3.8 KiB
Minutes from 56th OpenVDB TSC meeting, July 14th, 2020, (GMT)
Attendees: Nick A., Jeff L., Ken M., Dan B.
Additional Attendees: Johannes Meng (Intel), JT Nelson (Blender), Peter Cheng (DW), Andre Pradhana (DW)
Regrets: Peter C.
Agenda:
-
Confirm quorum
-
Secretary
-
Forum Issues
-
Siggraph
-
Fast Sweeping
-
Morphology Improvements
-
Next meeting
-
Quorum was confirmed.
-
Secretary was Dan Bailey
-
Forum Issues
User running into issues with Centos8 compilation, should have been fixed in later release of OpenVDB. Nick to suggest using CMake >= 3.12 to resolve subsequent Python3 issue.
There is an unrelated issue that Ken and Nick have run into attempting to use Mac OSX Catalina and a later Clang compiler(?) which manifests as a deprecation warning about a random number generator. Currently, our CI only tests one configuration on Mac which doesn't expose this error. More investigation needed.
- Siggraph
Have to decide between a 25 minute and 55 minute session. Our preference is for the free user event instead of the official Siggraph birds of a feather event which requires a paid registration.
Discussion about what the target audience is for such an event. Decision to opt for 55 minute session split into two - the first half an introductory overview of the library for new users and how to start contributing, the second half a more advanced session. Ken feels the most useful for us is the discussion with the community so would rather dedicate the advanced session entirely to this. Jeff mentioned that we should at least present a quick overview of new and upcoming features first. Dan likely not able to present due to being on an unreliable internet connection for this session.
Question about how well the ASWF format will lend itself to a group discussion rather than being geared towards presentations from the TSC. Ken to reach out to ASWF to ask about what we can do here to improve our ability to hear from the community. Dan mentioned that David Morin frequently uses a poll to allow the audience to pose questions and upvote on other questions asked by the audience, this could be a good way to direct the discussion towards the areas of most interest.
Need to put together title slides.
- Fast Sweeping
What to do when the algorithm fails. Jeff highlighted a number of cases where not having a zero crossing for a level set should be acceptable. One such example is disconnected level set regions arising from a fluid simulation. The algorithm shouldn't generate an exception in this case because it's not an exceptional case. Ken to revisit this.
Andre still awaiting DWA approval for the Extrapolate SOP, hoping to push up a PR soon.
Dan and Ken to discuss voxel slicing optimization offline.
- Morphology Improvements
Currently, a topology union gives different results dependent on whether the serial or parallel algorithm is used. Nick submitted PR751 to add a preserveActiveTiles flag to Tree::topologyUnion() to avoid always densifying active tiles. No objections to doing this. It is acceptable to treat this as a bug and to change the existing behavior, especially as the documentation does not mention tile behavior.
Nick also has improvements to dilation and erosion in PR754. Aim is to deprecate the existing dilateVoxels and erodeVoxel methods with new functions that provide a more uniform interface and support matching functionality. Adds support for edge/vertex erosion based on inverse dilation. Significant performance improvements to large dilations. Note that the result of the threaded dilation differs due to the behaviour of topologyUnion (as discussed above). This PR also fixes a bug where tools::erodeVoxels() always calls pruneLevelSet() regardless of whether the input is a level set.
- Next meeting
July 21st, 2020. 1pm-2pm EDT (GMT-4).