caleb

I finished this project up without using Vellum, which is a mild disappointment, but let me demonstrate how to use basic Trig to cheaply fake the influence of gravity. Here's how it ended up:

There is at least one issue with this, in that the post-process which NURBifies the polygons causes the branches to no longer be attached to the vines. There are a number of ways to approach this. The most appropriate being to not use the cheap NURBification at all. It's a dirty hack. The overarching point on the exercise is that the base structure of the pipeline is modular, and generally stays the same, while the functionality of each module is improved iteratively. I'm done with this one for now, but may return to it. Let me know of improvements! :-) -caleb

First dump from a home-brew Grammar system expressed iteratively as an evolving growth system. A decade of hunting down Blackberries to their roots - to remove them from my yard; A month of looking at and sketching Blackberry growth in the wild. A day with Houdini Apprentice, and a single point of geometry to seed the system. Then just turned it on. There's fewer than 100 lines of code, and I'd guess about a dozen substitution rules in the grammar system. Emergence is an effective approach for modelling organic processes. Blackberry_h17.13.hipnc

Caleb J. Howard 5 mins Yo smartipantses! I'm just noodling with vellum for the first time. Has anyone had success with a vellum solver integrating with a SOp solver that is changing the topology? In particular, I want the equivalent of a wire solver, so I'm using the hair constraint template. The wires are open polygons - growing by adding new points to the end every so often. I'd like the added edges to be added with the same bend and stretch constraints that the pre-existing edges have Can anyone point me at a solution or example hip which runs a vellum solver with a SOp solver, where the latter is adding geometry throughout the simulation? (The image shows the first point added just falling - with ineffective constraints. I've tried several variants with essentially similar results.) Thanks!

OK! Thanks! Trying again! CalebHoward__trace_foreach.otl.zip

Hmm. I attached a zip file, but am not seeing ti. Am I doing something wrong?

minor bug fix. New version pending on orbolt. Here it is here:

Have you ever worked with the ForEach Sop in Houdini? Have you ever nested them many layers deep? This past weekend I rewrote and documented a Sop I developed a year or so ago – the trace_foreach Sop. It will allow you to walk through any complex hierarchy of nested ForEach Sops and see the state of the loops at any point. It’s a trace for the ForEach Sop. It is here. Limited in its testing. Please let me know of any failures. https://www.orbolt.com/asset/CalebHoward::trace_foreach

Something like that. I put it on orbolt. I'll post a link now. Thanks!

Hi! Might help - I've just uploaded a trace_foreach SOp asset to orbolt. I'll post the link when it clears. -caleb

subject pretty much says it all.

Hey there! Can you point me at the examples you found searching on Plastic, please? I'm searching on Plastic, but finding zero examples in the Houdini 12 docs. Any help will be appreciated. Thanks!

Hey petz - thanks! I was looking for exactly this sort of example, and here it is.

Hiya! Before I dive into developing this myself, has anyone done anything like this: I have need for a more adaptive surfacing method to generate a mesh from an intermediary unorganized point cloud (with Normals), or from a volume primitive directly. I have already developed methods to generate a nonuniform point cloud from a (uniform) volume primitive - with the points where I need the detail. I would love a Delaunay triangulation from that. (The standard marching-cubes method doesn't have the nonuniform detail I need) Has anyone done this? Any similar experience, or pointers to software that might provide this solution? Thanks! -caleb [Cross-posted to sidefx.com - sorry for the bandwidth]

For the interest of anyone with curiosity, The SESI forum Thread I started got a more useful answer: http://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&t=17926 So far, there is no straightforward way to maintain my own Python class data to be persistent between cooks, but there is an acceptable workaround of converting Python data to strings by means of the repr() function, and stuffing those strings into hidden string parameters. I haven't tried this for large datasets yet, but this is adequate to my needs. -caleb