Obi Official Forum
Trying to fix the object - Printable Version

+- Obi Official Forum (https://obi.virtualmethodstudio.com/forum)
+-- Forum: Obi Users Category (https://obi.virtualmethodstudio.com/forum/forum-1.html)
+--- Forum: Obi Softbody (https://obi.virtualmethodstudio.com/forum/forum-12.html)
+--- Thread: Trying to fix the object (/thread-3451.html)



Trying to fix the object - JeffsonYu - 14-05-2022

I'm currently trying to fix the obi softbody object. I want it only occurs deformation rather than transformation on the object. I can't find a good way, how can I do this?


RE: Trying to fix the object - josemendez - 14-05-2022

(14-05-2022, 07:49 AM)JeffsonYu Wrote: I'm currently trying to fix the obi softbody object. I want it only occurs deformation rather than transformation on the object. I can't find a good way, how can I do this?

Hi,

This problem isn’t well defined: the shape of a deformable object cannot be determined using an affine transform. Transforms only make sense for rigid objects.

Think about it this way: if the softbody deforms a lot (for instance, a straight bar bends into a circle), is this purely deformation or is it also translation/rotation? If so, how much of it is translation, how much rotation, and how much deformation?
When determining translation/rotation, what point do you consider to be the center of the object? Its center of mass? The average of all points in it? Some arbitrary point?

There’s no clear answer to any of this, so you must be a lot more specific regarding what “deform” and “transform” mean in your specific use case.

Kind regards,


RE: Trying to fix the object - JeffsonYu - 14-05-2022

(14-05-2022, 08:39 AM)josemendez Wrote: Hi,

This problem isn’t well defined: the shape of a deformable object cannot be determined using an affine transform. Transforms only make sense for rigid objects.

Think about it this way: if the softbody deforms a lot (for instance, a straight bar bends into a circle), is this purely deformation or is it also translation/rotation? If so, how much of it is translation, how much rotation, and how much deformation?
When determining translation/rotation, what point do you consider to be the center of the object? Its center of mass? The average of all points in it? Some arbitrary point?

There’s no clear answer to any of this, so you must be a lot more specific regarding what “deform” and “transform” mean in your specific use case.

Kind regards,
Hello, Thank you for your reply! I'm currently using a hand to grab the obi softbody to obtain several information of grasping, but now it seems that even if I remove the gravity in the obi solver, the object will still move if I attach a force onto it. So I intend to keep the deformation of the surface of softbody, like depression of the barrels in the sample scene, and also fix the object so that it cannot move or rotate because of the force. I hope I made it clear this time... Again thank you for your help!


RE: Trying to fix the object - josemendez - 14-05-2022

(14-05-2022, 09:24 AM)JeffsonYu Wrote: Hello, Thank you for your reply! I'm currently using a hand to grab the obi softbody to obtain several information of grasping, but now it seems that even if I remove the gravity in the obi solver, the object will still move if I attach a force onto it. So I intend to keep the deformation of the surface of softbody, like depression of the barrels in the sample scene, and also fix the object so that it cannot move or rotate because of the force. I hope I made it clear this time... Again thank you for your help!
Hi!

Not much clearer really, what I tried to explain is that “moving” and “rotating” aren’t well defined for deformable objects (unlike for rigid objects). If you squash a cube flat on the floor, has it moved down? Or just deformed? it depends on where its center (pivot) is, right?

I assume that by “center” of the object you mean some arbitrary particle or particles of your choice, in that case you can just use an attachment to keep that particle(s) fixed in place. See:
http://obi.virtualmethodstudio.com/manual/6.3/attachments.html


RE: Trying to fix the object - JeffsonYu - 15-05-2022

(14-05-2022, 09:43 AM)josemendez Wrote: Hi!

Not much clearer really, what I tried to explain is that “moving” and “rotating” aren’t well defined for deformable objects (unlike for rigid objects). If you squash a cube flat on the floor, has it moved down? Or just deformed? it depends on where its center (pivot) is, right?

I assume that by “center” of the object you mean some arbitrary particle or particles of your choice, in that case you can just use an attachment to keep that particle(s) fixed in place. See:
http://obi.virtualmethodstudio.com/manual/6.3/attachments.html

Hi! Sry to bother you again...yesterday I didn't see the message

In fact, in my task, the softbody objects are actually similar to the rigidbody, that the deformation is very small. So what I want to do, is to fix the center of the mass, so that the center wouldn't move from like (0, 0, 0) to (1, 0, 0). 

Again take the barrel in the demo as an example. Now if we use the ball to hit it, it would fly and the surface would deform. Instead of that, I was trying to stick the barrel on the floor, like using a glue, so that when we hit it again, it would only occur deformation on the surface, and stay put. In most cases in my task, the objects are very similar to the barrels.

I would say that rotation would probably not occur in my task, for the hand grasping controls the whole object. I suppose it is similar with fixing the particles? Possibly I will look for the scripts to see if they helps.

Again thx for your reply.


RE: Trying to fix the object - josemendez - 16-05-2022

(15-05-2022, 06:55 AM)JeffsonYu Wrote: Hi! Sry to bother you again...yesterday I didn't see the message

In fact, in my task, the softbody objects are actually similar to the rigidbody, that the deformation is very small. So what I want to do, is to fix the center of the mass, so that the center wouldn't move from like (0, 0, 0) to (1, 0, 0). 

Fixing the center of mass wouldn't do what you want: The center of mass is the mass-weighted average of all particles' positions, so if you fix the center of mass, moving one particle to the right would move all other particles to the left.

I think you just want to attach some particles in place, which is what attachments do.

(15-05-2022, 06:55 AM)JeffsonYu Wrote: Again take the barrel in the demo as an example. Now if we use the ball to hit it, it would fly and the surface would deform. Instead of that, I was trying to stick the barrel on the floor, like using a glue, so that when we hit it again, it would only occur deformation on the surface, and stay put. In most cases in my task, the objects are very similar to the barrels.

To achieve this, you can just use attachments. Check out the RubberDragon sample scene, as it does what you describe here: the bottom part of the dragon is attached in place using an attachment component, so that you can drag to deform it, but it doesn't move around.


RE: Trying to fix the object - JeffsonYu - 17-05-2022

(16-05-2022, 10:11 AM)josemendez Wrote: Fixing the center of mass wouldn't do what you want: The center of mass is the mass-weighted average of all particles' positions, so if you fix the center of mass, moving one particle to the right would move all other particles to the left.

I think you just want to attach some particles in place, which is what attachments do.


To achieve this, you can just use attachments. Check out the RubberDragon sample scene, as it does what you describe here: the bottom part of the dragon is attached in place using an attachment component, so that you can drag to deform it, but it doesn't move around.

Yes! I have used the attachments on the particles on the bottom of the objects! That worked very well! Really appreciate it!