View Full Version : What makes a good rig? Masters Project Survey

07 July 2007, 04:19 PM
Hi all!

I'm currently doing my Masters dissertation at the University of Teesside. The subject of my dissertation is character rigging in Maya, and as part of the documentation/report, I would like the opinions of professional Character TD's about what makes a good rig?

Also, how many of you create custom control windows for your rigs? A window like this is also part of my dissertation, and I would like to know your opinions about these as well.

Please either reply here, or send me your thoughts to malcolm_kenworthy (at) hotmail (dot) com, with a subject line of Rigging Dissertation.

Many thanks for your time in advance guys and girls!

07 July 2007, 04:22 PM
The right balance of manual and automated controls, flexibilty and very fast viewport interaction. No counter rotation - very very fast understanding of the rig.

07 July 2007, 07:05 PM
wo, eek, you've summed it up brilliantly!!

check out jason schleifer's blog/forum, I remember he was asking people what they think would be hot to have in a rig

07 July 2007, 07:17 AM
The definition of a good rig is actually very simple: That the rig does what it needs to do with as much ease as possible for the animators. Simple but smart!

Deformations/finishing is also very important...

07 July 2007, 03:40 AM
The definition of a good rig is actually very simple: That the rig does what it needs to do with as much ease as possible for the animators. Simple but smart!

hehe, some of the simple things are the least simple :D

I think every animator has a slightly different version of what an ideal rig is anyway. And a clean and efficient interface is really hard to build. Ideally though, the animator should be able to easily show and hide the controllers he/she needs to work with. If I'm working on ruff body animation, I don't want to have facial and fingers controllers all over the scene, hehe. and then, when I work on fingers, because I worked layered, I only want to see finger controllers, nothing else. and if I do a separate pass on the tongue, I only want the tongue controllers to be visible/selectable. other people work less layered and they might need everything... that must be the hardest thing to setup, interface-wise.

Btw, I've seen situations like... one controller doing too many things - in facial setup. I think it's best if the rigger is also a little bit of an animator too, so he/she can realize what's animator-friendly. For the rigger it may look cool to have one controller doing many things... clean interface, right? But in fact it's hard to use in practice. The animator probably needs precision with that controller, and what seems easy to use when you're just moving things around to test deformations, can work very differently in the actual process of animating. For example, I need a controller that I can use easily to open and close the mouth, and only that. It might look cool to add more functionality, but in fact, I work with that one controller hours on end, and having to be careful not to touch the other function/s of that controller is counterproductive. So it all depends finally on the animator's workflow.

07 July 2007, 06:34 AM
Some points that come to mind for me when it comes to rigging:

1. Modularity - The rig has to be broken up in logical pieces so that modification and troubleshooting is easy, should the need arise for either. Organized hierarchies help a lot in this respect.

2. Consistency - This comes into play if you have multiple rigs of, say, biped character. You'd want the controllers and setup of each to be as identical as possible to each other, so you wouldn't have to relearn an entire rig before you could modify or fix it. This also helps when doing modifications to referencing source files. This include everything from naming conventions to hierarchy conventions.

3. Ease of transfer animation - You ought to have provisions of being able to easily transfer your animation in part or in whole to another version of your character, or another simillar character. This is especially helpful when you want to deal with animation libraries.

4. The capacity for local/global ik/fk controls - Although not necessary in all cases, it helps a great deal if limbed characters have the option to switch between different follow states for different scenarios, for example, hands rested on an object as opposed to hands following the body in a walk cycle. Some animators prefer forward kinematics in certain instances, some inverse kinematics. A rig has to be accesible in this regard. Also, the switching between systems should have provisions for a gradual transition, say, if you want to switch from IK to FK your rig animation shouldn't pop.

5. Avoidance of counter animation or double animation - If you have to fight against your controllers to achieve a certain pose, the control loses all meaning. Rigs should be kept accesible enough so that automatic setups can be turned on and off at the disgression of the animator.

6. Automation when necessary - Some instances of this would be a wing flap for a butterfly, swim attribute for fish, or the automatic raising of the clavicle joint when you raise your elbow above the shoulder line. A rig can be built in such a way that expected animation is taken into account.

7. Keeping controller numbers to a bare minimum - Provide controllers and attributes only to whatever is useful for animation, and avoid any additional controllers that will eat up the time it takes to animate. Think of the most efficient rig as having the least number of controllers for the animator to perform certain tasks like walk/run cycles, facial animation and such. Try to keep what contollers you need, and ace any superflous controllers or transform attributes.

8. Rig for the expected range-of-motion or character action - Try to keep your rigging provisions in-line with what is actually needed of your character. The expected performance of a character is what should determine the direction of the character setup. This not only keeps your rigs useful, but also minimizes on the time it takes you to rig things that may not actually be used in animation.

07 July 2007, 09:34 AM
IMHO, one of the most important aspects is the animator himself and his/her workflow. TD should collaborate as much as possible with the animator while creating a char rig. Not to depend and rely on animator's wishes, but have them in mind all the way.
Automatic stuff is great, but semi-automatic is even better! Always, always create control that can turn off automaticity on some rig parts, in case the animator wants complete control.
Also, as isoparmB said, ease of animation transfer is a very important thing, especially because it comes in play late in the workflow. So, not having it in mind, can cause pain later.
Personally, I prefer the approach of having a 'bind' rig on which the mesh is, and separate 'control rig(s)' which will influence the bind joints, through scripting,or point and orient constraints.

Hi all!
Also, how many of you create custom control windows for your rigs? A window like this is also part of my dissertation, and I would like to know your opinions about these as well.

Depends on the job, of course, but being an animator myself, I prefer to create them. But keeping them simple, maybe just for selecting/displaying controls, or for changing the smoothness of the mesh,for example, stuff like that.

Hope it helps, and good luck with your Masters dissertation!

07 July 2007, 11:32 AM
Thanks for all of your replies so far. They have all been extremely useful to me. Also, thanks for pointing me to Jason Schleifer's forum. I've had some great replies there as well.

Anybody else have a view on this matter?

Thanks again for your time!

CGTalk Moderation
07 July 2007, 11:32 AM
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.