>>2765193I'm working on it. I'm adding morphs, the physics aren't good yet and I'm trying to solve a bunch of rigging issues that arose from adding the arm IK canceller things. Well I'm mostly done with the rigging (fixed all immediate problems), but I still need to add like a dozen morphs, tweak physics and do a lot of testing with motions, those tasks being really boring. Now that I know what my problems are and how/what to test on, I don't wanna spam with every tiny update and post the thing when it's actually mostly done. My main aim right now is to get the rig working right in motions so I can start polishing some other models of mine and then simply use the script/bone data on them.
I'm also trying to improve the unsubdivider I wrote. With it, I reduced the polycount further from the 160k in "1.02" to 130k now without compromising topology much. The anon mentioned that my approach had problems was right, it worked only for 1 subdivision and failed in edge cases. I rewrote it and now I'm manually stepping along edge rings from the edge marks using interpolated vertex weights as reference, is but it's still not great. Pic related. Better than automatic decimation though. Of course a normal person would just dissolve unneeded edges by hand, but I want to be able to work only on the unsubdivided meshes without redoing any post-tweaks every time I wanna re-export. I am lazy.
