Centre-of-rotation skinning (optimised CoR, Le & Hodgins 2016) as an option: no candy-wrapper wrists or collapsed joints, LBS-compatible, default off #1581
Labels
No labels
amd
avatars
bug
builder
cannot-reproduce
crash
epic
feature
from-viewer
gl-removal
known-limitation
monitoring
parity
perf
platform
priority
critical
priority
high
priority
low
priority
medium
provenance
rendering
rt
status
in-progress
status
testing
status
up-next
triage
ui
upstream
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
sky_line/Slipstream#1581
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Add optimised centre-of-rotation (CoR) skinning (Le & Hodgins, SIGGRAPH 2016) as an option for rigged mesh, alongside today's linear blend skinning (LBS). Default off.
Why
LBS averages bone matrices, so twisting and sharply bent joints collapse:
These show on Bento hands and fingers, tails, necks and quadruped legs (Open Pony). Plain dual quaternion skinning (DQS) fixes the collapse but bulges at bends, and it cannot carry the bone scale SL depends on (shape sliders, fitted-mesh collision volumes, animated Bento scale). CoR keeps the volume without the bulge, and it's built for content weighted for LBS, which is all SL content.
How
The cost is a handful of extra ALU per vertex in the existing GPU skinning path. The RT BLAS refit consumes the skinned positions unchanged.
Done when
Risks