Jump to content


Popular Content

Showing most liked content on 01/04/2019 in all areas

  1. 2 points
    AMNH in NYC is looking for a TD to fill a year long spot on our next planetarium production. See details here: https://careers.amnh.org/postings/1718 We're also potentially looking for a few modelers and Maya/Houdini lookdev artists to help develop spacecraft models and planetary elements over the course of production, and a Nuke compositor later in the summer. Staff position is on-site, and freelance gigs are preferred on-site but we could work remotely with the right person. Our shows are regularly considered among the best in the field and emphasize accuracy and making cutting edge scientific data visualization accessible to a general audience. Two examples of previous shows: https://www.amnh.org/exhibitions/space-show/dark-universe https://www.amnh.org/global-business-development/planetarium-content/passport-to-the-universe This is a very small commercial size team working on a medium length production with a reasonable schedule, which could be a nice change of pace from either commercials or film, depending on which way you look at it. The content is typically pretty technical and substance always wins over style here. The TDs often end up doing a wide range of work from camera layout, to FX, to graphics programming and Python scripting. And for you Houdini nerds, Houdini is our core production package for most content creation and rendering. Email Gavin Guerra (gguerra@amnh.org) or me at my AMNH email (jpparker@amnh.org) for more info. -Jon
  2. 1 point
    Yep, it's a bitch! A lot of older tutorials will encourage you to use Copy Stamping, but it's mostly a deprecated workflow as of Houdini 16. If you want more information about all this template point / copying business, take a look here: http://www.toadstorm.com/blog/?p=493
  3. 1 point
    Again, don't use a copy stamp. You can define all your template points in advance, then use groups to separate even / odd points, then copy the appropriate geometry to each group. Alternatively, you can use a For/Each loop to define which geometry gets copied to which point. It's a little more complicated, but it scales up much better than Copy Stamping does, especially when compiled. I'm attaching an example of both methods. copy_sop_switching_toadstorm.hip
  4. 1 point
    here is the example connect the Python SOP directly to the geo whose first primitive is the VDB volume you care about import numpy node = hou.pwd() geo = node.geometry() vdb = geo.iterPrims()[0] vdbRes = vdb.resolution() vdbBounds = vdb.activeVoxelBoundingBox() values = vdb.voxelRange(vdbBounds) values = numpy.array(values) values3d = values.reshape(vdbRes, order='F') print values3d
  5. 1 point
    I don't believe is possible in off the shelf houdini without writing your own tool to do this.
  6. 1 point
    couple of things, I focused only at SDf force as that's the way to go: - you are using Isooffset to generate SDF, which is slow and may introduce artifacts, consider using VDB From Polygons - your volumegradient is not sampling the SDF, but the default cube geo, you have to point it to your geo - your volumegradient is sampling using v, it should sample at P position - you are adding P to result of your gradient, while the gradient direction is already parallel to the tube normal, so you shouldn't mess with it more, just negating and normalizing is fine - VOP Force is old and may be slow, in your case when using RBDs consider using POP Force Instead here is the file with all the changes also has both VOP Force and POP Force approach Custom_Gravity_fix.hip
  7. 1 point
    Divide and Conquer. If you want to do everything in one go...goodluck....I'm not at Jedi level so I'll take the Divide and Conquer approach. Here's my file, you should do it in stages below to get the cache out to proceed to next steps: 01) Sim the Ball and Trampoline by themselves, make sure the active solver (next to brain at bottom) is vellumsolver1. Save out the sim with filecache1 (Save to Disk), then ensure to tick Load from Disk. An extremely simple sim. 02) Now make sure active solver is vellumsolver2. This uses the cached out file in step 1 as colliders. So now the sim is for the ice only...(Ball + Trampoline are simply colliders) not Ball + Trampoline + Ice all active at the same time.....this would be a nightmare to control. you'll notice I use 'lines' only for the ice and not the orig broken up pieces. This is because it's much harder to make cloth that is 'stiff' as ice. While with Hair, using low res lines, it's easier to make it stiff. So with this approach, the 1st solver uses a measly substep of 4...while the 2nd uses just the default ONE !!! vu_vellumtrampolineice.hiplc
  8. 1 point
    what does your $HIP evaluate to? In other words, does your fileName path: "$HIP/../anim/sceneABCexport/"+ns[1:]+'.abc' result in a valid path?
  9. 1 point
    I like this solution, it seems to respond to stiffness fairly well. I added a slight sin based wind along the z-axis. sin($T*55)*2 ap_vu_simple_vellum_bush.hiplc
  10. 1 point
    It could be because the box isn't a volume. Try putting an isoffset before the scatter Let me know how you get on
  11. 1 point
    seems to work for me, this was executed in hython process run from command line aa = hou.node("/obj").createNode('alembicarchive') aa.parm('fileName').set("default.abc") print("Has UI: {}".format( hou.isUIAvailable() )) print("Subnet Content before update: {}".format( aa.children() )) aa.parm('buildHierarchy').pressButton() print("Subnet Content after update: {}".format( aa.children() )) output Has UI: False Subnet Content before update: () Subnet Content after update: (<hou.ObjNode of type alembicxform at /obj/alembicarchive1/gabc>,)