Forget it. I accomplished the desired result in Houdini by using a transform node with

origin("","..","TX") origin("","..","TY") origin("","..","TZ")
origin("","..","RX") origin("","..","RY") origin("","..","RZ")
origin("","..","SX") origin("","..","SY") origin("","..","SZ")

in the channels and then ticking "Invert Transformation" down below.

Is there a solution for Unreal?

I want to turn off "Push Transform to Houdini" on my HDA node in Unreal by default. I can't find it in the Type Properties window. How can I change the default to be unchecked?

Thanks for the tip! I put the name of the node I want to click a button on into the Type Properties -> Node -> Editable Nodes text field, but it only works in Unreal if the HDA in Unreal has the Cook Triggers -> On Parameter/Input Change button checked on. Can I set it up so that I can use my HDA button without having that box checked? It's super slow and annoying having it recook every time I change a setting.

I tried adding a python callback to render the VAT node. The line of python works in Houdini, but it seems like it's not getting run from Unreal.

Is there an easy way to run a python script in Unreal after an HDA is finished cooking? I see options from the Houdini side, but I want to automate the Unreal setup after the VAT ROP is run in an HDA - import the pos and rot textures, import the fbx, set up the material instance, etc. I know that VAT 2.1 has some scripts, and I'm using the materials that have been provided, but I want to run the setup from a python script, since it's basically the same steps every time.

I want to automate RBD destruction simulation in an HDA in Unreal. Any tips or tricks to make that work? The RBDSolver node seems to not like being in an HDA.

The goal is to feed geo from Unreal to the HDA through the Object Merge node and pass it through an RBD Material Fracture, then RBD Solver nodes, and export the simulation through the Labs VAT output node.

Edit: It works if you add a File Cache node after the RBD Solver node.

I'm trying to create an HDA that will do some RBD destruction of Unreal assets in Houdini using the Labs Vertex Animation Textures node and the RBD Material Fracture node. When I render the VAT node on a simple torus in Houdini it works fine - I get the textures, materials and fbx files. When I install the HDA in Unreal and try to Initialize for Engine and then Render the VAT, it does not work. I get an fbx and I get the _pos.exr texture. I do not get a materials folder or the _rot.exr texture.

I tried exporting the geo from Unreal as .fbx. When I import that .fbx into Houdini the verts are scrambled, as if the vert count is wrong. It's not 100% scrambled, but it's wrong. The cook log suggests that some attributes that Houdini wants are missing, e.g. name. The FBX includes various LODs.

I am using UE 4.26.1 and Houdini 18.5.408. I'm using version 2 of the Engine plugin.

Any ideas?

I have a very simple network that accepts geo from the HDA in Unreal, via the Object Merge. I put the Labs VAT node inside a ROPNet so that it can be included in the HDA definition. I can not get the _rot.exr or the material .json. I do not know why it is not working.

In the HDA I have exposed the Initialize for Engine button and the Render button that is on the Labs VAT node, so I can click those buttons from within Unreal.

Edit: One thing that seemed to help was to go into the type properties, and in the Save tab, turn on the option for Unlock New Nodes on Creation.

I think this problem might be due to running out of disk space on my root partition.

I am getting this error as well, rendering a flip simulation. The uniformvolume shader is right out of the box. I didn't touch it. My scene will render with the errors using the film cannister button in the lower left corner of the Scene View, but rendering by pressing “render” on the mantra node fails. The weird thing is that it renders two frames, then fails on frame 3. After I did a “delete channel” on the displacement bound the problem went away. I'm on Centos 7, Houdini 17.0.352.

How is selecting grains in the scene view supposed to work? I'm trying at home in Centos 7, with an Nvidia GTX 970, driver version 384.111.

The best I can get is this:

1) create a cube of grains and simulate a few frames.

2) box select some grains in wireframe view. All of the grains disappear.

3) toggle shaded mode on and off. All of the grains become visible again, and the selected wireframe grains are now yellow.

4) Try to select any other grains. Impossible.

5) advance a frame in the simulation, and you can go back to step 2.

This is the experience on Centos 7. In Windows 10 it's much more erratic for me.

Can someone please test this? Am I going crazy?

New graphics card driver was installed. Still not working correctly.

I'm having inconsistent behavior trying to select grains and particles, but not vertices on geo. I can't consistently box select. It works about 5% of the time.

I've got an Nvidia GeForce GTX 1080.

Driver version is 23.21.13.915, which seems weird to me. I thought they wer ein the high 300's now. That looks like an i.p. address
- LOL.

The OS is Win 10. I tried rebooting.

I'm also having issues where I dive into or out of a node and the points do not display. I have to jump around a bit more and then they'll appear.

Sometimes my grain pile is black, in the viewport, like a holdout, and sometimes they're there, but not really there. This is starting to feel like a graphics card issue.

Needless to say this is decreasing my productivity. Any advice?

Thanks for the info. I didn't realize the Indie license was so restricted.

I was using Apprentice on this machine for a while. I just bought an Indie license.

I'm trying to output a frame range of texture maps from an oceanEvaluate node. I'm getting a warning saying “Image resolution is larger than 1920x1080(Apprentice HD License limit for sequences of images)”. I am not able to output my 3000x3000 displacement map sequence.

But when I look at the license server I'm using the Indie license. How can I get around this problem?

Please see attached image.

I just wanted to say thanks for the file angel_FLIP_rest.hiplc! That's a great succinct example of a rest position setup.

