VRML - Faces missing in viewer

Author

Faces missing in viewer

Rhiana

2004-09-01, 12:17 pm

Hi,

I am modelling in Blender 2.34, exporting to VRML 2.0 and viewing with
Cosmo Player 2.0 and glview on Windows. I am modelling a building. I
am missing the faces of some of my walls. Sometimes they appear at a
certain angle or from a certain distance, some have been replaced by
shimmering lines on the floor.

I don't understand the cause of the problem. I am new to VRML.

Can anyone point me in the right direction?

Rhiana

Joerg Scheurich aka MUFTI

2004-09-01, 12:17 pm

> im missing the faces of some of my walls. Sometimes they appear at a
> certain angle or from a certain distance, some have been replaced by
> shimmering lines on the floor.

> I don't understand the cause of the problem. I am new to VRML.

Sounds like the nasty "thin wall" (z-fighting) problem.

http://www.csv.ica.uni-stuttgart.de...xtag/waende.wrl
http://www.csv.ica.uni-stuttgart.de...zelflaechen.wrl

Graphiccards do not account depth/distances exactly, they use a very much
faster algorithm which can fail, if too walls are too near to each other.

If you can not avoid the problem by using "solid TRUE" in the IndexedFaceSet
(if the back part of your wall is doublesided, not singlesided) you have
to make bigger distances between 2 walls 8-(

so long
MUFTI
--
Wenn Ihr CD-ROM Antrieb Brief d nicht ist, tauschen Sie ihren CDROM Antrieb
fuer d in der Fuehrung aus.
(aus einem Softwarehandbuch, Stichworte: drive und letter)

Rhiana

2004-09-02, 12:18 pm

Joerg Scheurich aka MUFTI <rusmufti@helpdesk.rus.uni-stuttgart.de> wrote in message news:<ch4a4f$lag$3@infosun2.rus.uni-stuttgart.de>...
>
>
> Sounds like the nasty "thin wall" (z-fighting) problem.
>
> http://www.csv.ica.uni-stuttgart.de...xtag/waende.wrl
> http://www.csv.ica.uni-stuttgart.de...zelflaechen.wrl
>
> Graphiccards do not account depth/distances exactly, they use a very much
> faster algorithm which can fail, if too walls are too near to each other.
>
> If you can not avoid the problem by using "solid TRUE" in the IndexedFaceSet
> (if the back part of your wall is doublesided, not singlesided) you have
> to make bigger distances between 2 walls 8-(
>
> so long
> MUFTI

Hi MUFTI,

This is not actually my problem. Some faces of walls are actually
missing altogether, (the walls are vertical) the shimmering lines are
on the floor where the walls would meet the floor. Like this:
______________
//////////////

Actually that makes sense.. the shimmering IS most likely caused by
the z-bufffer problem where the bottom face of the rectangular prism
(wall) and the floor meet, however it does not explain where my walls
have gone!

The faces of the walls do exist in Blender. What could this be? Is
this a common thing for viewers to do?

I have a series of errors in Cosmo Player all like this one:

* IndexedFaceSet has a face defined with less than 3 vertices
(starting with element 1417)

But I don't see how it relates to my walls which clearly have 4
vertices to a face..

Help appreciated,

Rhiana

Rhiana

2004-09-02, 12:18 pm

Many thanks to those of you who replied to my post I have found the
solution.

Simply to set the IndexedFaceSets solid field to FALSE.

Thanks

Below I have included the article in which I found the answer - for
those of you who happen to run into the same issue.

Kind Regards
Rhiana

In article <339b8b75.8369495@news.why.net> Shannon Davis,
in3d@spam.me.not.why.net writes:
>So it appears that VRML 2.0 has a problem with double sided polygons.

It's not a bug, it's a feature. VRML 2.0 IndexedFaceSets and
Extrusions
have a field called "solid", whose default value is TRUE. The effect
is
that the faces will only be rendered when seen from the "outside" side
(where "outside" is defined by the ordering of the vertices and the
"ccw"
field of these node types). Set "solid" to FALSE in these nodes to
cause
the faces to be rendered when seen from either side.

The purpose of this feature is to help the browser optimise the
rendering
of solid objects which the viewer is not intended to walk inside, and
whose faces can therefore be assumed never to be seen from the inside.

-- Richard Kennaway, jrk@sys.uea.ac.uk, http://www.sys.uea.ac.uk/~jrk/
School of Information Systems, Univ. of East Anglia, Norwich, U.K.

Rhiana

2004-09-08, 12:17 pm