Hi,

I'm trying to cookie two NURBS tubes to create a cannon barrel + the braces that go around it, need to cookie so that you can still see down the end of the barrel.

As far as I'm aware the cookie SOP only works for polygons, but obviously for a cannon I need a round surface.

Is there a cookie like operation/SOP that can be used on NURBS?

Surfsect.

Make sure you keep the seems in the Nurbs surfaces out the way of the are you want to join or cut.

Surfsect is the way to go, except that I've (very recently) found that if you have are intersecting dense NURBS geometry (I was intersecting a tube into a sphere and wanting a rounded fillet), the parameters are very hard to dial in correctly. I hade to unrefine my surfaces to get any usable solve at all actually. But when it works, it's a beautiful thing.

On real dense stuff you have to play around for ages just knocking the values around little amounts til it works. Then when it does lock it, cos the next day may be a different story…. :roll:

Yeah, I've taken to doing that too. Sometimes the trim and the extracted profile don't match too. It *seems* to me that have really small marching step size and a large number of steps should get you a better and better solution, but it rarely seems to work that way. A pretty confusing tool sometimes…

I really wish SESI would improve the NURBS modeling SOPs. In particular the SurfSect and Round SOP. I still prefer NURBS for a lot of modeling, but sometimes Houdini just won't do and I have to resort to other packages for (NURBS) modeling.

Also, it'd be nice to see some kind of solid modeling implementation. That would be a very powerful modeling asset, especially combined with Houdini's procedural nature.

I think that with the increasing popularity of subd's, nurbs modelling isn't as common as it once was.

I indeed tend to use sub-division surfaces most of the time, although I still use NURBS very often. Therefore I hope that the NURBS workflow and toolset will be improved and expanded in the future, it's actually a great desire of mine.

Solid modeling tools/workflow would truly make a great addition; especially in combination of Houdini's procedural nature, fully scriptable too, just think of the possibilities! Houdini would then be even more complete.

I think they both have their strengths. Subdivs are great for organic things, especially characters, whereas I still think when you're working with design, vehicles, architecture that NURBS are still where it's at. Certainly the whole design manufacturing market is all NURBS. If you're having to import and massage data from this market(as I have recently ), having a more robust NURBS importing, massaging and creation toolset would be a good thing. I still find the houdini toolset to have a lot of oddities in this regard.

AFAIK there's only one input path for NURBS(inc trim surfaces), and that's IGES. IGES lives in a special little circle of hell that has so many variations and lack of standards it's got to be the worst possible vector.

Cheers,

J.C.

Single biggest problem with Nurbs in Houdini is exactly what you describe Nurbs models come from CAD packages and CAD all work at double precision, Houdini works at single precision…. even if you can get an iges to import tons of stuff will stop working just for that reason. Rhino is pretty good at sorting out these issues but until everything is in double precision floats I think it will continue to be a problem even if you can read every form of iges known to man.
my 2 cents.

I wonder if SESI has any plans to expanding to double float precision, the NURBS modelling toolset and (geometry) data I/O.