Particles Self Collision

   15016   14   3
User Avatar
Member
72 posts
Joined: March 2014
Offline
We have very cool and useful feature in the Maya for nParticles.
It's Particles Self Collision feature.
In the Maya it's so fast and without any problem or error.

I know Dynamic System of Houdini is so perfect and in some special cases , it's actually better than Maya , But I think we don't have this feature in the Houdini , even with new Particles System ! Right ?
So how can I do that ?
Specially for sands and soils simulations ?
User Avatar
Member
1391 posts
Joined: Dec. 2010
Offline
I don't know what's better way and solutions for now.

But maybe this Topic can help you :

http://forums.odforce.net/topic/8061-inter-particle-collision/ [forums.odforce.net]

Also we have an example project in the Houdini Help about this.
Of course I think in this example we use RigidBodySolver for self collision of the particles , So each particle simulating as different RBD object and I think it's not very fast !!!
Maybe it's good only for learning :wink:
https://www.youtube.com/c/sadjadrabiee [www.youtube.com]
Rabiee.Sadjad@Gmail.Com
User Avatar
Member
72 posts
Joined: March 2014
Offline
Specially thanks for your Help Sadjad.
It's exactly something that I want , But as you said before , it's really slow !
I wanna do that very fast , because I have too many particles for sands !
User Avatar
Member
443 posts
Joined: Sept. 2012
Offline
You can use popproximity which create nearestdist attribute that can be used to trigger popinteract.
User Avatar
Member
72 posts
Joined: March 2014
Offline
PradeepBarua
You can use popproximity which create nearestdist attribute that can be used to trigger popinteract.

Specially Thanks , Yes , That's another way , But I think it's slow too , maybe slower than using Bullet Engine for particles !

I think Sidefx should add this feature in the new Particle System ! :cry:
User Avatar
Member
8509 posts
Joined: July 2007
Online
also note that the technique in popswithrbdcollision example is old and therefore the slowness
to get particles with rbd collisions in >=H13 you can just use packed rbd's which combines power of particles and bullet and works without any special setup
Tomas Slancik
FX Supervisor
Method Studios, NY
User Avatar
Member
72 posts
Joined: March 2014
Offline
tamte
also note that the technique in popswithrbdcollision example is old and therefore the slowness
to get particles with rbd collisions in >=H13 you can just use packed rbd's which combines power of particles and bullet and works without any special setup

Really Thanks , But how can I use RBD Pack ?
I should add new particles in the AutoDopNetwork , then use RBD Fractured Objects ?
User Avatar
Member
443 posts
Joined: Sept. 2012
Offline
Check this expample file:
http://www.sidefx.com/docs/houdini13.0/examples/nodes/dop/rbdpackedobject/EmittingObjects [sidefx.com]
User Avatar
Member
72 posts
Joined: March 2014
Offline
PradeepBarua
Check this expample file:
http://www.sidefx.com/docs/houdini13.0/examples/nodes/dop/rbdpackedobject/EmittingObjects [sidefx.com]

Thanks , I saw this help before , But I don't know how can I use it for Particles !?
User Avatar
Member
443 posts
Joined: Sept. 2012
Offline
You need not to use particle. If you see in detail view Packed RBD Object is just single point. Since these are like points so all the popdop nodes work just like particles. The way you use forces and other operations to popdop, same you can use for packed rbd objects too.
User Avatar
Member
443 posts
Joined: Sept. 2012
Offline
If you want to use particles only then check this:

http://vimeo.com/82039752 [vimeo.com]

In the end of video Ari did something similar.
User Avatar
Member
443 posts
Joined: Sept. 2012
Offline
Here is a video by Matthew Puchala.
https://vimeo.com/111449179 [vimeo.com]

May be he can help you more.
User Avatar
Member
157 posts
Joined: Nov. 2012
Online
Sure, I would love to help. I have been developing a self collision solver that can be implemented as a solver on POPs in DOPs. I am extremely happy with results for simple collisions such as in the first example of my video. My algorithm was designed to handle things such as crowd avoidance, or particle simulations that need volume preservation, and do it quickly. However I recently decided to stress the system under “high density” situations, such as sand or ball bearings in a cup. It dramatically slows down the sim time, and does not preform well at all. From what I observe in testing, my solver does preform the calculations to maintain a rough volume and collide. However, it struggles to find an equilibrium and settle into a rest state. I am still working on it and will post any updates if I get a chance to play with it some more.
Matt Puchala
www.mattpuchala.com
User Avatar
Member
7 posts
Joined: Aug. 2016
Offline
Hi everyone,

I know this is a very old post, but I stumbled across it while looking for a solution.

I actually found that giving your points a @pscale attribute (i.e. via POP Wrangle) right after the POP Source and then using the POP Grains Node with the same seperation (i.e. 0.1) gives you a pretty fast way of getting the same effect as with nParticles.
To visualize it just use a POP Sprite (set to sphere_matte.pic).

I may upload a file when I'm home…

Cheers,
Chris
User Avatar
Member
2 posts
Joined: Jan. 2018
Offline
Thanks Chris, that helped me!
  • Quick Links