PDA

View Full Version : how to parent a wire curve to geometry in maya


denis333
06-07-2008, 06:09 PM
hi,everyone.
i try to using wire tool in maya for my animation.
but,after i parent the curve and cluster to the geometry.when i rotate the geometry,the wire tool works like the image.
could anyone help me?
thanks for your any help.http://b5soww.blu.livefilestore.com/y1pCR2LI-8Fv7TbhT9cFNTuKwl1EWa5tMNjZkE4CxkLnvPiekqdALvxgtgfvsHc8LMShCu9mP2awTCfwQH_KikSAQ/11.jpg\

http://b5soww.blu.livefilestore.com/y1pCR2LI-8Fv7SlmYhrHhW_ZcrmYhVTcvFt5lXyWFFrfMUq-xcgoN_2XCd0Q0985LpmuWV4wtXeD5pKHPxiK-xo5w/22.jpg

FridayImInLove
06-09-2008, 02:05 AM
It's double transformation problem.Try to imagine you use deformation tools to control vertices of geometry(shape) and use geometry(transform) to control your deformation tools.Try to unparent cluster ,wire curve and base wire curve from geometry then group cluster wire and geometry togather (make it same hierachy) and use that group as global control(make sure your cluster is relative).


sorry for my very very terrible english.

SkullboX
06-09-2008, 01:52 PM
The hierarchy looks correct for what you're trying to achieve, although I wouldn't parent anything to the actual model. 'FridayImInLove' (your english is fine by the way) is right about the clusters, they have to be in relative move (select a cluster and check relative in the attribute editor, that should hopefully work).

However, if you're in maya 8.5 it will not work, ever. This is a huge bug in this version apparently fixed by maya 8.5 sp1. We switched over to maya 2008 at school at the same time, where all is well.

Good luck!

denis333
06-09-2008, 04:14 PM
FridayImInLove, SkullboX
Thank you very much for your help.I tryed group the geometry the curve and the wire base.It works super .thank you very much.i try to parent the curve and the wirebase the geomery .it seems works too.

http://www.highend3d.com/maya/tutorials/character/276-3.html
just when i learn this tutorial.He said "parent the clusteur and the curve to greometry.but not parent the the wire bas.

FridayImInLove,
SkullboX is right,your english is fine.better than me.
finalment .thanks for your help.really helpful

Anothe problem on this is ,if i use wire tool in the face of a charactor,how can i parent or group the clusters and the curve.i try to parent the clusters ,curve,wire bas to the skeleton of head,the wire tool works well.but when i rotate the root .the problem is double transformation.

http://b5soww.blu.livefilestore.com/y1pCR2LI-8Fv7QNITfIMRbGTtek2vMZS2jryZg8VPZJd-XoGtsmOltHtsiWdWEbr-oG5_9bIXe5WkLl-VjikFVUyA/di-1.jpg
the charactor
http://b5soww.blu.livefilestore.com/y1pCR2LI-8Fv7S0E_6xV2MEWgjPRSjMYRWotENPVkZer5iwNIop8onbVDuzlrGNTzxoszNgSISFZ21MIrH-Yh4roA/Untitled-2.jpg
wire tool curve and cluster

help!!!

yenvalmar
06-17-2008, 08:42 AM
is just apply the rigged wire face as a blend shape to a face thats skinned onto the character with regular bones. trying to make it so wire rigs can move through space in rigs combined with joints will drive you insane.

so you just let that wire rigged face sit there and use it to animate the expressions on the mesh thats skinned on the character.

this gives you a face control panel thats stationary relative to the char- considered a <high end> rig feature lol. and no possibility of double transforms or anything weird happening.

i actually like wire face rigs a lot but i have only got it to work by not putting the wire rig within my moving character. fWIW.

SkullboX
06-17-2008, 02:11 PM
Anothe problem on this is ,if i use wire tool in the face of a charactor,how can i parent or group the clusters and the curve.i try to parent the clusters ,curve,wire bas to the skeleton of head,the wire tool works well.but when i rotate the root .the problem is double transformation.
Don't parent the base wire to the skeleton, it should remain static in your scene somewhere. I always parent it under some group I never move around. You should parent the clusters and the curve to the head though, but never the basecurves. From the tutorial:
4. Parent the faceCurves and clusters: So that the rig follows the head, you need to parent the curves and clusters under the skull joint so they follow along with the head. Do not parent the baseCurves. These were created when the clusters were formed and shouldn’t move. You can group them and hide them away in the same level as the head mesh (that shouldn’t ever be moved) for organizational reasons.
I've build that rig twice now, and it animates great once everything works. I don't think Maya 7.0 has anyproblems with using curves as deformers.

And no offense yenvalmar but your solution is pretty bad to be honest. The great thing about this curve facial system is that you can control everything intuitively. I actually animated one scene without using any preset expressions, I just animated everything freeform directly using the clusters' controllers. So using a seperate head kind of ruins one of the main advantages of building this rig rather than using blendshapes...

