In the attached scene I have a 10 frame vdb sequence that I want to loop. Dive into geo1 and render the geo rop, then go up and into the lop network.
The first rop reads the vdb and wraps it in usd, run it and move to the next step.
The reference lop loads the usd, it gets looped until frame 100, and put behind a variant, so I can turn it on and off.
If you swap the reference for the sublayer, the clip still works, but breaks when fed to the variant.
Is there any reason for this? I use sublayer when I'm sure the heirarchy isn't going to change, but I might start using references instead if sublayers cause odd behavior.
looping a sublayer with clips breaks with variants
3168 3 2- mestela
- Member
- 1804 posts
- Joined: May 2006
- Offline
- mtucker
- Staff
- 4525 posts
- Joined: July 2005
- Offline
I'm not 100% certain, because I don't have a full grasp of exactly how Value Clips work, and where they fit into the “LIVRPS” hierarchy, or if (as it seems) they are somehow outside that hierarchy… But the difference between Reference and Sublayer in your hip file is that when you feed a node into an Add Variant LOP, we do layer flattening of the input so that we can do a direct copy of the SdfSpecs from the second input into the variant portion of the scene graph. In the reference case, flattening the input leaves the reference intact as a reference arc, so the value clip opinions “win”. In the sublayer case, the data from the sublayer is flattened and that data gets copied inside the variant. Somehow the presence of the opinions in the save layer where the value clip is defined prevents the value clip opinions from winning.
You can see the same thing happen with a Graft LOP instead of an Add Variant, as a Graft LOP does the same thing - flattening the secondary input data and copying it directly into the active layer. The value clip wins against data pulled in by reference, but loses if the opinions are set in the active layer.
You can see the same thing happen with a Graft LOP instead of an Add Variant, as a Graft LOP does the same thing - flattening the secondary input data and copying it directly into the active layer. The value clip wins against data pulled in by reference, but loses if the opinions are set in the active layer.
- mestela
- Member
- 1804 posts
- Joined: May 2006
- Offline
- nicolaas.tanghe
- Member
- 44 posts
- Joined: March 2023
- Offline
-
- Quick Links