Re: [osg-users] osgviewer stereo bug

2009-07-21 Thread Christopher Mire
I've started looking at code, but have no familiarity with this codebase,
and time is very limited.  I was wondering if someone code offer
assistance on code, as in what functions to look at.   I was looking
at osgGA src, and found Animation path code, and in
osgViewer/ViewerEventHandlers.cpp the part where it switches
to animation path camera, but I'm not sure where to look to
start playing around with code, since I don't know how it
works yet.   Just looking for additional details that help me
get started on problem.  Thanks.


 Hi Christopher,

 I think the problem is that fusion distance is not being calculated.

 Most of the manipulators calculate fusion distance using an
 intersection test or distance to the center of the bounding sphere of
 the specified node.

 Probably the camera manipulator to handle animations is not updating
 the fusiondistance so the eye separation is incorrect.

 A workaround could be set an EventCallback inside the root node and
 inside the FRAME event calculate properly the fusion distance. In an
 EventCallback you have the node, the Viewer and probalby all you need
 to compute it correctly. You can approach the fusion distance using
 simply a distance to the center of the bounding sphere of the node or
 if you want a more accurate eye separation (for example for terrains)
 you can calculate a test intersection.

 Hope this helps,
 Rafa.



 On Tue, Jul 7, 2009 at 4:19 PM, Christopher Mirecnm3...@gmail.com wrote:
  Hello, I have encountered this bug where stereo rendering is messed up
  when using saved animation path file. ?This bug causes the separation
  between left and right eye images to be way off, messing up stereo
  effect. ? Upon initial loading of model, the images are not proper,
  but only when using the -p ?flag with animation file. ?Without using
  -p option, the rendering is fine. ?I am using an OBJ file, and have
  duplicated this bug with OSG version 2.8.0 from fedora 11 repository,
  as well as 2.8.2-r2 and svn head as 7-6-2009. ?I have duplicated bug
  when default stereo mode, as well as QUAD_BUFFER.
 
  Since two images are off on initial loading, I'm guessing its some
  initialization issue, I suppose with camera. ? The two images move in
  sync as camera is rotated and translated, but since they are off
  initially, stereo effect just doesn't work. ?It looks like the eye
  separation is just really huge, maybe like 10x more than it should be
  or something. ?Please advise on further steps to resolve this issue.
 
  Thanks.


___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


[osg-users] osgviewer stereo bug

2009-07-07 Thread Christopher Mire
Hello, I have encountered this bug where stereo rendering is messed up
when using saved animation path file.  This bug causes the separation
between left and right eye images to be way off, messing up stereo
effect.   Upon initial loading of model, the images are not proper,
but only when using the -p  flag with animation file.  Without using
-p option, the rendering is fine.  I am using an OBJ file, and have
duplicated this bug with OSG version 2.8.0 from fedora 11 repository,
as well as 2.8.2-r2 and svn head as 7-6-2009.  I have duplicated bug
when default stereo mode, as well as QUAD_BUFFER.

Since two images are off on initial loading, I'm guessing its some
initialization issue, I suppose with camera.   The two images move in
sync as camera is rotated and translated, but since they are off
initially, stereo effect just doesn't work.  It looks like the eye
separation is just really huge, maybe like 10x more than it should be
or something.  Please advise on further steps to resolve this issue.

Thanks.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org


Re: [osg-users] osgviewer stereo bug

2009-07-07 Thread Rafa Gaitan
Hi Christopher,

I think the problem is that fusion distance is not being calculated.

Most of the manipulators calculate fusion distance using an
intersection test or distance to the center of the bounding sphere of
the specified node.

Probably the camera manipulator to handle animations is not updating
the fusiondistance so the eye separation is incorrect.

A workaround could be set an EventCallback inside the root node and
inside the FRAME event calculate properly the fusion distance. In an
EventCallback you have the node, the Viewer and probalby all you need
to compute it correctly. You can approach the fusion distance using
simply a distance to the center of the bounding sphere of the node or
if you want a more accurate eye separation (for example for terrains)
you can calculate a test intersection.

Hope this helps,
Rafa.



On Tue, Jul 7, 2009 at 4:19 PM, Christopher Mirecnm3...@gmail.com wrote:
 Hello, I have encountered this bug where stereo rendering is messed up
 when using saved animation path file.  This bug causes the separation
 between left and right eye images to be way off, messing up stereo
 effect.   Upon initial loading of model, the images are not proper,
 but only when using the -p  flag with animation file.  Without using
 -p option, the rendering is fine.  I am using an OBJ file, and have
 duplicated this bug with OSG version 2.8.0 from fedora 11 repository,
 as well as 2.8.2-r2 and svn head as 7-6-2009.  I have duplicated bug
 when default stereo mode, as well as QUAD_BUFFER.

 Since two images are off on initial loading, I'm guessing its some
 initialization issue, I suppose with camera.   The two images move in
 sync as camera is rotated and translated, but since they are off
 initially, stereo effect just doesn't work.  It looks like the eye
 separation is just really huge, maybe like 10x more than it should be
 or something.  Please advise on further steps to resolve this issue.

 Thanks.
 ___
 osg-users mailing list
 osg-users@lists.openscenegraph.org
 http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org




-- 
Rafael Gaitán Linares
Instituto de Automática e Informática Industrial  http://www.ai2.upv.es
Ciudad Politécnica de la Innovación
Universidad Politécnica de Valencia
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org