On Thu, Apr 10, 2008 at 8:07 PM, Daniel Sperka <[EMAIL PROTECTED]> wrote:
> I ran into a problem like this a while back. Not as simple as I'd have
> liked it to be, that's for sure.
>
> You can't just make a quad "extending to infinity" because ultimately
> you have to feed actual coordinates to opengl.
well, I think it is possible to make an infinity plane using homogeneous
projective-3 coordinates... I realize after testing it that any coodinate
with w=0 represents only a direction within homogeneous projective-3d
space.. so, if you have 4 vertices in a quad with w=0 than you have 4
directions... but then you will not be able to do something like a regular
quad : (1,0,-1,0), (-1,0,-1,0),(-1,0,1,0), (1,0,1,0). It is because
(1,0,-1,0) and (-1,0,1,0) are not perpendicular directions.. so, in this
example you have 4 directions but only 2 perpendicular directions... the
directions are collinear 2 by 2...I thnk if you do something like
glNormal3f(0,1,0);
glBegin( GL_QUADS );
glVertex4f( 1,0,0,0 );
glVertex4f( 0,0,-1,0 );
glVertex4f( -1,0,0,0 );
glVertex4f( 0,0,1,0 );
glEnd();
you'll then have 4 perpendicular directions within the xz plane, and then
your infinity plane is created! ;]
since w=0, the camera translation has no effect within the relative position
of the plane, only the rotation since
p' = [Proj][R][T][ 1 0 0 0 ]^t (the las col of translation matrix will
multiply 0)
... but it think theres a problem... if you put textures within this plane I
dont know how the perspective correction of opengl will look like and I
think the relative position of the textures will never change related to the
observer.. so if you have tiled texture, it will seem that the floor is not
moving, only rotating... some how very strange
so one will need to do something pretty like you've described, something
like a projective grid within the clip space.
>
> You can make a really big quad, assuming you can ensure that your
> viewing volume will never extend beyond the edges of it.
yep. for now it is exactly what I did... I only needed something very simple
by now ;]
>
>
> I needed my ground plane to have random dots on it, and the camera moves
> around the ground plane, so it was impractical to generate a single
> ground plane prior to the application running. My solution was to create
> a custom node core OSGGroundPlaneCore. When it is rendered, this core
> fetches the camera beacon and the camera's viewing volume parameters
> (near plane, far plane, etc). I use all that to compute the viewable
> region of the plane, and generate graphics which cover only that region.
> I use pre-computed grids with plain colors and random dots, though the
> technique could be extended to to use textures or simple colors/patterns.
>
> That's considerably more complicated than the simplest solution, so I
> won't bore you with the details unless you really need them. Let me know
> - I'd be glad to share the code with the list.
well, it would be great if you can share your code ;] Thanks again
>
>
> Another option, though its one I know little about, is to use a terrain
> generator.... it seems there has been some discussion of these on this
> list, and I've seen or heard of them before. These are for generating 3D
> surfaces, not planes. YMMV.
;]
Thank you very much Dan!
>
>
> Dan
>
>
>
> Pablo Carneiro Elias wrote:
> > Hi there.. is there any way to create a plane extending to infinity in
> > every direction? as the ocean surface or something like that... I've
> > tried w = 0 using osg::Vec4f , but it doesnt work...
> >
> > any hints?
> >
> > its is just a geo quad extending to infinity (creating horizon line).
> > In fact, extending to infinity means that translation has no efect,
> > only rotation... it means that if there is any whay of specifying that
> > my coordinates are already within camera coordinate system it will be
> > perfect.. is there any way to do this or I'll need to take the inverse
> > translation of the camera and set it as the node's translation? In
> > OpenSceneGraph there is a flag that sets a transform as being within
> > camera coordinate system.
> >
> >
> > thanks
> >
> > []'s!
> >
> > Pablo
> > ------------------------------------------------------------------------
> >
> >
> -------------------------------------------------------------------------
> > This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
> > Don't miss this year's exciting event. There's still time to save $100.
> > Use priority code J8TL2D2.
> >
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > Opensg-users mailing list
> > [email protected]
> > https://lists.sourceforge.net/lists/listinfo/opensg-users
> >
>
> --
> Daniel J. Sperka, Ph. D.
> UC Davis Center for Neuroscience
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
> Don't miss this year's exciting event. There's still time to save $100.
> Use priority code J8TL2D2.
>
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> _______________________________________________
> Opensg-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/opensg-users
>
-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
Don't miss this year's exciting event. There's still time to save $100.
Use priority code J8TL2D2.
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
