I've encountered the same problem, and setting the
context-getState()-setUseModelViewAndProjectionUniforms(true);
for each context didn't work in any case (with a multi-pass offscreen
rendering).
So the simple solution for me was to use the #version 150 compatibility
profile, which let's you
I'm quite desperate on this one,
any advises how to setup the camera in the second pass, anyone?
Hello again,
In my project I need to setup a second render pass that performs the
following:
Render the scene from the original camera's view (as in the first RTT
pass).
Set the second camera
Hi
you might use osg::Switch
see
http://www.openscenegraph.org/projects/osg/browser/OpenSceneGraph/trunk/examples/osggameoflife
example
( see ProcessPass::ProcessPass(osg::TextureRectangle *in_tex,
osg::TextureRectangle *out_tex, )
and then there is osgPPU
Hi,
forgot so follow up - concerning osg::Camera::FRAME_BUFFER_OBJECT
for attaching depth etc to frame buffer object you might want to
check this page
http://www.opengl.org/wiki/Framebuffer_Object_Examples
Regards
Sergey
___
osg-users mailing list
Hi Ga,
First up I would recommend moving away from GLUT as it just
complicates things, the basic osgViewer::Viewer is far more flexible,
feature rich, scalable and easier to use.
Second, the vertex data in the OSG is stored in a osg::Vec3Array which
is implemented with a std::vectorVec3 so you
Hi Pete,
This rather sounds like a driver issue. Could you try running your
tests on another system with different hardware and drivers.
Cheers,
Robert.
On Thu, Sep 29, 2011 at 9:05 PM, Peter Amstutz
peter.amst...@tseboston.com wrote:
Hello all,
I have a performance regression with 3.0.1.
Hi Sebastian,
as for depth buffer - you might try ( not sure though ) two approaches
a) use depth replace shader ( replace current depth with provided depth )
b) use depth testing in shader - such that you manually compute depth
value and compare it with attached depth texture in shader see
Hello Sergey,
Thanks for the hints so far. I'll check the approaches.
Hi Sebastian,
as for depth buffer - you might try ( not sure though ) two approaches
a) use depth replace shader ( replace current depth with provided depth )
b) use depth testing in shader - such that you manually compute
Hi,
What happens if you change the threading mode?
I am just wondering if its related to the GPU stats issue and stuttering,
changing the threading mode seems to clear it.
Have a search in the forums, for the subject 'stats issue with SVN'?
Regards
Martin Naylor
-Original Message-
From:
Hi Wang,
Don't worry about this. The title may not be able to tell the methods
used directly. In fact there are already several examples showing how
to integrate OSG and OpenGL in different ways:
[Ch2] Using draw callbacks to execute NVIDIA Cg functions: use camera
draw callbacks to call NVCG
Hi,
I'm trying to get osgShadow to work with Vizard, an OSG based VR toolkit. I
wrote a plugin which inserts a shadowedScene node and it all seems to be ok,
but the shadows are not there. I made a simple scene (ground, duck and 2 balls)
as shown in the first picture. There is one light,
Hi,
I had the problem in my application, that the picture was frozen for certain
camera perspectives. This is caused due to the near far auto computation, that
also takes the geometry of the osgPPU::Unit's into concern.
For the faulty case, updateCalculatedNearFar(..) in
Hi,
I had the same issue - please find my fix here:
http://forum.openscenegraph.org/viewtopic.php?p=43140#43140
It prevents the osgPPU::Unit's from beeing taken into account for near/far
computation.
--Alex
--
Read this topic online here:
J-S,
Thanks for the reply.
I hadn't sat down and compared matrices yet, but I had debugged to ensure I was
getting something. And yes, I'm accommodating for the fact that it's a
ModelView (and not just a Model) matrix. I've also been using the osg_
attributes to this point, in lieu of the
Hi Alexander,
Please find attached fix for osgPPU::Unit that prevents it's geode from be
taken into account for near far computation.
Nice to see you fixed this at the source, I had just overridden the
Processor class and disabled near/far computation for the whole Unit
subgraph, but
Hi Joel,
So far as your optimization tip goes, I assume you're meaning to apply the
shader at the transform node as well? I've been doing that sort of thing
already, just not on my Transform nodes...
No, I mean that all non-Transform children under a Transform will have
the same
Hi, I have some doubts about my approach to create a GL3 non FFP context.
I have compiled osg with all the cmake settings for GL3 as explained in
lots of other posts.
When I create an object ( a simple two Vertex line ), usual camera
manipulation is working without me having writrten a single
Hi Peter,
I don't know about your other questions about the GL 3 context, but I
can answer this since I've been looking at it recently.
How do I get these default osg_ Uniforms into the StateSet ? The only
place where I find code for creaating them is osgUtil::SceneView, which
is marked as
I agree, your model has blending enabled, and the face has non-1.0 transparency.
It'll be a couple days before I can dig into this. When I fixed the rgb part of
this issue, I looked through the flt spec, and don't remember seeing anything
about how transparency should be handled. But I'll take
Hi,
Cool to see this fixed. Did you post this on the osgPPU trunk? I don't seem to
see any recent change there...
By the way, what version of OSG are you using to compile it? I just tried
osgPPU 0.4.2 with OSG 3.0 and there seem to be a few changes to make to make it
compile.
Luca
Hi Luca,
Cool to see this fixed. Did you post this on the osgPPU trunk? I don't seem to
see any recent change there...
I think Alexander just posted a modified file, the changes haven't been
integrated into osgPPU yet.
By the way, what version of OSG are you using to compile it? I just
J-S,
Of course. It's not that I didn't know state is inherited by child nodes, but
for some reason, my testing of uniform inheritance failed, so I was manually
propagating uniform pointers... Took 30 seconds to disable that code and prove
your point. Thanks again.
- Joel
22 matches
Mail list logo
