Terren,
No - in as much as Java 3D only ultimately renders triangles.
I don't follow your reasoning that "it would be completely obvious if it was
just a triangle mesh, no matter how dense", though. Theoretically the mesh
could be so dense that adjacent vertices ended up being mapped to the same
screen pixel, which can't be beat...(?)
You are quantizing a projection of your "mathematically correct" sphere onto
a pretty coarse grid of pixels anyway - and that is after the renderer
converts your mathematical description (NURB, I'm guessing) into a triangle
mesh for projection.
NURBS aren't supported in Java 3D but you could write your own NURB
tessellator for Java 3D. There is an implementation in one of the Graphics
Gems books.
See the gluNurbsProperty in the OpenGL docs for how OpenGL *dynamically*
controls tessellating NURBS, but tessellate them it does. The default
tessellation creates polygons with a length of 50 pixels, certainly visible.
Sincerely,
Daniel Selman
[EMAIL PROTECTED]
http://www.tornadolabs.com
> -----Original Message-----
> From: Discussion list for Java 3D API
> [mailto:[EMAIL PROTECTED]]On Behalf Of Terren
> Sent: 17 January 2000 21:43
> To: [EMAIL PROTECTED]
> Subject: Re: [JAVA3D] pure spheres
>
>
> What I'm hoping for is that I'll be able to extend java3d in such
> a way so that
> I could create a new 'primitive' - a sphere, mathematically
> correct. I want to
> model a spherical mirror, and it would be completely obvious if
> it was just a
> triangle mesh, no matter how dense. I wish I had the time to
> immerse myself in
> j3d to see if what I want to do is possible, so I'm hoping
> someone here can
> tell me one way or another if non-linear geometric shapes can be
> modelled with
> it.
>
> Terren
>
> --- The Casteels <[EMAIL PROTECTED]> wrote:
> > I have attempted to use Java3d to evaluate volumetric data
> which has some of
> > the same problems you're running into. What I ended up doing
> was creating an
> > IsoSurface from the volumetric data, by using the marching
> cubes algorithm
> >
> > The result is a triangle surface mesh. The density of the mesh
> is dependent
> > on
> > how fine the cube mesh was.
> >
> > I know this isn't what you want, but it gives you an idea of
> what you're up
> > against. You'd have to do something similar by evaluating each ray into
> > recursively smaller and smaller pieces as you get closer to the
> threshold.
> > Once
> > you reach the threshold you'd have to generate a cube around
> the point to
> > establish the surface appearance, shading, normal ect.... then you can
> > establish the pixel color and plot it onto a 2D Canvas.
> >
> > A daunting task....
> >
> > In short, although I'm sure it could be done, this isn't what Java3D is
> > really
> > meant for. Java3D is designed to generate VR type worlds,
> interactive, going
> > beyond VRML to give the programer tons more control and
> functionality, but
> > still it is an interactive environment sacrificing realistic
> appearance for
> > interactivity.
> >
> > Another note, the above mentioned volumetric viewer was slow as a dog.
> >
>
> __________________________________________________
> Do You Yahoo!?
> Talk to your friends online with Yahoo! Messenger.
> http://im.yahoo.com
>
> ==================================================================
> =========
> To unsubscribe, send email to [EMAIL PROTECTED] and include
> in the body
> of the message "signoff JAVA3D-INTEREST". For general help, send email to
> [EMAIL PROTECTED] and include in the body of the message "help".
>
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
