View Full Version : a neat forarm setup

03 March 2003, 10:48 PM
hey all, I came up with a forarm setup for my last charactr which i feel is a little different. Ii thought i would share it with you all, and see what you thought.

first i start with a regular ik arm setup. with a forearm bone that rotates the wrist.

then i parent a bone above the elbow, and another below. i try to placethem in an area that would represent the 2 bones of the forearm.

then i parent 2 more bones to the wrist, these to represent the ends of the forearm bones' positions at the wrist.

Thne i create a 3 node camera [create-->cameras-->camera,aim and up]

I point constrain the "up" node to the elbow joint

03 March 2003, 10:51 PM
then, i point constrain the "aim" node to the coresponding wrist joint.

finally, i point constrain the camera to the bone i parented to the forarm

Next, i create a peice of geometry to represent the bone itself, the geometry will be used an influence objects forthe main skin geometry

Then after positioning it so each end is near the bones on the wrist, and forearm, i parent the geometry to the camera itself.

i repeat the whole process for the lower bone on the forearm anf Voila! all i need to do now is make the bone geometry into influence objects and play with the weighting and its done!

here is a video of the setup in action in a character i rigged recently

i like it because i was ablwe to set this up on a character that was aleady skinned, and it took only a few minutes to set up..

what do you fellas think?


03 March 2003, 03:15 AM
intruiging use of a camera :applause:

03 March 2003, 12:24 AM
cool camera trick! :)

03 March 2003, 01:14 AM
thanks fellas, just remember not to render through them! lol


03 March 2003, 06:20 PM
Why did you need to use cameras? Couldn't you just use an aim constraint and set the up vector to Object? Are you doing any twisting of the geometry of the bones themselves? Also, your animation only shows about a +- 45 degree rotation. How well does it hold up with the more natural +- 90 degree rotation?

Michael Duffy

03 March 2003, 07:12 PM
i used cameras because an aim constraint caused weird rotations when i bent the arm in different ways.....with the camera setup, it seemed a lot more stable. and becauce of the "up" node of the camera, the influence objects moved the same way no matter what position the arm is in.

here is a newer wrist rotation video...with new and improved rotation range!

let me know what you think of this one..


03 March 2003, 07:14 AM
Peace yo you rock uh man this is so sweet yeah!

Heres to post counts - yours and mine!


03 March 2003, 04:12 PM

Im new to Maya, and have been seeing a lot of tuts for Wrist setups. As a Messiah user what we do in there to have some nice wrist deformation, is to have 2 Wrist bones, then with the aid of Expressions we make those 2 wrist bones rotates as the hand rotates, one rotates 50% of what the hand rotates and the other (near the elbow) rotates only 25% or less.

To prevent double rotations to the Hand (from the wrist bones) we have the ability to use an expression to "parent" the hand bone to the wrist bone, the Expression is called MOVE TO, its not a common Parent, it will parent the items but only translate the movements not the rotations. I dont know if this is possible with MAYA, and i dont know how to make some kind of expression to simulate that in MAYA, but im a MAYA noob so... :rolleyes:

Anyway, here it is the info, if some iluminated MAYA user wants to try that and share it, i will be apreciated :applause:



Peter Reynolds
03 March 2003, 03:58 AM
Originally posted by sedric
thanks fellas, just remember not to render through them! lol


Then again, if your character is inclined to punch some other character in the face, you might get some interesting "punch in the face CAM" renders from them.

Though you might then have to have a quick word to your character to let them know that violence wont solve their problems...

Actually, on the subject of violence, while I don't encourage it, except perhaps in the study of movement - I thought it was interesting that a 3d modelling fight club had started.

After watching a few photoshop tennis matches, I always thought a 3d animation fight club could be interesting.

2 animators submit their models in a common file format so that both can use each others models.

Each animator then animates a scene of their character inflicting violence upon their oponents character within a time limitation (e.g. 3-5 sec).

After a set number of rounds cgtalk members vote on a winner.

Then the next animator steps up to take on the reigning champion.

Could be pretty funny. Can't you just imagine a Godzilla vs Bambi and Bambi wins?

03 March 2003, 04:08 AM
i dont remember where, but there is a MEL bots club...look for it on the web. ;)


03 March 2003, 05:33 AM
Heya Sedric!

How stable are the wrist up/down and left/right wrist rotations? Do the bones stay in position? Does your set up interfer with an IK/FK switch setup?

Eat your heart out jSchleifer.

-The rusted nut

03 March 2003, 06:47 AM
rusted nut:

it works pretty well IMHO, the "up" node and the "aim " node keep the camera under total control...its really solid. i tried aim constraints and they didnt rotate correctly and caused a lot of tearing in the mesh. overall im really happy with the three node camera, and am brainstorming new uses for it!

the extra bones are parented to my "bound skelaton" but are not actually bound to the mesh. They are just there to provide an anchor for the cameras aim and up if i change between ik or fk the setup remains stable

BTW: I posted this stuff to show everyone interested in rigging something I had tried, and to discuss different ideas on character setup....not to make anyone "eat their hearts out"

thanks for the feedback tho.....cya in class tomorow...teach!


03 March 2003, 06:33 PM
I'm still not following how a camera with the up and aim nodes is any different than an aimConstraint with an up object set (which would give it an up and an aim node as well). Both are simply taking a cross product of two vectors to get a perpendicular vector, putting the perp+aim+up vectors in a matrix, orthonormalizing the matrix, and then calculating the Euler angles from the matrix. Unless the camera is using some other method to decompose the Euler angles. We have found that single chain IK solvers tend to run into less gimbal flip than aimConstraints, so there's a chance that the camera is using the IK math rather than the constraint math.

Are you sure you're setting an up object for the aimConstraint? Any chance of your posting your setup files somewhere so we can take a look at them?

Michael Duffy

03 March 2003, 07:22 PM
there is no aim constraint......each node of the camera is point constrained to the locations i showed above....the aim node acts as a built in "aim constraint" for the camera..but to control the other 2 axis' of rotations, the up node is used.....then the influence objects are parented to the camera.....

ill see if i can make a quick scene file to show you what i mean.

here it is!...hope this helps
this is a new more reliable web space.... interested in hearing what you think of it...


03 March 2003, 07:48 AM
cool idea!!!!!!!
:bounce: :applause: :thumbsup: :love: :wavey: :beer:

03 March 2003, 06:30 PM

Thanks for the file! :) Here is the same arm, setup with an aim constrain instead of the camera setup. Your original rig is still in the file, just hidden. I put some animation on both arms to test their twisting. On the new rig, I just created some nulls under the forearm bone origins, aim constrained them to the target bones at the wrist, set the up vector to be the ElbowJoint object, and then parented in copies of the bone geometry. Did you find any problems with this kind of setup that the camera approach overcame? (

Michael Duffy

CGTalk Moderation
01 January 2006, 06:00 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.