Jump to content


  • Posts

  • Joined

  • Last visited

  • Days Won


Everything posted by Atom

  1. Try focusing on the rendered output, not the viewport. You could try VDB smoothing along the axis of projection or supply a velocity field for volume based motion blur.
  2. You can use Labs InstantMeshes. Sometime, a simple subdivide with one recursion level works.
  3. That piece of code is assigning a 1 or a 0 to a group called "selected" The result of check>=0 will return a zero or a one. If you read the documentation on the intersect function you'll see this... Basically, if you get a -1, it won't be added to the group.
  4. That does help, but I'm still playing around with the OP's hip file and can't seem to eliminate the edge. What input min/max values are you using inside the FIT nodes? I assume the fit remaps to 0-1 to drive the ramps?
  5. I played around with that and got some help on a Discord. It's hard to see in the video, but inside the VOP blending the @vel, there are two types of volume samples. The top one is a volumesamplev (used to sample vectors such as velocity) and the bottom one is a volumesample (used to sample floats such as density). Try swapping your top volumesample node with a volumesamplev.
  6. I don't think your GeometryWrangle is not working as intended. Nothing ever shows up in group B, so that force has no effect. It looks like you're expected a 1:1 match on @ptnum from your various input, which are also skewed. Remember, input 0 is the points you're running over. Try moving your fetch slots down 1 so your code reviews from input 1 and 2, not 0,1. You may have to match point numbers by fetching nearpoints on your inputs instead of a direct fetch from @ptnum.
  7. Group the points by bounds, before you copy to points, then break off a network line and add a delete node that references the points in that group. Choose Delete Non- Selected, then have your font node reference the result of that delete instead. Which will be all the points in the box. You'd have to animate the box on the group node, however.
  8. You can use the hScript npoints() function for that. Make sure to surround the function with the backticks so the font node will evaluate the function, and not display it.
  9. Check out the nearpoint vex function. https://www.sidefx.com/docs/houdini/vex/functions/nearpoint.html
  10. It looks like a custom function inside the Vellum Object. Unlock those HDAs and poke around. You could try taking the code and placing it inside a GeometryWrangle running over primitives bound to the constraint geometry.
  11. You don't have to use a setpointattrib on input 0, because you're already running over those points. You can assign the attribute directly. i@start = @neigh[1];
  12. Here's an attempt at that setup using only groups. ap_some_particles_follow_some_curves.hiplc
  13. I don't think PNG files support floating point values. 16b Fixed works for me, though.
  14. Alembic files typically contain "packed" primitives. This means a single point represents a block of geometry. Try this. Drop a Delete node after the alembic and type the number zero into the group field. Your rig is one of those packed primitives. If it's not point 0, then try typing a 1 into the field. This is a crude way to "solo" objects inside your Alembic. Once you have isolated the point number for the care, change the Delete to non-selected to isolate that car mesh. It may be made up of several points. In the end your delete field might contain several numbers separated by spaces, like 0 3 7 etc... Once you have isolate the meshes, drop down an Unpack node to turn that single point into the actual geometry. You might want to check out this link: https://www.sidefx.com/forum/topic/67731/?page=1#post-287800
  15. Try putting @Time on the phase of the wave inside the chopnet.
  16. Try following the peak node with a point wrangle running over group1, and add this code. float y_boost = fit01(rand(@P),0.1,0.3); @P.y += y_boost;
  17. I don't think it works that way. If you're trying to break up the uniform voronoi pattern, lower the count on the scatter and drop down a point replicate after. Set the point replicate to a small count value and scale the clusters out a bit. This will give you a cluster of points at each fracture section. This helps generate size variations in the final fracture pieces.
  18. Atom


    You can review data types here. https://www.sidefx.com/docs/houdini/vex/lang.html
  19. Start with reviewing Igor's HIP files https://www.sidefx.com/contentlibrary/flip-viscosity-by-attribute/
  20. I had to try building that. Here's my setup. Now if there was a way to get whitewater to show up on those cresting waves..? ap_ocean_tube_040123.hiplc
  21. Sometimes modeling is followed up with simulation. You can submit all those spheres to a DOPNetwork to deintersect them. Another way might be to scatter by density and base pscale off of the density value, making dimmer values smaller. ap_scatter_without_intersection.hiplc
  22. Thanks for that link Sean. I've never played around with the instance tab before. It looks like you can put an ocean on anything, then path deform it later.
  23. What if it's the actual name of the file. Are you naming the images "0.exr" etc? what if you add a prefix before the number? I was using $F3, perhaps adding some zero padding might get you over the hump. I assume you tried the obvious and directly loaded image 7 into the file node.
  24. I can scrub my jpg sequence fine. You're not mentioning your hardware or image sizes. Certain Houdini licenses do have image size limitations. Can you post your troublesome EXR sequence, the first 7 frames..?
  25. Try passing the direct path instead of 0. string handle_path = "op:/obj/geo1/dopnet1:crowdobject1/Geometry"; float len = agentcliplength(handle_path, @id, "walk"); f@attr_len = len;
  • Create New...