Hello,
I'm in the process of developing an HDA for our in-house AOVs, and I'm handling everything through the Sdf API. The issue I'm facing is that it generates all renderProduct productName time samples at once, which creates time dependencies. This causes Houdini to cook the node for each frame, even when it's not necessary.
I could modify my code to generate these value clips only for the current frame instead of the entire render range, but that seems less efficient. Is there a way to inform Houdini that this Python script node, despite creating time dependencies, is not actually time-dependent?
Thank you for your help!
Handling Time Dependencies in Python Script Nodes
412 3 0-
- dkeraudren
- Member
- 8 posts
- Joined: July 2019
- Offline
-
- Heileif
- Member
- 225 posts
- Joined: Jan. 2015
- Offline
-
- tamte
- Member
- 9130 posts
- Joined: July 2007
- Offline
Creating timesamples shouldn't make your node time dependent, in fact I believe this is a preferred way and many nodes author timesamples for the whole range without introducing Houdini timedependency
What can make the node time dependent is if it depends on the actual houdini scene time, or any of the nodes/data it samples from the scene is time dependent
But I'd assume creating timesamples from scratch or based on incoming stage timesamples should not create node time dependency
What can make the node time dependent is if it depends on the actual houdini scene time, or any of the nodes/data it samples from the scene is time dependent
But I'd assume creating timesamples from scratch or based on incoming stage timesamples should not create node time dependency
Tomas Slancik
CG Supervisor
Framestore, NY
CG Supervisor
Framestore, NY
-
- dkeraudren
- Member
- 8 posts
- Joined: July 2019
- Offline
Thank you for your help! I am now certain that the time dependencies were not caused by setting the time sample. After reviewing my code a few days later with a fresh perspective, I realized the issue was due to calling the eval() function on some parameters. It all makes sense now. I suspect Heileif might have encountered the same issue and was fixed when reading range from metadata.
-
- Quick Links