>>2832879But... why?
>>2833038It used to be somewhat interesting when filesize limits were stricter.
>>2833039One of the few good things to come from MARUSU is this tutorial:
https://pixeldrain@com/u/ykq91DRWHowever, I have a few points to add based on my own experience of using the juicy fruits of it over the years:
1. The jiggle should be parented to leg (hip) bones instead.
2. Weights should be fairly smooth, with more steep progression above. My model uses a 1-3-5-10-14-18-22-25-25-25-20-15-8-2 progression from bottom to top, with some 100% 足 bone space above the knee and the transfer from lower body weights to leg above and in the first two rings/rows of the jiggle. I would probably have expanded at least several rows further up, but she had SDEF hip joints at the time I incorporated this feature. Here's another reason to go for another Blender remodeling round, I guess?
3. 0.5 movement/rotation decay (jiggles longer) and Phys+Bone mode (to stick with the thigh) work better for jiggle rigids. The last bit is particularly important if you're using skirts or stuff that you want to collide with the thicker-than-leg jiggle rigid to help against clipping.
4. See pic related for better joint parameters. Slide should definitely be highly restricted in most use cases of MMD physics, and it's not a good idea to use it with low weights at all due to how much the rigid detaches from the mesh (=wonky collision).
