Solid Modeling

[mrwolfy]

There used to be a software called Gig that offered animation and modeling with "solids". The concept basically is that you have mathematical representations of geometric forms, instead of vertice based models. The advantage was that you had resolution independant forms, instant, perfectly crisp boolean operations, and highly smooth metaballs-like transitions. Not practical for producing character animation or whatnot, but quite interesting for some types of animation.

 This type of modeling software exist today?

[joojaa]

All Mechanical engineering software work on this paradigm. So sftware of this kind include:

Catia
Inventor
Pro|e
Solidworks

But a lot more exist.

The concept basically is that you have mathematical representations of geometric forms

Actually thest not the idea, the idea is that you have functionality to ensure thats what you can get. In parctice they work just like surface modellers EXCEPT they go areound validating the suer all the times. So in parctice they dont doffere rom using maya and modelling with nurbs booleans all the time, exccept yyou are not allowed to make mistakes. Wich lessens chance of bboolean failing. IN parctice tough the booleans still fail from time to time.

[mrwolfy]

Hmm, perhaps the Gig software was different, a different definition of Solid?... As I remember there was no chance of failure of a boolean operation. In fact the boolean only appeared in the render. You would just interset shapes, and the raytracer would calculate the boolean, if that is the right word. Everything could be animated on the fly. Was good for certain types of animation.

[joojaa]

Aaah yeah they are called implicit surfaces. Different paradigm.

PRman does them, any shape of them in fact. As well as maya fluids. Blobby particles are also implicits but just round ones.

I sufggest also mathematica as its very cool for implicit modelling. I belive cinema3d has a host of them too. Thing is tough the triangulator thet makes implicit surfaces is easy to do, its just potenttially slow. You can get code examples from NVIDIA.

As I remember there was no chance of failure of a boolean operation. In fact the boolean only appeared in the render.

Theres ALLWAYS a chance for boolean to fail (yes even implicits fail to sometimes create the limit surface the boolean thechnically dindt fail but the end result did), but solid modellers are VERY good normal users will never spot a boolean failure. IN fact booleans dont fail much in maya either IF and only if the user is making sure the environment is correct, see maya doent have theese checks.*

You can do rendertime booleans in mentalray and prman. They will allmost never** fail except when theres aopen side to them in wich case they will just make render artefacts

• for example i can not remeber when ive last failded to get a boolean to work when i have a surface thats actually clean and able to boolean. Cant remeber a situation where i couldnt clean the surface. Its just a matter of skill thats all. See its jut a matter of modelling something that can actually work.

** almost never means for you never if the surface is closed.

[mrwolfy]

thanks!