multiple locators to control a single object without deformation

multiple locators to control a single object without deformation

Anonymous
Not applicable
3,282 Views
15 Replies
Message 1 of 16

multiple locators to control a single object without deformation

Anonymous
Not applicable

Hey everyone, 

 

This must be a simple issue, but i can't seem to figure this out. Let's say I have 3 or 4 locators from mocap that represent different points on the head of the character. How do I let those locators control translation and rotation of 3d head without deforming it?

 

Thanks,

 

Pavel

0 Likes
Accepted solutions (1)
3,283 Views
15 Replies
Replies (15)
Message 2 of 16

J450NP13
Advisor
Advisor

Set driven key?

0 Likes
Message 3 of 16

Anonymous
Not applicable

can you please elaborate on this? As I understand it there should be a way to translate movement of one locator in regard to the other into rotation for the child object. So the if left locator moves forward and right locator moves backwards it causes the child object to rotate. What I can do it select a set of vertices and create a cluster, which is parent constrained to the locator, but this method creates deformation of the object.

0 Likes
Message 4 of 16

J450NP13
Advisor
Advisor

I can't elaborate on it now...I am at work.

 

I'll leave this here though.

 

https://www.youtube.com/watch?v=kF2CWD63AUQ

0 Likes
Message 5 of 16

Anonymous
Not applicable

Ok, thanks. But I fail to see how this is relevant here. I already have animated locators that move around without rotation. Movement of multiple locators has to be translated into translation and rotation of a single object. This could perhaps be done with a combination of parent, aim and orient constraints. Maybe I'm missing something here?

0 Likes
Message 6 of 16

J450NP13
Advisor
Advisor

The locators are already animated or key framed from the MoCap?

 

You could use those as the driver as you would anyway....then select what you want those to drive and set it up.

0 Likes
Message 7 of 16

saihtam
Collaborator
Collaborator
This can be setup with a different number of setups. joints and IKs constrained to the locators. Aim constraints etc. It really depends how the locators from the mocap are and what they're supposed to represent. Could you possibly share a scene with these locators? Preferably with animation.
- Mathias
0 Likes
Message 8 of 16

Anonymous
Not applicable

here is the scene. There are 3 locators for shoulders and chest and 3 locators for the face. Let's say I'd like to make a mask for the character's face.

0 Likes
Message 9 of 16

saihtam
Collaborator
Collaborator
If you've done the marker setup yourself, then I'd recommend taking some time to think of where to place makers and how to recreate it in Maya.

For this what you can do for a super simple setup is to pointConstraint a locator(or what you want) between left_eyebrow and chin. Then aimConstraint that locator towards forehead.

The thing here is that if the person has moved their eye brow then you will get wrong results. Normally I believe people use hats for this. You can easily use a caps with markers on it. Triangles are king. Google mocap shoots etc or see professional kits that you can buy for some hints.
- Mathias
0 Likes
Message 10 of 16

Anonymous
Not applicable

Thank you. It's called an eyebrow only arbitrary. In fact it's just an upper part of the left side of the forehead. 

So the setup you are suggesting is this: create a locator between the eyebrow and the chin and orient constraint the new locator to the forehead. After that parent the "mask" to that new locator?

I don't think it will allow rotations in all possible directions. Like let's say if you tilt the head forward.

I think you need something like two aim constraints on the same locator, let's say I take the main forehead as the main locator then eyebrow would control the rotation of the y axis and the chin locator would control the rotation in the other two axis. But I can't seem to get two aim constrains on the same object.

0 Likes
Message 11 of 16

saihtam
Collaborator
Collaborator
Accepted solution
No orient constraints. Just one aim.

1. PointConstraint a locator between chin and left_eyebrow
2. aimConstraint that locator to forehead using left_eyebrow as the object up in the aimConstraint.

That should give you a quick and simple setup that will follow all rotations of the head.
- Mathias
Message 12 of 16

Anonymous
Not applicable

I made a mistake in the previous post, I meant to say aim constraint instead of orient. The problem was inside the aim constraint options. That's why it didn't work the first time.

 

thank you very much for all the help. What would be the general outline of a more precise method? 

The methd you suggest works, but precision is not that great. For example on frame 307, the orientation of the object is not the best it can be in regard to the locators.

0 Likes
Message 13 of 16

saihtam
Collaborator
Collaborator

Thinking a bit more, you could try to just pointConstraint the locator to forehead, then using left_eyebrow and chin as aim and object up.

Other than that might do a setup with IK's, don't have the mental capacity to imagine that right now. I do believe also that the quality of your track could be questioned. It seems a bit jumpy. Perhaps try to smooth it a bit or something.

 

EDIT: So the aim would be towards the chin with left_eyebrow being object up

- Mathias
0 Likes
Message 14 of 16

saihtam
Collaborator
Collaborator

I had a look at your maya file(completely overlooked it previously), and did the new setup that I described in my last post. I'd say it works exactly as it should. The problem is just with the accuracy of your track/mocap.

 

EDIT: You should probably run this through motion builder or something similar to tweak the mocap. Mocap is never a magic bullet. It needs tweaking. In motion builder I believe you can specify three points to be a static object and motion builder will solve for this. Again, I question the accuracy of your mocap.

- Mathias
Message 15 of 16

Anonymous
Not applicable

Thank you. You're probably right. I will take a look at the mocap again tomorrow. Thanks again for the advise!

0 Likes
Message 16 of 16

saihtam
Collaborator
Collaborator
Happy to help. Did a quick test by finding the longest and shortest distance between forehead and left_eyeybrow and the shortest dist = 3.975 and the longest dist = 4.762.

So you have a pretty big error margin, and that's not even thinking about rotational offset.

Good luck, cool project though!
- Mathias
0 Likes