All Activity

Wrap new nodes into one subnetwork and use the logic as before. Alembic, USD... Save script as a shelf tool. Improve script to be able to select "subnet_path", fully create "attribdelete" instead of copy-paste. Google Python examples, you will find close enough to this tasks.

Does anyone know how to change node handle size of camera manipulator in Houdini? The standard * & / does not affect camera manipulation handle size. preproduce issue: in SOP or in LOP When select camera and press enter, multi handle appears and its uncorrelated to translate or rotate standard size that is controlled with * and /. Houdini 20.5 Thank you for your time!

Ok. Right. I ran the script again and found it was that indeed. My bad. Also...while importing the fbx, I came across a better import option that places materials inside the existing matnet instead of creating a new one inside OBJ network. It also creates all parts directly at the OBJ level, rather than nesting everything inside a subnet (which is what my earlier method was doing). Would it be possible for you to tweak your Python script to work with this setup instead? Here's the updated project file. https://www.dropbox.com/scl/fi/e9np7kw2n0n90s5myjrmn/wristwatch_updatedone-Copy.hiplc?rlkey=slt4wfz55snmtf2ljitfi4wzc&st=dtmr0k66&dl=0 Also, I had a broader question.. Is there a no-code way to achieve something like this in Houdini? Right now, this approach requires knowing Python, remembering the logic, and executing it each time. Is saving these scripts in a file the standard workflow, or is there a more efficient way (like shelf tools, digital assets, or something else) that you’d recommend?

I am trying to create a "USD MaterialX Builder" subnet in my Material Library LOP and failing. The only working solution is: builder = matlib.createNode("subnet", "material name", run_init_scripts=False) But it creates a regular Sunbnetwork, instead of a Material X Sunbnetwork. How to get proper subnet (shown is a screenshot, I created it manually)?

Thank you Konstantin! big fan of your videos!

3 games 3 min. ridiculous games3min.hipnc

attribdelete is set as render node already not sure I understand what this issue is, jump to operator shows exact material applied to the node

Another thing. Is there a direct way to select materials of each piece rather than checking what material is applied to it in the render tab by selecting the obj node and then going to the matnet and then searching for the name amongst 50 materials used in the project, and then clicking the material node? If I click on this button on the right side of material it opens a window containing list of all materials inside the model. Doesn't open the directly the material network of the material applied to the geometry piece. Jump to operator button right next to it only shows this panel.

Thank you!! Wow! This worked quite well! Is it also possible to set the display flag to attribdelete inside each of the parts in bulk too?

1. Create an attribdelete with required parameters 2. Windows > Python Source Editor selected_nodes = hou.selectedNodes() source_node = selected_nodes[0] subnet_path = "/obj/Hublot_King_Power_F1___Watch_fbx" subnet = hou.node(subnet_path) file_nodes = [n for n in subnet.allSubChildren() if n.type().name() == "file"] for file_node in file_nodes: parent = file_node.parent() copied_node = source_node.copyTo(parent) outputs = file_node.outputs() copied_node.setInput(0, file_node) 3. Select attribdelete to copy to 4. Appy script It will copy-paste selected node into each node in Hublot_King_Power_F1___Watch_fbx subnetwork and attach it to file* node

I'm trying to import an fbx model with embedded textures but the textures don't show up in the render. They get hidden in render because of this shop material attribute that I have to delete for each and every piece of the watch. Is there an automated way to delete this attribute instead of entering the node for every piece inside the wristwatch? I converted the embedded materials by using ODtools. Here's the project file. https://www.dropbox.com/scl/fi/e8803p5cmz2pw3vsct820/wristwatch-material-challenge.zip?rlkey=kys5oyueemjzkv9k7d86vrb9j&st=4wpwuhkt&dl=0

Nowadays you would rather use the group expand node, though: https://www.sidefx.com/docs/houdini/nodes/sop/groupexpand.html group_expand.hip

It depends which solver you use. If you use the sop level solver, the velocity should be inside the collision vdb as much as I know. If you use the dop workflow, there is no need to rasterisze velocity since the collision velocity is taken from the incoming points.

we are in 2026 and this help a lot !

Do I need to rasterise velocity into collisionvel field at the collider level before simulating, or is it ok to not rasterise it, and let the solver take the velocity from the points? the collision source sop doesn't offer this option while the vdb from polygon does. What is the difference if any between rasterising it before and not? is it reducing sim time?

I accidentally bumped into this wonderful info on the geometry spreadsheet (simulation time & memory usage) and want to extract it, any idea how?

I've searched in the geometry spreadsheet for substeps couldn't find it, is it available from where or do I need to calculate it some how?

Thank you so much!

Oh yes I forgot to mention this because I created my own circle shape and called it rect to override the default rect node shapes.

Hi, Of course, when PDG submits background job or farm job - it submits in a command line env way so no UI is available. New flipbook node or openGL node give you the option to submit flipbooks to the farm on GPU enabled machines, but this option is extremely limited in a huge way - grabbing the viewport exactly how it is in it's current state. What I mean is - being able to display hard templated geo, templated geo, ghosted geo, viewport background easily, black vs blue background, all those functions are hard coded or just gone. Particularly I'm interested in being able to see hard templated geometry since this is very useful when wanting to see multiple geos without merging and messing up attributes in the viewport - it's also much faster to preview this way as well. So now comes the issue. Since PDG submits farm jobs without the ui available - this means we are not able to flipbook using flipbookSettings or hscript('viewwrite') commands in the background. This forces us to be restricted to flipbook node or openGL rop nodes. I'm asking if there is a way around this limitation? To spawn a job on the farm but WITH the ui enabled so that things like the sceneViewer.flipbookSettings and sceneViewer.flipbook() commands could actually work without crashing on the farm? I was looking into hython, studying up on how TOPS works, but I don't know how to do this without more information. Can anyone give me more background?

Omg... you need to edit the name INSIDE the json.

Yes, it is off. Desperately installing latest daily. Maybe it's something off in the install or something. I'm using Education version, do you think that has anything to do with it? Installed through the launcher.

Did you also turn this option off?

Yes, I copied circle, bone etc and renamed them to rect.json in the user folder. But they all show up as the standard rectangle like in the screenshot. So again just to remain sane: I went into the Houdini install folder, copied different jsons (like bone, circle) to Documents\houdini21.0\config\NodeShapes and renamed them to rect.json just to test. But they all render like the standard rectangle shape in Houdini (round corners). Am I missing something here?

thanks a lot ! that's exactly was I was looking for.