Jump to content


  • Posts

  • Joined

  • Last visited

  • Days Won


TheDude last won the day on February 23 2014

TheDude had the most liked content!

Personal Information

  • Name
  • Location
    Los Angeles

Recent Profile Visitors

4,125 profile views

TheDude's Achievements


Newbie (1/14)



  1. Average proximity would be one way, check how far away your nearby particles are. If the 10 nearest particles to your current point are greater than 1cm away, then split this into a seperate group. mesh only those that were not in this group, and leave the rest as points.
  2. I can't view your attached files, not at a computer. My initial reaction would be to take whatever mesh, convert it to non optimized quad mesh, sample nml, up and cross. Through a single axis along the centroid of the geo progressively push the point along the cross by the percentage it is along that axis. You'll want to run this in a for loop also referencing the original mesh so the ones that are pushed a long way can remain fairly accurate through complex surfaces towards the max end of the axis your pushing geo around.
  3. scatter the new points per frame and set @BirthTime to -1, in a solver, merge these new points with the previous frame, if @BirthTime is not -1, set @BirthTime to @Frame
  4. Sounds cool, to answer your question directly, you could measure distance to collision object with xyzdist, below a certain distance you could then switch its position data to then always stick on the object you found within the distance. To do this, use primuv, and track the location found from xyzdist, and keep your current position to match that parametric uv of the object from when it was within range. As a general comment, I found grains in the past to be awesome for sticky high detail saliva fx!
  5. TheDude


    Put down a pop drag, and keyframe animate the drag as needed. You can also use vex expressions in this if you need to control specific groups / areas / procedural approach based on other data
  6. Hey Chris good to see you here. In that situation what I've done in the past is use a solver, to in this instance, flip a value from 0 to 1 if this colour turns red. If this triggers you mark down the current time. Timeshift to the end of your timeframe, and transfer the "triggerTime" value to whatever you are doing. You now have the time at which your colour will change. Use that value as an offset to your alembic and it will know when to start playing. If you had a bgeo cache and thousands of points, you could use this to be able to know what offset all your animations would need to be on particles
  7. output picture on mantra -- this dictates where your image will save. You are rendering frame 1 - 50 as 1 file, which is shown second from top of your file browser. this also has no extension. make sure to include extension on your output picture (as in .jpg, .exr, .tiff, whatever you want), and i presume you don't want to overwrite every frame into the same file, so add in _$F4 before your extension. $F means current frame number, $F4 means it will always output 4 digits representing current frame number.
  8. I was wondering if OPscript file is no longer supported for setting desired node preferences in your network view? I have also seen that now under Network View Display Options, under Themes tab, it now has pull downs / user made settings for shapes, colors, and wire style. This however saves a pretty gross .json, which is fine if you are not looking to edit this in the text editor (but that was the benefit of OPscript). It also appears that, unless you are going to edit the json through a text editor, you now have to make a new theme whenever you'd like to change your default settings, and this has no way of replacing your current custom theme. These are all changes I don't like D: Is OPscript no longer supported?
  9. To explicitly answer your question: Take your current volume, and place 3 "name" nodes, in each one link it to your current volume and name in the string parm "vel.x", "vel.y", "vel.z" -- place a merge, and combine your original volume, along with the 3 volumes that have been renamed. You'll now have your original volume, and 3 fields for vel. I think with more context though, there will be a better answer.
  10. What bunker recommended will work well. To address why this is happening, I am guessing, is that you have a significantly high heat value around your source and the temperature dissipation is significantly high along with buoyancy. This means you have fast moving voxels in the Y value, carrying high values that will consistently devalue at each step. Solutions being to increase steps, add noise to your source fields with appropriate frequency enough to mess up this obvious banding.
  11. To answer your question specifically, to replace particles with animated geo -- copy your animated geo onto your particles, and if your particles are already looped, make sure your animation cycles of your geometry match the duration of your particle animation and youre done! Keep in mind the fundamental use of the sin wave -- it is super functional for creating loops in all types of animations
  12. There are lots of variables here, so it's hard to speculate. Here's a few things to check. Alpha Consistency from different angles. Near / Far clipping plane on camera. Density consistency in the geo of the file you are using. Check in shader for unusual manipulation of the fields imported.
  13. So once you've got your points on your deforming geometry, you can then inherit your UVs from the object, and displace by the texture
  14. Right, the only balancing act if you wanna do it the hack way is simply how many grains thick you make it along with constraint parameters -- That being said I've found cloth in houdini to be getting better and better each version, despite innumerable changes. Or maybe I'm just better at not doing dumb stuff
  • Create New...