vijinr@sidefx.com
21.0.474
First of all, thank you so much for your work, the support you provide is incredible!! Congratulations.
I’m the one who has been reporting the issues related to the “crash while re-parenting/copying multiple prims.”
I’ve been checking version H 21.0.474, and those errors are now fixed.
IMPORTANT:However, there is still a strange behavior/bug that, if you were able to solve it, would fix what I personally consider the biggest problem of Solaris when trying to work on a full pipeline involving multiple departments inside it.( be able to make full layout and manual dressing in solaris viewport --- using stage manager )
In short:The problem is that if an asset has a previous transformation (including a scale change), when you use a Stage Manager, for example, to transform or duplicate that asset and move it, the transformation and the pivot go crazy. making this node useless if you want to use it by different usdlayers.
In the attached video A, you can see what I mean.I uploaded a videoB too, with the same behavior with 2 stage manager nodes.with a simple example.
Why is this so important?Because it prevents the use of multiple Stage Managers within the same USD. And this limitation makes it impossible to properly use Solaris at each stage of the pipeline in a real production.
Example 1)Several assets have been created and published by the modeling and surfacing departments for a sequence of 3 shots.
The layout department uses a Stage Manager to load those elements and position them in the scene, scaling, transforming, and duplicating them, in order to publish a sublayer called layout_env_blocking, where the environment blocking pass is worked on (still without working from camera).
In another step within the layout department, once the blocking is approved, the goal is to work on that environment blocking from camera, refining positions and duplicating or relocating things so they look better from the different cameras.
To do this, we would load the layout_env_blocking sublayer and create three branches in Solaris: shot01, shot02, and shot03.
What we need and want is to be able to use another Stage Manager at this point—ideally a different one for each shot_camera—to perform step 2 of layout from camera, adjusting the blocking we already had and adding more detail from the camera’s point of view.
For this we would use another Stage Manager that reads the previous elements and allows us to move them again, transform them, duplicate them, or add new ones.
The problem is that with the bug/behavior shown in video A, this is not possible. Since the assets already have previous transformations, the Stage Manager in shot01 does not behave correctly, and we see the broken behavior shown in video A.
Fixing that behavior would allow us to work full pipeline inside Solaris, since the Stage Manager is the main tool to be able to do layout in the viewport precisely (the Layout node is less precise).
Example 2 of using Stage Manager in a real pipeline:
I work at a company where we had to develop an alternative solution to Stage Manager because of these problems.
In this company there is a layout department and a set dressing department. The set dressing department works from the USD published by layout, which contains the blocking pass and the cameras.
If layout uses a Stage Manager to place elements in the scene, transform them, duplicate them, etc…
When we then load that sublayer in order to work on top of it—putting a new Stage Manager to edit the layout, change variants, move them, etc.—the error shown in the video appears.
CONCLUSION:If the Stage Manager could fix the behavior shown in the video, and be able to read the previous transformations of the assets while still allowing us to edit transformations, change variants, duplicate, and add new elements (which it already can, except for the bug), that would give the green light to work layout and set dressing fully in Solaris USD within Houdini.
If I haven’t explained myself well or you have any questions, please don’t hesitate to ask—I’m very interested in seeing this solved or improved.
At first this comes from my personal/home projects, but it would also be very good news for the company where I work, which uses Solaris and USD across multiple departments in the pipeline.
attachments :
videoA - bug
videoB - another bug example
hip file with production tests examples , to show who important is the stage manager will be fixed.Best regards, and thank you so much.