Subject:
|
Re: Sphere tesselation
|
Newsgroups:
|
lugnet.cad.dev
|
Date:
|
Mon, 11 Sep 2000 19:24:12 GMT
|
Viewed:
|
502 times
|
| |
|
|
"Rui Martins" <Rui.Martins@link.pt> wrote in message
news:Pine.GSU.4.10.10009111622520.12940-100000@is-sv...
> On Sat, 9 Sep 2000, Travis Cobbs wrote:
>
> > Does anyone know how to perform an arbitrary sphere tesselation which • ends
> > up with the tesselation used by the LDraw sphere primitives when set to • a
> > detail level of 16 (16 being the number of facets around the • circumference
> > at the equator). I'll have to be able to run it for an eighth of a • sphere,
> > of course, since that is the base sphere unit in LDraw.
>
> As far as I know, the sphere is incorrectly tesselated in Ldraw, because • it has
> non-planar quads.
>
> So if there is ever going to be a matematical tesselation, which aproaches • this
> form, it won't surelly be a sphere, more like "spheroid", similar to a • sphere,
> but not quite.
Actually, I don't believe the existing one is quite a sphere. I managed to
programatically generate the edge quads (ignoring planar-ness), and some of
the points I generated (which did fall on the surface of a sphere) couldn't
be made to align with the file's points and still remain spherical (unless I
did it wrong). Specifically, the fourth point of the corner quads which
isn't along the edge doesn't appear to be on the surface of a sphere. This
is probably just rounding, though. With a sphere radius 1, 2 decimal places
is NOT enough (the file rounds to 2 decimal places).
> MAybe this primitive(s) should be corrected, to use planar polys (tris or
> quads). Then it could be made a matematical approach, which will tesselate • with
> the same or similar points.
I'll put my vote in for that here. In fact, if I come up with the
programatic way of generating it, I'll even provide the replacement file.
> > I want to do a primitive substitution for the
> 1-8SPHE.DAT primitive, but
> I'd > really rather not have to hard-code the points.
> > It would be nice at some point in the future to be able to increase the
> > complexity and still have my spheres and cylinders meet exactly.
>
> Not Ever !
> Even now, if you rotate the cylinder or the sphere, relatively to one • another
> (in steps different from multiples of 90 degrees, it won't match.
The current version (and the modified version we are discussing) at least
will always meet when it is the same diameter and has the same rotation (or
90 degree difference). Since all of the spherical "end caps" should fit
these criteria, any replacements should continue to allow the polygons to
line up in this base case.
> The "perfect" solution for polys, is to make the objects interpenetrate • each
> other, a let the renderer to the "pixel match work".
True, but that is only possible if the cylinder and the sphere have
different radii. If they both have the same radius, the polygons along the
edges of the sphere need to be capable of lining up with the cylinders.
--Travis Cobbs (tcobbs@san.REMOVE.rr.com)
|
|
Message has 1 Reply:
Message is in Reply To:
 | | Re: Sphere tesselation
|
| (...) As far as I know, the sphere is incorrectly tesselated in Ldraw, because it has non-planar quads. So if there is ever going to be a matematical tesselation, which aproaches this form, it won't surelly be a sphere, more like "spheroid", similar (...) (24 years ago, 11-Sep-00, to lugnet.cad.dev)
|
9 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
