Welcome to od|forum

Register now to gain access to all of our features. Once registered and logged in, you will be able to contribute to this site by submitting your own content or replying to existing content. You'll be able to customize your profile, receive reputation points as a reward for submitting content, while also communicating with other members via your own private inbox, plus much more! This message will be removed once you have signed in.


  • Content count

  • Joined

  • Last visited

  • Days Won


Community Reputation

21 Excellent

1 Follower

About KarlRichter

  • Rank

Personal Information

  • Name Karl

Recent Profile Visitors

447 profile views
  1. Ahh... For non-Newtonian you may want to look at pressure. You can turn the pressure field into a particle attribute by using the "gas field to particle" gas microsolver. Then use that to drive the viscosity or something. See attached for a quick example. pressure.hip
  2. That would probably work, although I have found the POP collision node to be unreliable with FLIP if your particle count is changing rapidly (like if you have reseed on). You may consider using a sop solver with the "xyz distance" node to make your own collision detection.
  3. If you subtract the velocity of the previous frame from the current frame, that would give you the direction and magnitude of the "force" that affected your sim that frame.
  4. @jurueta benne5 has the right idea. When I need to make a pyro sim go "faster" I use a combination of the solver time scale and the dopnet time scale. The dopnet time scale can make the biggest difference, I have found, but you must also increase the substeps on the pyro solver the proportional amount. Also, when simulating fast moving fluids, i always set my min and max solver substeps to the same value. In your example scene, I was able to make your flames go very fast by setting the dop time scale to 3, and the min and max substeps to 8.
  5. @Shalinar I believe Kishen PJ found the reason your packed primitives change their pivot, and that the intersecting geo issue is not relevant here. By default, Bullet will let geometry that starts with intersections stay intersected until they are affected by other forces in the simulation. Incidentally, you can change this behaviour by creating and setting the "found_overlap" point attribute on your packed prims. I think your real issue here is how you are remapping your high res geo. See the attached example on how to use the dopimport node to drive high-res geo. dop_points_jump_example.v02.hipnc
  6. Interesting problem. You can use a SOP solver with a multisolver node to process the mesh point attributes, you just don't want to affect the P attribute directly. Attached is a solution to your intersect problem using a VDB in a sop solver to dynamically set the targetP attribute. It changes the look of your sim a little, so it still needs tweaking. For more info on "solid object" point attributes, see here: http://www.sidefx.com/docs/houdini/nodes/dop/solidconfigureobject Note that this works well enough that you can almost get by with a single substep. Also it almost completely takes place of the internal collision calculated by the finite element solver, try turning off self collision in the attached example to see what I mean. flask.v1.Odforce.FF.VDB.hiplc
  7. This old, but still relevant, post explains GPU acceleration very well, and has tips on how to optimize your sim for openCL. There is also an attached file that clearly shows the kind of speedup you can expect: https://www.sidefx.com/forum/topic/25234/ With that test scene, i'm easily getting 4x to 5x speed up with Quadro M4000.
  8. You needed to normalize the axis input to the rotate node. I also included a very simple approach using the align node. WTF_fixed.hiplc
  9. If you use a multisolver node, you can affect your cloth object with both a cloth solver and a sop solver. In the sop solver you can process your cloth mesh any way you want, in your case adding animated noise to the "targetstrength" attribute. See attached for an example of how you might set this up. animated_cloth_noise.hipnc
  10. If you add a float point attribute to your target mesh named "targetstrength" (as its now called) it will act as a multiplier to the target strength parameter. This is true for most of the cloth object parameters.
  11. You're almost there. Use a trail sop to calculate the velocity of your surface mesh, then use an attribute transfer to transfer the velocity to your scattered particles. Use that as the source for the whitewater. If you want the particles to float, you will need to make a surface VDB from the water mesh and plug that into the whitewater solver as well.
  12. All you need to do is increase the density of your collision geo. Houdini calculates velocity from the mesh points for collision and you original box only had 8 points. see attached for example. stick_rotation_fixed.hip
  13. Use the whitewater source node with some velocity added to the source particles. See attached for example. whitewater_from_object.hip
  14. The mask needs to be a volume that has a value from 1 to 0. See attached for example. CrownSplash Anim C_mask.hiplc
  15. See attached for example.... vdb_morph_mask.hip