Wrong bounding box for copied packed spheres

   244   2   0
User Avatar
Member
1 posts
Joined: June 2025
Offline
Hello world,

I’m very new to Houdini and losing my mind over this and really need some help 😅.

I’m 'working' in Houdini 21 and I’m running into a strange issue with Copy to Points.

When I copy spheres onto points that have additional rotation, the bounding box becomes completely wrong (much larger than it should be).

However, if I switch the copied geometry to a cube, the bounding box is perfectly correct.

Same setup, same rotations , only the primitive changes.

Is this expected behavior with spheres ? Am I missing something obvious about how Houdini computes bounding boxes for rotated spheres?

Attachments:
2026-01-06_22h38_43.png (813.7 KB)
2026-01-06_22h39_46.png (796.7 KB)
2026-01-06_22h52_08.gif (4.0 MB)
wrong_bbox_for_rotated_copied_packed_sphere - Copy.hipnc (90.2 KB)

User Avatar
Member
9406 posts
Joined: July 2007
Offline
when geometry is packed you will get bounding box of the transformed bounding box of the original geo

so in your case try to replace your source sphere with it's bbox, then you will see that the final bbox on transformed geo is the same as it was for the packed sphere itself

otherwise the bounds sop would have to unpack all packed prims if it wanted to compute perfect bbox based on every packed shape contained
Edited by tamte - Jan. 7, 2026 19:26:20
Tomas Slancik
CG Supervisor
Framestore, NY
User Avatar
Member
73 posts
Joined: April 2013
Online
Hi David,

Deleted my previous post, the answer by Tomas clears up why it does what it does.

Forgot that I made a simple hip for you that has working solutions just incase you needed by packing
outside of the copy node where the bounding box looks like it works with a packed object as expected.


Image Not Found

Attachments:
Bounding_Box_TEST_David.hipnc (370.4 KB)

  • Quick Links