JYoung
06-17-2008, 04:50 PM
Not sure if you're trying to follow along exactly, but that tutorial doesn't use the wire deformer. The curves are added as influences via skin>edit smooth skin>add influence

yenvalmar
06-17-2008, 11:12 PM
And no offense yenvalmar but your solution is pretty bad to be honest. The great thing about this curve facial system is that you can control everything intuitively. I actually animated one scene without using any preset expressions, I just animated everything freeform directly using the clusters' controllers. So using a seperate head kind of ruins one of the main advantages of building this rig rather than using blendshapes...

ok you should pass that on to every rigger that has ever sett up a rig with a facial control panel then. some of us find being able to see our characters face at all times no matter what position they are in without moving the camera is nice. of course you can parent a cam on the characters face in the rig but then its the same thing as what i do.

if you rig it my way you still animate it all by hand with the clusters driving the wire deformer to create the expressions "live". there is nothing in the way i rig to prevent animating the way you want- in fact that is the entire point of my rig. and in fact it still IS a wire deformer solution, technically. just with a little extra step that makes it a billion times more usefull :)

i agree canned blend shapes suck as a facial animation tool. thats not at all what i am doing. having a separate face rigged head driving the head rigged on the body has nothing at all to do with using canned expressions - instead of blend shaping a static morph target the lame old way you are blend shaping RIG targets. the animation is all done live using the rigs as if the rig were on the character itself like the original poster is trying to do and like you do already. get me now? its brilliant.

when you define the blend shape to the rigged face, any animation you put on the rig, will propagate to the final head. the rig stays live, you arent blend shaping a static target. you are blend shaping to an ANIMATED target! so you do your animation using the deformers as you want to. the magic of mayas construction history AT ITS FINEST!

this has so many advantages its hard to imagine, you can do 10 different face rigs using every technique as needed, for the requirements of the shot, and mix them seamlessly. if you want flexibility this is the ultimate.

(mind you it can get confusing to keep track of what rig is doing what if you go too overboard :) )

and none of this has anything to do with a straight ahead vs pre-posed face animation work flow so i say your COMMENT is crappy! :buttrock:

SkullboX
06-18-2008, 07:00 PM
denis333: Yungimoto is right about not using a wire deformer, but rather the option to add cvs to a skin solution. I figured you used the term incorrectly, considering you were following that tutorial.

ok you should pass that on to every rigger that has ever sett up a rig with a facial control panel then.
What does this have to do with using a control panel? Using curves as influences in the way shown in the tutorial has the freeform control as an addition to any control interface, not a replacement. What I prefer personally has nothing to do with the capabilities of the rig, I personally used either just the preset interface, just the freeform controllers or a combination of both.
trying to make it so wire rigs can move through space in rigs combined with joints will drive you insane.

so you just let that wire rigged face sit there and use it to animate the expressions on the mesh thats skinned on the character.
I'm primarily an animator and I couldn't stand having to mess around on a static head if I wanted any freeform controls, if you could still call it that. I want to be able to pose the eye target, rotate the jaw and any other joint controls and then drag control objects all over the place to sculpt the final pose, then tweak and repeat. I don't want to go back and forth between a second head for freeform controls - which is what makes this rig so animator friendly.

Also if your solution is based on a system that avoids combining joint and wire deformations, what deformation information would the static 'freeform' head actually contain? Just the wire deformation, so you couldn't see any jaw/lip/eyelid or any other conceivable facial joint deformation on the static 'freeform' rig? Or would this freeform rig also contain all the rotation of facial joints above the head?
this has so many advantages its hard to imagine, you can do 10 different face rigs using every technique as needed, for the requirements of the shot, and mix them seamlessly. if you want flexibility this is the ultimate.
This would be just as easily possible with the solution in the tutorial, working in exactly as you described, just without actually having to combine any data other than just blend between animated blendshapes. The rig the OP is trying to build can do all of the same yours could, adds actual freeform control, is less complicated to build and last but not least your solution is heavier, if only by a little.

If you can indeed build a rig that works with the same amount of control, with the same freeform capabilities, in the same time, with the same load as the one we're discussing I take back everything I said, but for now either I can't see your solution, or it doesn't work properly.

JYoung
06-18-2008, 08:57 PM
I don't think Ian was saying you'd have separate control heads. The control head would still have all the necessary free-form/on face controls as well as joints/clusters/whatever, only it would remain static. You end up with the exact same control, only you don't have wire deformers attached/following your character rig.

At least, that's what I think he was suggesting.

yenvalmar
06-19-2008, 01:39 AM
apparently some animators dont like control panels for the face- in which case this somewhat comes down to a personal preference. i have never met these animators face to face but its a big internet out there. i have sinceerely learned something from this!

i came about this entire idea by looking at some face rigs that were floating around out there and getting a lot of positive comments about the functionality, freaking out when i saw some of the actual scenes under the hood and saw how complex the rigging was (hey guess thats why you gotta buy the DVDs that come with the rigs normally), and seeing how i could do the same exact functionality (and better even) as they did far more simply.

