopes in a vain attempt to try and work with them...
will do the initial block phase like you suggested, and as i go along, i'll pick each bone, make the envelopes radius 0, and hand-pick the vertices and weight them 1....
and i'll have to learn that mirror tool within skin modifier.... i'm doing things arbitrarily at this point, and that doesn't bode well at all.

---

Cogito, Ergo Sum

Not that I know of. thats why its usually best to make the bones really thin so the envelopes dont affect anything when they first appear. Plus its important to know that by default there is the check box "weight all vertices" checked in advanced parameters. That one is means that even if the envelope doesnt get all the verts there are no verts with zero influence. they are automatically weighted to the nearest bone. If you are hand weighting it may be better to turn that off and manually weight everything.

The mirror tool is pretty straight forward. press the mirror mode button adjust the mirror offset and threshold until the blue verts match the opposite green verts (no red verts on the opposite side). then hit the paste to blue/green verts button. depending on which way you want to mirror.

---

ok something just occurred to me....i've been working all this while selecting vertices and assigning them weight 1.0 to an appropriate bone nearby...but i've been doing it to the base mesh...won't all of it get screwed once i meshsmooth it?

---

Cogito, Ergo Sum

No, don't worry. You will want to leave your turbosmooth modifier on the top of the stack. mostly you will animate looking at the base mesh with the turbo smooth off in the viewport. If you right click it in the modifier stack select "off in viewport" basically your stack would be something like this from top to bottom.

turbosmooth (off in viewport)
skin morph (for fixing problems with joints collapsing etc.. good for maintaining volume or making muscle bulges) (tricky though)
skin modifier
morpher modifier (facial expressions)
base mesh.

there is also the skin wrap modifier for adding meshes onto an already skinned character. good for driving much higher res geometry such as chains or ornaments that you want to deform with the characters skin.

---

thanks man so on top of the skin modifier, i'll just add the turbosmooth/meshsmooth. and once i'm done blocking out the vertices with 1.0 weights, i'll try the weight blends that you suggested...the 0.75, 0.5, etc as it approaches the border with the next bone's vertices.
i've been mucking about with the weights table also to an extent, it's actually quite useful if you have the patience

---

Cogito, Ergo Sum

a bit of a problem now.....i turned meshsmooth on, and the mesh looks crazy in a few parts. this is because i'd imported my uuka as a .3ds when i began this whole rigging process. that seems to have messed up something (maybe some quads became triangles, or pentagons became quads who knows).... so to rectify it, i ambitiously imported my older, proper uuka mesh into this new scene with the rig, etc.... i tried copying the skin modifier from the new, messed-up mesh to the older one, but it doesn't quite do it. i then tried selecting all vertices in the weights table, and copying them, but apparently that "copy" only lasts till i close the weights table.....not if i want to paste it over to the other uuka's skin....

so, long story short.....i'm back to having a rig, and a mesh, and have to redo the weighting procedure.
in spite of having over 50 saves throughout the uuka creation process.

---

Cogito, Ergo Sum

yeah, that sucks, usually i would say that cutting your losses and starting again is the best course of action for a clean result. I usually check my mesh over and over, make sure its at the right scale and perform a reset xform just prior to any skinning to avoid painfull problems later on. However you could try a few things such as the skin utilities (utilities tab>more) which allow you to extract and import skin weights for swapping between meshes. If you just want to use an existing skin modifier you can also try importing the clean mesh align it to the existing mesh and under the skin modifier in the stack attach it to the existing mesh. then select the first mesh as an element and delete it.
If that doesnt work Just take a deep breath, load up the clean file, do all the testing and welding and reset xform then start again. It sucks but its all good practice.

---

yeah man it's not the skinning that's caused the problem. it's that i was working on the uuka modelling (hair, etc) and i wanted to make a rig in a separate file, and not touch the modelled uuka scene. so i exported the uuka as a .3ds and imported into a separate new scene. that's what caused the problem. i just hadn't realised it all this while as i was working with the base mesh.
i'll try what you said, attaching the original mesh, and deleting the new one.
thankfully, even if it doesn't work, i don't have too much work to redo. just a little touching up of the uuka, and the skin-weights.... if i was a little further down the road, i'd have been in much deeper trouble

---

Cogito, Ergo Sum

for future reference if you are only using max and not wanting to open the model in another package the "save selected" option will save out a new max file with whatever you have selected. Its much better to work with the maxfiles unless you really can't such as when you need to work in zbrush it has to be .obj format etc..

---

yeah man i realise that now...i thought that since .3ds is a proprietary format, that everything would be ok. i should've just used merge scene or something simple.
btw, for that creature of yours, did you link the start tail bone to the pelvis? or did you link the tail spline to the pelvis? because the tailspline path constraint is at 100, and the tail doesnt budge if the tailspline doesn't..... i linked the tailspline to the pelvis, but the tail helpers don't move... do i link those to the pelvis as well??

---

Cogito, Ergo Sum

.3ds is a very limited format.

Avoid it unless you just want to export geometry.

---

Please support us and post a [Link to www.digitalartsfront.com] on your website.

--Matt

Yeah you just link the first point helper that gets created with the spline IK to the hip control or to the root bone.
The rest of them move with the first point helper "point01" then you animate the tail by moving the children point helpers.
or add a few spring controllers to them for automatic secondary motion on the tail.

---

thanks man i've now more or less finished the initial 1.0 weight stage (again!) and am trying to mirror it...
there's a small glitch in that i adjust the mirror plane's offset to align it to the center of the model, and hike up the threshold really far, but the blue vertices don't cover half the model... see the pic below. if i offset the mirror away from the center of the model, so that i somehow get blue half and green half, and mirror it, weird things happen when i try to rotate the head etc......
have you ever encountered this problem?

Edit: in the pic, the mirror is aligned with the center of the model, and you can see the blue verts are less than the green verts....
Edit 2: i know i should've aligned the model's pivot to World center before applying Skin, but i didn't. so can't do that now

---

---

Cogito, Ergo Sum

rough skinning done

---

---

Cogito, Ergo Sum

well it looks like its working pretty well. I can't say whats going on with your mirror plane and vertex weighting without having a look at the file
but as you say if you didnt centre the transform to the mesh then reset xform you can safely assume that is the reason for any weird stuff.
Looks like a pretty good attempt at skin weighting to me. Its not the easiest thing in world.

---

Can we see it in action?

---

[Link to www.3dprevis.com]

am currently doing the weight blending stage, will post a vid as soon as i finish skinning!

---

Cogito, Ergo Sum

at long last. i think this is all the weighting i have the patience for now, but still feedback would be appreciated
i'll move onto materials after this.

Caption: "I need to get new slippers"

---

---

Cogito, Ergo Sum