I have Houdini 15.5.607, educational version, on a Centos 7 machine. Everything was working fine, but recently I've run into a problem. When I click in the network view, hit tab, and start typing, the autocomplete feature is not working any more. It's actually kind of a pain to use Houdini without that little convenience. Does anyone have any idea if there is maybe a KDE desktop setting I messed up, or is there a Houdini fix? Any ideas appreciated.

I re-installed Houdini. That did not help.

Note to self: I installed nvidia driver 367.44 and things seem to be working much better. These resources helped a lot:

https://www.youtube.com/watch?v=he3s_JRnXWE [youtube.com]

http://www.dedoimedo.com/computers/centos-7-nvidia.html [dedoimedo.com]

https://www.linkedin.com/pulse/rhel7centos-nvidia-drviers-updated-christopher-meacham [linkedin.com]

Did you find a solution? I get a similar error before Houdini starts, right after the splash screen.

I just installed CentOS 7. I downloaded and installed Houdini. When I try to run Houdini I get the splash screen and then it crashes. I have an nvidia GTX 970 graphics card.

I'm using the default nouveau driver. I have not installed any graphics drivers. I hope that's not my problem/solution. I googled how to install the nvidia drivers on CentOS 7 and it looks like a scary world of pain. I've tried running Houdini from the command line, and from the start menu thingy, and it crashes both ways.

Any ideas? Anyone?

Here is the crash report from the /tmp folder:


$ cat crash.untitled.jpatterson_4929_log.txt
Crash report from jpatterson; Houdini FX Version 15.5.607
Uptime 0 seconds
Sat Oct 1 11:52:14 2016
Caught signal 11

Traceback from 4929 ThreadId=0x7f413ff0b300
AP_Interface::coreDumpChaser(UTsignalHandlerArg) <libHoudiniUI.so>
AP_Interface::si_CrashHandler::chaser(UTsignalHandlerArg) <libHoudiniUI.so>
signalCallback(UTsignalHandlerArg) <libHoudiniUT.so>
UT_Signal::UT_ComboSignalHandler:perator()(int, siginfo*, void*) const <libHoudiniUT.so>
UT_Signal::processSignal(int, siginfo*, void*) <libHoudiniUT.so>
_L_unlock_13 <libpthread.so.0>
QReadWriteLock::lockForWrite() <libQtCore.so.4>
QGLWidget::deleteTexture(unsigned int) <libQtOpenGL.so.4>
QGLContext::~QGLContext() <libQtOpenGL.so.4>
QGLContext::~QGLContext() <libQtOpenGL.so.4>
QGLWidget::~QGLWidget() <libQtOpenGL.so.4>
re_QGLWidget::~re_QGLWidget() <libHoudiniUI.so>
QObjectPrivate::deleteChildren() <libQtCore.so.4>
QWidget::~QWidget() <libQtGui.so.4>
QWidget::~QWidget() <libQtGui.so.4>
QObjectPrivate::deleteChildren() <libQtCore.so.4>
QWidget::~QWidget() <libQtGui.so.4>
re_QWindowWidget::~re_QWindowWidget() <libHoudiniUI.so>
RE_QtWindow::~RE_QtWindow() <libHoudiniUI.so>
RE_QtWindow::~RE_QtWindow() <libHoudiniUI.so>
UI_Window::~UI_Window() <libHoudiniUI.so>
QT_Window::~QT_Window() <libHoudiniUI.so>
FUSE_AnonymousStats::~FUSE_AnonymousStats() <libHoudiniAPPS3.so>
FUSE_AnonymousStats::~FUSE_AnonymousStats() <libHoudiniAPPS3.so>
FUSE_App::~FUSE_App() <libHoudiniAPPS3.so>
FUSE_App::~FUSE_App() <libHoudiniAPPS3.so>
FUSE_HoudiniBinariesApp::~FUSE_HoudiniBinariesApp() <libHoudiniAPPS3.so>
__run_exit_handlers <libc.so.6>
__GI_exit <libc.so.6>
UT_Exit::doExit(int, bool) <libHoudiniUT.so>
RE_OGLRender::initGLState() <libHoudiniUI.so>
RE_OGLRender::makeCurrent(bool) <libHoudiniUI.so>
RE_OGLRender::lockContextForRender() <libHoudiniUI.so>
RE_QtWindow:penWindow(int, int) <libHoudiniUI.so>
QT_Window::doOpen() <libHoudiniUI.so>
UI_Window::handleEvent(UI_Event*) <libHoudiniUI.so>
UI_Queue::processNextEvent() <libHoudiniUI.so>
UI_Queue::drain() <libHoudiniUI.so>
UI_Queue::drainEvents(bool, bool) <libHoudiniUI.so>
OPUI_MainApp::loadFilesAfterUiOpen(UI_Event*) <libHoudiniAPPS2.so>
UI_Queue::processNextEvent() <libHoudiniUI.so>
UI_Queue::drain() <libHoudiniUI.so>
UI_Queue::eventLoop() <libHoudiniUI.so>
main <libHoudiniUI.so>
__libc_start_main <libc.so.6>
_start <houdini-bin>

I just stumbled on this thread, and I am looking at the downloaded file, and trying to understand why it works. Can someone correct my thinking here:

1) box node - we create a box

2) facet node - we break the points up from 8 to 24 unique points per box face.

3) att node - we create an attribute that holds the face number on each prim.

4) attribpromote2 node - we transfer those prim number attributes to the 24 points.

5) scatter1 - we scatter points on the prims(?). Why/how does our new attribute get assigned to the points we scatter? Why do we have to promote the attribute to points, if the scatter is operating on the primitives? Maybe each primitive is passing on it's points' attributes to the scatter points it is helping to create?