Soft bodies for animating a tail

[chriscurra]

I am having a really hard time making my tail move in maya. I am trying to use an ik spline curve and making it a soft body, but I really dont know what I am doing. Can anyone help?

[mantrid]

OK, well, there are many ways to animate tails. Personally, I would never use a soft body IK spline...reason being, you have absolutely no control over it and the result is that it becomes a secondary animation...but in reality, the tail does not dangle off of creatures, it is used and controlled for their ballance. So in otherwords, a soft body tail will look cr@p and not at all realistic, and you certainly can't pose it as you would like for specific things.

But anyways, if you insist on using IK spline soft body, you need to layer up 3 or 4 sets of springs in order for it to move "...ahem...corretly" without stretching. What you would be better off doing however, is to make your own simpler curve with maybe four control points and use that as the IK spline, and animate it directly.

Other ways of animating the tail are to link the rotations of joints every so many joints, and say, you have linked the first 3, put a controller in like a poly object or something, that you use to drive to animate the rotation of the next joint, which is in turn driving the rotations of the next 3 joints, and so on and so forth.

You can also do it by controlling just the first joint and giving the tail a whipping flowing motion via an expression that makes joint2 follow the rotation of joint1 with an offset of so many frames...etc etc.

Generally speaking, tails are not easy to animate convincingly, but its down to personal preference and practoce really. Never the less, softies are not the way to go, usually, because you just don't control em.

[nonquad]

I have developed something of a reputation as a rigger at my school. And so I get questions. The other day, I get:

"Hey john how would you rig a tail?"
Well I use soft splines...
"Yo Chris, he says to use soft splines."

So I guess I got this thing started...
And now you're making it sound like I don't know how to rig. Can't have that.

Well allow me to rebutt!

Soft splines would suck if you weren't allowed to control the softbody goal. But that's not the case.

Done properly, you put the softbody on top of a clustered spline controlled by keyable handles, and tie the softbody lag controls to channels on a pickable control object. You get all of the advantages of the "simpler curve" you proposed, with the softbodies as a second layer.

So the softbodies don't usurp control of the tail; they just provide a way to fill out the motion of the tail, giving smooth, dampened casual motion with minimal keyframes. When you need more spark, you just key the softbody down or off. I would take this setup over a following expression any day.

It's a good system. I got most of it from Steve Cody... very sharp guy. Dunno where he got it. Probably here =p

[rivard]

I accede with pootang (n.b rare thing) and his little ballance theory.

However regardless the rig is quite interesting and can be used for other things, i.e bristles for a toothbrush etc. so here is.

-Create your joints, and IK spline.

-Make the curve soft with a hidden goal.

-Select the hidden curve (say it's called curve1) and open the hypershade, press on input-output connections.

(this code does it well, you could assign it as a hotkey

tearOffPanel "Hypershade" "hyperShadePanel" true;
hyperShadePanelMenuCommand("hyperShadePanel1", "showBottomTabsOnly");
hyperShadePanelGraphCommand("hyperShadePanel1", "showUpAndDownstream");

-You'll see a connection between curve1 and ikHandle1, break that connection and plug the soft curve in the IK's inCurve

connectAttr copyOfcurveShape1.worldSpace[0] ikHandle1.inCurve

(I think you can only do it with MEL, tip: when you delete the connection press Undo so you get the code popup "Undo: disconnectCallback....."

-finally grab your particles in component mode and open the component editor, and tweak the goalPP.

And that's it! What's nice about this rig is that the particles can collide with objects which allows you to do all sorts of crazy little things.

Hope this helps.

[mantrid]

"...a reputation as a rigger at my school"

...hmmm, yes...I suppose your rigging experience at your school is just totally superior to my years of experience as an animator for big budget hollywood movies!

Your idea is interesting and certainly, like all ideas, worth a try...I'm just trying to think why on earth you would want soft body splinIK at all if you are using control points or cluster controls!...all you'll achieve is either nothing at all or funny wriggly jiggly sections of tail....wierd. I think if you meant about getting the volume of the tail to move as though it has some skin/muscle/fat weighting in it, then it is obvious to use soft body...on the tail itself, not on the IK spline. Your idea is a little bizare to me but I am interested to know what you are getting at...kindly explain in further depth as to the motion you expect to achieve?

Anyways, I stress that with a tail, the idea is not to generate secondary motion, but rather to animate the tail on its own. It is a creature's tool, not some floppy dangly bit of tissue hanging off the back. You need to be able to both pose it, and animate it in a flowing movements. A creature uses a tail for balance and counter-weighting. Try to remember that when you are rigging it and animating, else the animation will simply look cr@p and unprofessional, as every animator who watches it will certainly notice instantly.

[nonquad]

Bah. I took things personally which shouldn't have been, to no good end. And of course I defer to your authority Pootang; not from the size of your laurels but from the good advice I have seen you give on countless occasions.

My appologies.

However regarding the question at hand; if you haven't tried a soft spline, then perhaps you should give it a try. I doubt you could give it such a harsh review, if you had used one in person.

