search

Bismuth-Consultancy-BV / MLOPs

Machine Learning Toolset for Houdini
BSD 3-Clause "New" or "Revised" License
356 stars 53 forks source link

Camera to Points HDA issues #112

Open luisfos opened 11 months ago

luisfos commented 11 months ago

  1. currently won't pick up "objects_to_render" if they are not directly in /obj e.g nested or within object networks. camera_to_points/set_objects_to_import python SOP has hard coded "/obj" preappended to objects for render. in python sop replace "/obj" with "scene_path" parameter to allow it to correctly find any nested objects.

  2. Modes are flipped in X inconsistently Correct: Depth, Normal, Albedo Flipped: Segmentation, Shading

mnlrbr commented 8 months ago

What Luis said. There is one Flipp _along_Y point wrangle for the segmentation and in the python ogl rendering there are some transpose operations...

oliverban commented 3 months ago

Was this ever fixed? I am having major problems with FBX scenes with imported characters that come in as as huge hierarchies with bones and what not. The scene works fine but the Camera To Points doesn't seem to be able to render them when they are in a Subnet. Right Clicking and Extracting Content directly into OBJ just crashed the Camera To Points node! :( :(