「利用者:Sebbas/GSoC 2017/Reports/Week 6」の版間の差分
< 利用者:Sebbas | GSoC 2017 | Reports
(→What's been done) |
細 (1版 をインポートしました) |
(相違点なし)
|
2018年6月29日 (金) 06:22時点における最新版
Week 6
This week I cleaned up the "Pointcache for liquids" and also started work on float and tracer particle support.
What's been done
- I fixed the issues that I had with caching particles in the "Pointcache" format. The "best case scenario" I described last week is now complete: smoke and liquids run cache smoothly with the Pointache. All in all, I think the move away from caching liquids with .uni files to caching with Pointcache files should pay off. Caching feels much more "Blender-like" now. I also think this will make the transition to the master branch easier. (rB72311c6357e0)
- The FLIP particle option is back (I temporarily removed it when setting up drop particles). So in the UI you can now choose between FLIP and/or Drop particles. The corresponding particle systems will get allocated/removed automatically. (rB93d31ac344c5)
- I am currently working on the remaining two secondary particle types "Float" and "Tracer". On Blenderartists [1] we decided that it's best to setup separate particle systems for each type. I am therefore now making my particle code more modular (it's better to not just duplicate code when setting up multiple particle systems). (rB47056f2b3a4f)
Next up
- For the float particles I would like to have a first version next week (maybe I can squeeze in initial tracer particle support too). So what I'll do is work on the Mantaflow side again and refine my secondary particle sampling functions.
- On the Blender side, I'll see that I have a very generic particle data getter setup. Again, when reading mutiple particle types it's best to not duplicate too much code.
Questions
- No