To that end, check it out for yourself:
http://mywebpages.comcast.net/jstewart912/tails3.zip

Granted, the rig has its limitations, but in general I believe you will find that your negative expectations were unwarranted.

[mantrid]

I have used soft body splines for many things and see no use of it for a tail. Anyways, moving on, I am glad you see it that way, and don't be so humble, I'm no cg god like other people on this forum, just a fairly average animator. None the less, you have shown much integrity, and let that be that. Try to remember, we are just making pretty pictures and thats about the size of it. These forums are here so we can help eachother out, not to see who's got the biggest schlong ...which I believe is an award that will always belong to Gmask, amen.

[nonquad]

Yup.

You know you're a 3d geek when your best pickup line involves a pole vector and soft/rigid body dynamics.

Oh and don't miss the rig ^^^. I'm interested in your evaluation.

Also, in response to Shydo; MEL isn't necessary for this. You can just select the soft curve directly for the IKhandle tool (without the 'create curve' option). That's what I tend to do in these situations; create the curve first (in position) and create the joints in a clean row along a major axis -- rather than starting with the joints.

But that's just a personal preference.

[mantrid]

lol...not to mention when you get feverish over the mention of luid dynamics and any process begginning with the phrase realtime, and when you have spent so much of your spare time as well as professional time digging into 3D that you actually become aware that not a single term of 3D-speak is actually recognised in any dictionary ...I sware right...try to find realtime or raytrace in the dictionary....it just ain't there ...I suppose both these examples began life as two words, but hey, I was playing scrabble earlier...not cos I'm that much of a geek, but because I was having coffee with Mum...and I had a bunch of words I wanted to use that wre all from

[joojaa]

Potang were you having a bad day there for a monment cos this post and a few others seemed to hint me something like that

[mantrid]

...well, yeah lol

[chriscurra]

You are the man for rigging and I want to see where you stand as far as what kind of character rigs you perfer over others..basically i need to know the best way to rig a character..and you have been doing it forever and I would do exactly as you say master
email me chriscurra@hotmail.com

[nonquad]

yeesh ... classmates messin with me.

There's no rigging class here, so you basically get celebrity status if you can troubleshoot a blendshape. Must try to avoid ego trips =/
gack.

[mantrid]

....who?...me? If you mean me, I am very flattered lol I don't do too much rigging, its not main job, I just do lots of it in my spare time. In film work, there is always a person or 3 who are true rigging specialists...character TD's. But as animators, we learn lots all the time through using and modifying the TD's rigs. I am of course happy to help. Main thing to always bare in mind though, is that character the best character rigs are not necessarily the most sophisticated. The simpler they are, to use, the more practical they are to a production environment. Its making them simple that can get sophisticated, but I always reccommend using work-arounds and cheat methods...its often faster and bug free. The key to making good rigs is to test them really thoroughly and propperly make sure there are no bugs.

[chriscurra]

well my problem was i was having a hard time doing arms and legs which is stupid. But was good at like face stuff. so i found a couple of auto rig programs and let those work and then just added on the facial controlers i made from scratch..what do you think about auto rig programs for maya. the one in perticular I am using is the setup machine.
chris

[mantrid]

I am not a fan of auto rigging. How can any automatic default rig be suited to your unique character?? You may not be able to build specific specialist controls in cos of the way the rig is designed. You may only find out your idea isn't working till its too late, you just can't predict these things unless you design the rig yourself, or have a good rigging td on the job. You can't predict the things the rig won't do, cos you are unaware of the rules it followed in creation. And how can you animate the character efficiently and in the right time scale if you are unaware of the mechanics in the rig?...perhaps you'll find out during animation that joints are not oriented to suit the movements, etc etc, and to fix things at that stage gets really messy.

Its all very well having a quick rig so you can get on with the animtion (if your not a good rigger), but your animation will suffer as a result of a less then custom-designed rig unsuited to your UNIQUE character. Perhaps if you spend time learning just a little more about good rigging, you might find that you are a much better animator than you thought, as a result

[nonquad]

Yup. You always learn more doing it hands-on.

But there is something to say for automation; it does ensure a certain level of standardization -- which is nice if you're dealing with many characters and many animators. In those situations it's nice to have everything as interchangable as possible.

I think you could narrow it down to this: sometimes it makes sense to use a generic rig, but only if you fully understand what the rig does, so that it isn't going to surprise you, and so that you can be confident modifying it if necessary.
The only way to get to that point, is either to script the rig yourself, or get good enough at rigging that you at least understand everything it does.

Either way, the place to start is rigging by hand.

[mantrid]

Yes I agree. If you have a ton of characters to do, you could as you say standardise the rig, and especially considdering the rigs will all be pretty much the same for all characters and therefore easy for people to switch between. But also as you say, you would still have to create the basic rig for duplication/automation of creation of standard components yourself, as opposed to using someone else's auto rigger that you are not gonna know some of the problems with till its too late, and also, you need to fully understand it. Agreed, and if you did it that way, it certainly has merrit IMHO.