Hello Carsten,

> > > hm, I'm afraid nothing comes to mind that would explain this. One
> > possibility is that the Stage used to render into your FBO is itself being 
> > culled
> > - can you set the volume of the stage's node to infinite?
> > 
> > The size of the stage is set to (0.0f,0.0f,1.0f,1.0f). Is that what
> > you meant? Or do i have to set it to something like
> > (-OSG::Inf,-OSG::Inf ,OSG::Inf ,OSG::Inf )?
> 
> no, that is the portion of the FBO written to by the stage (kind of like the 
> viewport for a window). I meant that since a stage is a type of NodeCore it 
> has a Node which places it in the scene (usually somewhere close to the root) 
> - I was referring to that Node's bounding volume. It also means it is subject 
> to frustum culling, just like any other node. So if the stage is a child of 
> the scene root, it is in the origin and has a bounding volume that contains 
> the scene to be rendered into the stage (i.e. the stage's children). 
> Depending on what you render into the FBO and how you use it, you may want to 
> render it even if the stage is not inside the camera frustum, in that case 
> you can make the node of the stage have an infinite bounding volume.
> Don't think that is the problem here though, see below.
> 

Ok, in my case the stage is the root node right now, but this will
change later.

> > > Beyond that I'm not sure where to look for the problem, it sounds a bit
> > as if somewhere the modified near value is cached and not updated
> > correctly, but why does it update the first time you change it? Can you run 
> > your
> > program under an OpenGL debugger [1] to see if the scene is being frustum
> > culled by OpenSG (i.e. no draw calls are being made), is outside OpenGLs 
> > camera
> > volume (i.e. camera values are not correctly updated) or if for some
> > reason the read back stops working? That would be very helpful to figure out
> > where to look more closely at the code.
> > >
> > 
> > I've run the programm and send you the files off-list.
> 
> Thanks! Looking at the log I've noticed that at one point this changes:
> from: glClearDepth(1.000000)
> to: glClearDepth(0.000000)
> 
> So the problem is that you are clearing the depth buffer with the value for 
> objects 'in' the near plane and nothing will pass the depth test afterward. 
> Do you at some point change the clear depth of a background perhaps?
> 

Yes, I changed the clear Depth. But this does just effects, that the
values after the near-plane is behind the geometry are not '0' but
'255'.

It does not affect, that the image values are not changing after that.

Perhaps I might describe the behaviour again, I could have made it not
clear enough.

I start my program and setup the scenegraph. Then I render the scene 4
times in one function, with different states. After the 3rd rendering I
print out the stencil values.

When I first start the rendering the image shows, what I expect.
Then I render more times with different near plane values and the image
does not change (but it should) until i reach the point, where the
nearplane is beyond the geometry.
Then the buffer shows just '255'.

After that I change the nearplane value back to where the geometry
should be visible, but nothing happens with the image.

I hope this helps a bit to understand what I mean.

I thougt it could be, that I forget to change the States to the standard
values after rendering my 4 times, but I can't explain the sudden
change, when the nearplane gets behind the geometry, with that
explanation.


thank you
Constantin








------------------------------------------------------------------------------
Download new Adobe(R) Flash(R) Builder(TM) 4
The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly 
Flex(R) Builder(TM)) enable the development of rich applications that run
across multiple browsers and platforms. Download your free trials today!
http://p.sf.net/sfu/adobe-dev2dev
_______________________________________________
Opensg-users mailing list
Opensg-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/opensg-users

Reply via email to