Hi Anto,

> Well I meant something like in image. Selected one is exported as IGES from MOI,
> uniformly re - parametrized in Softimage, imported back into MOI. Points are on
> the same position, but shape is different, after Softimage re - parametrization. <...>

I see, so I guess what Softimage is doing in that case is throwing out what is called the "knot vector" part of the NURBS structure and instead creating what is called a "uniform knot vector" which basically means every control point has the same amount of influence on the spline shape.

But that's a sort of unusual thing to do with NURBS surfaces really, because as you mentioned it can cause the surface shape to mutate, sometimes in fairly odd lumpy bunched up ways if the original surface had a lot of non-uniform knot distribution...

Most of the time with NURBS surface modeling there is much more of an emphasis on preserving the shape of the curve or surface itself to some tolerance level rather than trying to preserve the particular position of the control points.

Maybe I could help you some more if I understood the particular reason _why_ you want to use this Softimage function on your surfaces to begin with... What is it that you're looking to gain by doing that?


The main focus for MoI for control point editing is to do that on curves, not really on surfaces. Once you have got some curves drawn and edited, then when you create surfaces from those curves it tries to make those generated surfaces be accurate to follow the overall shape of the curves. Some operations like intersections and sweeps work by an iterative refinement method where they add as many control points as needed to the generated curve in order to achieve a certain level of accuracy to the desired "ideal" shape. With that overall workflow, the specific arrangement of surface control points isn't really all that important, it's more about viewing the surface as a sheet of geometry that you know follows your profile curve shapes to a high level of accuracy.

And yes, this whole workflow is indeed quite different from the 3D point control-cage-focused type process that you'll be focused on with a polygon mesh modeler.

In a sub-d modeler, the process is kind of more like sculpting in a way, where you're doing a lot of squishing of points around on the 3D shape. MoI is designed with a pretty different workflow in mind from that, in MoI it's sort of more like illustration rather than sculpting, where you focus more on drawing 2D curves to define key profiles of the shapes, and then surfaces are generated directly from those profiles and not really squished around themselves directly in 3D.

The MoI workflow tends to work very well for man made objects that are often times well defined by 2D profile curves. By being able to do most of the work in 2D these kinds of shapes can be constructed very rapidly with MoI's main intended approach.

For shapes that are more lumpy organic type things, those are usually not very well defined by 2D profile curves and those type of shapes are usually better to do in a sub-d modeler instead of in MoI.

There is some limited amount of 3D surface control point squishing in MoI as well, but it's not really primarily designed to do things that way.

- Michael