Describe the bug
The test 0131 shows discrepancies between hydra and the procedural due to the fact that hydra does not skin the normals. The following character used for the crowd contains normals and a subdivisionScheme set to none. Storm doesn't use the normals and is also not interpreting correctly the handedness.
Using subdivision, the character looks correct:
The usdskel imaging code shows that when a mesh is skinned, hydra skips the normals, it is explained in .//usdImaging/usdSkelImaging/skeletonAdapter.cpp UsdSkelImagingSkeletonAdapter::_UpdateSkinnedPrimForTime
However in our procedural we skin the normals, hence the discrepancies observed in the test 0131.
Describe the bug
The test 0131 shows discrepancies between hydra and the procedural due to the fact that hydra does not skin the normals. The following character used for the crowd contains normals and a subdivisionScheme set to none. Storm doesn't use the normals and is also not interpreting correctly the handedness.
Using subdivision, the character looks correct:
The usdskel imaging code shows that when a mesh is skinned, hydra skips the normals, it is explained in .//usdImaging/usdSkelImaging/skeletonAdapter.cpp
UsdSkelImagingSkeletonAdapter::_UpdateSkinnedPrimForTimeHowever in our procedural we skin the normals, hence the discrepancies observed in the test 0131.