Ambiguous or incomplete explanation of NURBS curves?

Community USD
1

The state of Universal Scene Description: UsdGeomNurbsCurves Class Reference seems a bit unclear and ambiguous when it comes to periodicity for both the points data and the knots vector.

I suppose the first thing that needs cleared up is the number of points expected in periodic NURBS curves. This isn’t really stated and the knot vector calculation provided would be ambiguous depending on the answer to this question.

  • The first calculation given for knots, (numSegments + 2 * degree + 1), never describes what numSegments actually is. NURBS are usually described with the (#CPs + order) form, which is what USD seems to want, or with the smaller (#CPs + degree - 1) form, but never in terms of segments. Is the segments term just a left over from the basis curves documentation perhaps?
  • For periodic NURBS, how many points need to be defined? There’s very few examples out there [1][2] but it seems like we should duplicate the first degree # of CPs at the very end? But the docs imply that only manipulation of the knots vector is required to express a periodic curve but detecting the type of curve like this would be ambiguous.
  • Are both periodic and non-periodic NURBS allowed in the same “batch”?

[1] Full github search: knots path:*.usda NOT path:generatedSchema.usda NOT path:schema.usda
[2] maya-usd/test/testSamples/basisCurves/nurbsCircleExport.usda at dev · Autodesk/maya-usd · GitHub

2

Bump? Still hoping to get an ultimate clarification here.

3

Sorry about the lack of reply. I’ll mention this in the geometry working group and see if anyone has an answer.