the exact moment when i found the solution i have been using for years now i was wrestling with exactly the problem the original poster was having, double transforms on a wire rig, and i was also rigging a facial control panel (using much more complex ways of connecting it), so i felt pretty clever when i killed two birds with one stone. so i had to mention it in this thread. if there are still simpler ways to rig, of course there are, just use all blend shapes or all bones. :)

as to free form control the only thing you "sacrifice" with having your facial animation rigged head be positionally static in the scene is actually considered a benefit to most people i have worked with up till now- the ability to always see clearly what the characters face is doing no matter what else is going on in the shot. you tear off a window and set it on the face, and position the face way underground so its not in the way, its super convenient. but maybe the trend is going back the other way and ppl want it all on location now...

since the rigged head is STATIC in space mixing and matching deformers is REALLY SIMPLE so you will have much less problems mixing stuff.. if you really want you can mix wires, bones, clusters, an FFD lattice and soft mods if you want, all on one rig, much more easy when the whole head remains static in space. or, as i have been getting into, make several simple rigs that are TOTALLY DIFFERENT, one for phonemes which uses blend shapes one for expressions using wires, one with joints, each simple. for me personally thats easier than making wires be able to scale, and not double transform, and work with joints etc etc, when they are in a heirarchy of a rigged full body character. YMMV.

also keep in mind that you can (and i have) drive all these different rigs from one unified set of controls, hooked up to the rigs via set driven key, so you dont need to manage 5 rigs to make one animation. in the end you can manage everything from one set of controlls and just see the final result, exactly as if it was a "normal" single uber rig that does everything and can add ability at any time. tell me thats not cool for something using first-year maya tools :)

i realize a loot of people instinctively want to do it all on one rig, but once you stop thinking that way life gets much easier. remember when you tried to make everything work in a single render before you learned about layers and compositing? same thing applies with the face rigs. the main increased complexity is "mental", that you have to just grasp the concept which is not based on "reality" and thus stretches our brains a tiny bit.

a parting shot- (based on something thats happened to me) if you are halfway through the project and the director suddenly decides the realistically rigged characters head has to do something totally insane like the movie the mask in the middle of a shot which already has animated lip synch, and you have a rig which has the character face thats directly rigged on the animation rig using a wire and bone that cant do the deformation desired, you will cry a lot more than i will, cause i can just rig up a new face to do the gag, add it to the blend shape, and animate it, i dont have to revise anything on the lip synch rig EVER really. what would you do? basically by putting in that blend shape early on you buy insurance against any situation. you can start animating with a rig thats good enough for the first shot and not worry about anything coming down the pike. its saved me so much time so many times..

(you could do the same thing and add a new rig as a blend shape too in that situation, but if you set it up with a blend shape from the beginning your final rig will be a lot smoother in performance and you avoid order of operation headaches)

in any case i wish the original poster luck i have made my *opinion* clear and its up to them really! one of the cool things about maya is theres many ways to do the same thing.

SkullboX
06-19-2008, 03:23 AM
i realize a loot of people instinctively want to do it all on one rig, but once you stop thinking that way life gets much easier. remember when you tried to make everything work in a single render before you learned about layers and compositing? same thing applies with the face rigs. the main increased complexity is "mental", that you have to just grasp the concept which is not based on "reality" and thus stretches our brains a tiny bit.
I appreciate the condescending tone, but the method your describing isn't any different from any rigging workflow where different defomers are combined into a single object, whether this is through direct defomers, static or animated blendshapes through other deformers. It's hardly hard to grasp 'your' method of rigging.

As an animator who also rigs I don't really see how you could argue the way you do. The point of the discussed rig is a freeform type of facial control which your case simply doesn't offer to nearly the same degree. While I'm sure you will remain highly convinced that your way of rigging is superior to that in the discussed tutorial, your method fails to deliver the type of desired animation control, and that's what a rig should do. Accommodate the animator, not showcase technical knowhow.

Garciaone
07-07-2008, 01:03 PM
I agree with both, i think the method that yenvalmar talks about is the better way to go when setting up JNT or curve based face rigs, But the actual animation controls can still follow the face of the animated mesh in the same way as the conventional method. So isn't this the best of both worlds , the animator would never know the difference?

zhenyang
07-10-2008, 03:02 PM
I appreciate the condescending tone, but the method your describing isn't any different from any rigging workflow where different defomers are combined into a single object, whether this is through direct defomers, static or animated blendshapes through other deformers. It's hardly hard to grasp 'your' method of rigging.

As an animator who also rigs I don't really see how you could argue the way you do. The point of the discussed rig is a freeform type of facial control which your case simply doesn't offer to nearly the same degree. While I'm sure you will remain highly convinced that your way of rigging is superior to that in the discussed tutorial, your method fails to deliver the type of desired animation control, and that's what a rig should do. Accommodate the animator, not showcase technical knowhow.

I think you understood him wrongly.

What yenvalmar was saying is... you'd have 2 head geos, one that is rigged with whatever method you'd like to use (wire, facial joints, cluster.... etc etc), and another final head that is deformed only by body joints and blendshape (from the first head).

So for your case, you rigged up the "freeform" facial rig on the first head, and apply this head as blendshape to the final head.

:)

CGTalk Moderation
07-10-2008, 03:02 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.