+- 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: Attached Mesh Tearing Problem (/thread-3682.html)
Attached Mesh Tearing Problem - patriot4947 - 06-12-2022
[attachment=1592]Why after I move attached mesh it is tearing?
RE: Attached Mesh Tearing Problem - josemendez - 06-12-2022
(06-12-2022, 02:42 PM)patriot4947 Wrote: Why after I move attached mesh it is tearing?
Hi,
Sorry, I don't know what I'm looking at or what you mean by "tearing" in this context. Could you elaborate, maybe show a before/after moving comparison?
RE: Attached Mesh Tearing Problem - patriot4947 - 06-12-2022
(06-12-2022, 03:00 PM)josemendez Wrote: Hi,before[attachment=1594] after:[attachment=1593]
Sorry, I don't know what I'm looking at or what you mean by "tearing" in this context. Could you elaborate, maybe show a before/after moving comparison?
RE: Attached Mesh Tearing Problem - josemendez - 06-12-2022
(06-12-2022, 03:33 PM)patriot4947 Wrote: before after:
If you move a surface-only softbody fast enough, it can turn inside-out if self-collisions are disabled. Simplest fix that doesn't require to use self-collisions is to sample the volume of the shape as well, so that the particles inside of the shape keep it from turning inside out.
See: http://obi.virtualmethodstudio.com/manual/6.3/softbodysetup.html
RE: Attached Mesh Tearing Problem - patriot4947 - 06-12-2022
If you move a surface-only softbody fast enough, it can turn inside-out if self-collisions are disabled. Simplest fix that doesn't require to use self-collisions is to sample the volume of the shape as well, so that the particles inside of the shape keep it from turning inside out.
See: http://obi.virtualmethodstudio.com/manual/6.3/softbodysetup.html[/quote]
I set voxels to volume sampling then nothing works well.
[attachment=1597]
[attachment=1598]
RE: Attached Mesh Tearing Problem - josemendez - 06-12-2022
(06-12-2022, 07:25 PM)patriot4947 Wrote: If you move a surface-only softbody fast enough, it can turn inside-out if self-collisions are disabled. Simplest fix that doesn't require to use self-collisions is to sample the volume of the shape as well, so that the particles inside of the shape keep it from turning inside out.I set voxels to volume sampling then nothing works well.
See: http://obi.virtualmethodstudio.com/manual/6.3/softbodysetup.html
Just click “bind skin” button in the ObiSoftbodySkinner component.
After regenerating the blueprint you need to bind the skin again since the amount/placement of particles has changed, as explained in the manual/videos.
kind regards,
RE: Attached Mesh Tearing Problem - patriot4947 - 06-12-2022
(06-12-2022, 07:25 PM)patriot4947 Wrote: If you move a surface-only softbody fast enough, it can turn inside-out if self-collisions are disabled. Simplest fix that doesn't require to use self-collisions is to sample the volume of the shape as well, so that the particles inside of the shape keep it from turning inside out.I set voxels to volume sampling then nothing works well.
See: http://obi.virtualmethodstudio.com/manual/6.3/softbodysetup.html
Thank you for your answers.
After bind skin fps decrease under 2fps.
And I have an another problem.When I release another sphere from up to down into the attached sphere, they intertwine as you can see in the image.
[attachment=1599]
[/quote]
RE: Attached Mesh Tearing Problem - josemendez - 06-12-2022
(06-12-2022, 10:06 PM)patriot4947 Wrote: I set voxels to volume sampling then nothing works well.
Thank you for your answers.
After bind skin fps decrease under 2fps.
And I have an another problem.When I release another sphere from up to down into the attached sphere, they intertwine as you can see in the image.
Binding skin has barely any impact on performance, so the issue has absolutely nothing to do with this. +4000 ms/frame clearly indicates a deeper problem.
Could you share a screenshot of your blueprint settings, as well as your solver settings?
Edit: the screenshot shows +150.000 triangles being rendered with just a cube and 2 spheres on screen… how dense are your sphere meshes?
RE: Attached Mesh Tearing Problem - patriot4947 - 07-12-2022
(06-12-2022, 10:34 PM)josemendez Wrote: Binding skin has barely any impact on performance, so the issue has absolutely nothing to do with this. +4000 ms/frame clearly indicates a deeper problem.
Could you share a screenshot of your blueprint settings, as well as your solver settings?
Edit: the screenshot shows +150.000 triangles being rendered with just a cube and 2 spheres on screen… how dense are your sphere meshes?
I solved a little bit fps problem using blueprint volume sampling resolution data on blueprint. The last case tearing and intertwinement(2 sphere is intertwined) problem is continue. The below link contains all your need SSs.(ı couldn't send more than 5 ss due to forum limitation fyi)
https://we.tl/t-85F4cm2j58
Note: I also used default sphere at Unity. All I want to do attach a ball to kinematic rigidbody and the other balls around crash and stick to it(Ball or Kinematic RigidBody) while soft body works on them.(I am imagine like sample ball pool scene(In package sample.) but after they crash to each other they stick to each other.)
RE: Attached Mesh Tearing Problem - josemendez - 07-12-2022
(07-12-2022, 07:12 AM)patriot4947 Wrote: I solved a little bit fps problem using blueprint volume sampling resolution data on blueprint. The last case tearing and intertwinement(2 sphere is intertwined) problem is continue. The below link contains all your need SSs.(ı couldn't send more than 5 ss due to forum limitation fyi)
Hi,
Two of the screenshots are the same, none show the actual blueprint generation settings (volume/surface, resolutions, etc) and none shows the solver settings which is what I asked for, so they weren't of any use, sorry.
If reducing the blueprint's resolution worked for you, and there were 150k triangles being rendered, I can only guess there was a huge amount of particles in there due to high resolution volume sampling and that was killing performance.
(07-12-2022, 07:12 AM)patriot4947 Wrote: The last case tearing and intertwinement(2 sphere is intertwined) problem is continue.
"Tearing" should have completely disappeared when using volume sampling, not sure what you're referring to. About intertwinement, make sure:
- that particle collisions are enabled in the solver, and there's enough iterations spent on them
- that collision filters are set up properly (in the blueprint editor, select all particles and check that their collide with property is set to "everything").
(07-12-2022, 07:12 AM)patriot4947 Wrote: Note: I also used default sphere at Unity. All I want to do attach a ball to kinematic rigidbody and the other balls around crash and stick to it(Ball or Kinematic RigidBody) while soft body works on them.(I am imagine like sample ball pool scene(In package sample.) but after they crash to each other they stick to each other.)
Okey, so attachments will only help when attaching a rigidbody to a softbody, but not with glueing softbodies to each other. This can only be done using constraints (specifically, stitch constraints), which is a very advanced use case.
The manual contains a page that explains how to create constraints at runtime, but you have to be familiar with C# scripting and vector math to use it:
http://obi.virtualmethodstudio.com/manual/6.3/scriptingconstraints.html
kind regards,