Flip colliding with another cached Flip

   4425   4   1
User Avatar
Member
120 posts
Joined: Jan. 2012
Offline
Hi there,

Im trying to use one Flip sim as collision object in another Flip sim.

So I use Surface field of first sim, as Collision field in second sim.

It kind of works, but a lot of particles still go through it. I can display Collision field in second sim, looks correct. Maybe I should apply it in a different way, inside solver ? Im not so familiar with Gas micro solvers….

Attached HIP and screengrab to get the idea what have I done.

thanks!

Attachments:
splash1.hipnc (1.5 MB)
snapshot2.jpeg (86.1 KB)

Michal Tas Maciejewski @ www.vfxtricks.com
User Avatar
Staff
818 posts
Joined: July 2006
Offline
Your file is very close, just a couple of tweaks:

1) The Particle collision method only collides with other DOP Objects, not collisions brought in via SourceVolume, which directly manipulates the collision field. You need to use MoveOutsideCollision in this case.

2) Your SourceVolume DOP will solve twice per frame since you have the FLIP substeps to 2, but the original sim is only outputting data once per frame. So you're not getting properly interpolated data out of the first simulation for any sub-frames, which degrades the collision quality.

You could substep the entire DOPnet on the first FLIP sim and cache out the half-frames. Or just interpolate the output volumes from the first sim as that's probably good enough for collisions.

The attached takes the second approach, using a WhitewaterSource node to handle the interpolation, as it's already set up to handle that in case you want to substep a Whitewater sim. It also rebuilds the surface SDF within a small distance to help ensure good collisions in the second sim.

Attachments:
splash1_wwsource.hipnc (1.5 MB)

User Avatar
Member
120 posts
Joined: Jan. 2012
Offline
Thank you so much. Im glad i can avoid Micro solvers for now.

Yes, I always forget to check if substeps are matching. Cool trick with new whitewater source node. I already looked into it to see accel/curv emission.

Works like a charm

cheers
Michal Tas Maciejewski @ www.vfxtricks.com
User Avatar
Staff
635 posts
Joined: July 2005
Offline
tas3d
Thank you so much. Im glad i can avoid Micro solvers for now.

Yes, I always forget to check if substeps are matching. Cool trick with new whitewater source node. I already looked into it to see accel/curv emission.

Works like a charm

cheers

Noticed that your hip file was NC. Can you talk about your plans for Houdini fluids at Scanline, if any?
User Avatar
Member
120 posts
Joined: Jan. 2012
Offline
oh, im just playing with H at home, recreating some workflows that we use in 3ds.
Scanline is definitely not interested in using H.
Michal Tas Maciejewski @ www.vfxtricks.com
  • Quick Links