United3dArtists Archive

Ok... since they were introduced in version 5 I've never really understood what a NURB is and how it could actually help me in my modeling. All I do understand is that it's like an automatically quad-divided primative and I can understand how it may help in organic type stuff (character modeling etc), but what advantages would it give me for solid modeling like for example a building or a piece of furniture?

Deep subject!

NURBS are a form of spline modeling, where you use the splines to create a surface profile.

Think 'slices' or 'cross section' and you're getting close. You can have one or more splines that define the shape of a surface in each of the three axes.

You can start with NURBS primitives or draw your own with the 'Add Curve' tool. You can start with or without a Drawpanel- think of it as an invisible, non rendering sheet of paper that you draw the spline on.
You can draw a shape- say the profile of a piece of crown molding, and then either turn that into a closed polygon object and /manipulate/sweep/ bevel/ etc. it with normal polygon editing tools; or you can keep it a nurbs object and use a different spline to define it's length.

I really recommend a slow read thru the relevant chapter in the manual. It explains it much better than I do!
There's a lot to it and it's definitely quirky in places.

There may be copies of some of the training vids on yootoob.

I saved one of the Caligari site tutorials and have it posted here.

Powerful set of tools, but they can be headache-inducing until you 'get it'. There are certainly better spline modelers out there, but there is still lots of cool stuff you can make with them.

I'm slowly starting to understand. I was never really clear on nurbs and splines when I was learning this stuff on my own back in the day.

Hey lrdsaytr8.

NURB stands for: Non-Rational Uniform Rational B-Spline. It's a specific type of curve, defined by a mathematical equation. The final 'NURB' curve is a summation of pre-defined weighted curves in 2 dimensions. The curve has control points; these are the 'weights'. Each control point has access to the curve over a specifc domain, or range. So if you tug on a control point in the middle of the curve, that adjustment does not affect the endpoints, since its 'power' over that area = 0. The power the NURB curve is, the user (modeler) only has to adjust the control points, not every vertex in the mesh, in order to get smooth, organic looking shapes. So dramatic changes can be achieved very quickly with little effort by the user. When the user tugs on a control point, only the vertices that are within that control point's domain get updated by the CPU. The rest of the vertices in the mesh are ignored. As a quick example: it's faster to update 5% of the NURBS surface, as opposed to 100% of the NURBS surface.

When you have an array of NURB curves, you end up with a nurb surface (NURBS).

There's other stuff such as trim curves, when you 'cut a hole' in the NURBS surface. That's some fairly complicated math stuff indeed.

That may not be the answer you were looking for though.

Froo