Welcome, Guest
Username: Password: Remember me
  • Page:
  • 1

TOPIC: Known Issues (updated 25th November 2019)

Known Issues (updated 25th November 2019) 1 year 2 days ago #77

  • Sebastian
  • Sebastian's Avatar Topic Author
  • Offline
  • Administrator
  • Administrator
  • Posts: 44
  • Thank you received: 12
This is an ongoing list of Known Issues - as of the version 9.3.2 of our Carbon For Houdini plugin:

  • multiparm warnings in DOP with Carbon Binding or Carbon Welding: Loading a scene with DOP Carbon Binding / DOP Carbon Welding, or creating a DOP Carbon Binding / DOP Carbon Welding causes warnings similar to the ones below to be printed to the Houdini Console.
    <internal> Warning(33): incorrect number of #s in 'objectA' for multiparm depth of 1
    It has no known effect on the simulation.
  • Invalid parameter name sharedata warnings in DOP with Carbon Rigid Body: Loading a scene with DOP Carbon Rigid Body, or adding a DOP Carbon Rigid Body causes warnings similar to the ones below to be printed to the Houdini Console.
    emptyobject OP_Parameters::evalIntT(): Invalid parameter name sharedata
    It has no known effect on the simulation.
  • Geometry Guides don’t update: Reset the simulation to see the Geometry Guides correctly drawn.
  • Relationship nodes such as Carbon Binding, Carbon Stitching, … etc. don’t respond well to be included inside a digital asset as the list parameters don’t call their callback scripts when they are modified programatically: Use either opparm -C or hou.Parm.pressButton() on the list parameters such as cloth#, group# …etc.
  • Renaming obj nodes such as cloth/colliders breaks relationship setup as they keep the old names: There are currently no known workarounds to this issue.
  • Houdini Help browser prevents downloads: The default Houdini Help browser does not support clickable links for file downloads. Fix by setting environment variable HOUDINI_EXTERNAL_HELP_BROWSER = 1 to use your system browser rather than Houdini’s internal one.
  • [RESOLVED as of version 9.2.0] Points of a goal posed object that are kinematic at the object initialization stage will remain kinematic throughout the whole simulation, even when animating Goal Skin parameters, such as Upper Threshold, to turn them non-kinematic: To improve solver speed, we apply internal optimizations where a lot of soft constraints for the kinematic points of a goal skinned object, are not created. Even if the Upper Threshold is animated so that certain/all points are not kinematic for parts of the simulation, those constraints will not be created during the simulation. Therefore, points that are not kinematic for the whole sim, should not be kinematic at initialization. Best to start with all points non-kinematic and set the ones you want kinematic at frame 1.
  • [RESOLVED as of version 9.2.0] Feeding non tet object into Carbon tet/simulation node will crash Houdini: When you are setting up a tet object, make sure you set it up with tetrahedralization before you connect it to a Carbon simulation node, otherwise it will crash Houdini
  • Rendering issue: When you hit esc to terminate a SOP simulation you may hit a point where the last frame render has been deleted and next frame render has not yet happened - so no scene is rendered. Just step back or forward one frame and all will be correct.
  • [Versions < 9.2.0] Carbon Binding and Carbon Welding wrongly using reference instead of start pose positions: In some cases, Carbon Binding and Carbon Welding wrongly use the cloth reference pose positions for constraint reference calculations. This is fixed in Versions >= 9.2.0.
  • Perfect Regular grids of Quads or Triangles take longer to initialize:
    If you create a large cloth (>100,000) with a PERFECT grid of quads or triangles, then it can take a noticeable time to initialize, every time you load the scene or reset the sim 150,000 quads can take 30 seconds. It seems Carbon's collision pre-processing is very recursive on perfect grids. Move one point off the grid and performance increases 30x !
    It's rare to have such a perfect grid unless you are literally testing with a large sheet, but we will look to improve this in future releases.
Last edit: by Sebastian.
The topic has been locked.
  • Page:
  • 1
Time to create page: 0.090 seconds