Newbee: How to group colliding points of a RBD object?

   4498   7   1
User Avatar
Member
39 posts
Joined: 7月 2010
Offline
Hello forum!
I'm very new to Houdini and this is my first post here. I'm doing my first
steps with dop networks and just wonder how I can store colliding points in
a group to use that information anywhere else in the network.

Is there a node wich can help me out?


best regards, Seb.
User Avatar
Member
112 posts
Joined: 2月 2010
Offline
Hi !
I tried it a lot but it didn't work . Uploading the hip files hoping that someone will point out my mistake.

1)
I was studying sopsolver and came across this in the help

By using an expression like stamps(“../OUT”, “DATAPATH”, “../.bjname/Geometry”) in an Object Merge SOP, the output of the previous timestep can be used as the starting point for the next timestep within the SOP Network.

So I thought we can achieve what you want through the sopsolver (file attached) I was transferring the colour from grid to the sphere by attribute transfer sop and feeding the frame back to the input by using the stmps expression hoping that part of sphere which has collided with grid will get coloured and then I can group it.

Can someone please point out my mistake? Is it possible this way?

2)
Also tried using rop output driver.

Imported the geometry from dop (sphere colliding with the grid, File attached)
Applied colour to both the geometries , used attribute transfer to transfer the colour from grid to sphere and then used rop output driver to render out .bgeo sequence. Each frame is then fed back to the input this time used switch operator to feed in the rendered sequence. Hoping that part of sphere which has collided with grid will get coloured and then I can group it. ( not very accurate way to get the pts which are collided, still)
Even this didnt work

Please let me know if there is any other solution to this problem

Please help :cry:

Attachments:
grup pts of dynamic obj collided.hipnc (143.1 KB)
grup pts of dynamic obj collided rop.hipnc (115.4 KB)

User Avatar
Member
112 posts
Joined: 2月 2010
Offline


Figured out one mistake. (Using rop output driver)
The geometry to which transferring attribute should be stationary. That is If I transfer the attribute from sphere to to grid , will get the area on the grid (attached the file)which has come in contact with sphere but not that of the sphere.

Can we get the area of sphere which has come in contact with grid by applying any change in the network?

Attachments:
grup pts of dynamic obj collided rop reverse.hipnc (115.5 KB)

User Avatar
Member
33 posts
Joined: 5月 2006
Offline
I tried this, but I have an offset. Anyways checkout the the file

Khermah

——————————————————————
http://www.youtube.com/user/Khermah [youtube.com]
——————————————————————

Attachments:
Group Collided Points.hip (130.3 KB)

User Avatar
Member
112 posts
Joined: 2月 2010
Offline
Thanks khermah for uploading the file.
The points are changing from frame to frame . The new collision points on the sphere should be added to the previous collision points on the sphere.
I want all the surface of sphere which has ever come in contact with the collision object. ( what if there are many rbd objects like sphere and the collision object is not stationary)
User Avatar
Member
112 posts
Joined: 2月 2010
Offline
I think I have figured out my mistake and now I can get all the surface of the rbd object which has ever come in the contact with other rbd object using sopsolver and can group it (file, sopsolver group collided pts.hipnc)

It is working on multiple rbds objects as well (like rbd point objects, you can see the surface of sphere being colored red, turn on the wire frame mode) but in that case machine becomes extremely slow(even if there are only 2 simple rbd point objects , low res simple spheres ) (file , test.hipnc)

Please let me know if there is any other efficent way…

Thanks

Attachments:
test.hipnc (242.0 KB)
sop solver group collided pts.hipnc (138.2 KB)

User Avatar
Member
33 posts
Joined: 5月 2006
Offline
I don't know whether this helps. I did it with 15 RBDs and was fast……

Attachments:
15RBDs 1_SecPerFrame.hipnc (176.4 KB)

User Avatar
Member
112 posts
Joined: 2月 2010
Offline
So using point values for center of mass and density did the trick (rbd point object > physical ). Even overriding gives you the same effect. As one can guess ‘Add to the existing value’ needs more calculations and so the the machine becomes slow (?)

Thanks Khermah.
  • Quick Links