Hellor Robert,
Every day I try to think a bit less, but it does not work for me... ;)
But seriously:
OSG is a framework, which tries to fit several different use cases. And it is
natural, that it will not fit perfect for every possible job. Thus I want to
figure out, how I can use it in the
Hello Sebastian,
unfortunately I cannot create bigger sets with more GL geometry in them, cause
I need the geometry to be addressable. :(
Sebastian Messerschmidt wrote:
Hello Alexej, Robert
I think the reason for Alexej to think that disabling the culling would
improve render time in
Hello Sebastian.
Sebastian Messerschmidt wrote:
For this i used the setComputeBoundingBoxCallback on the geometry in
question to set it to the correct bounding box.
Maybe you could use this on your first-class Node?
Hm, maybe.
Unfortunately the documentation to
Sebastian Messerschmidt wrote:
The callback will be called if the bounds are marked as dirty. (i.e. if
a new PrimitiveSet is added).
In my use case I simple set the callback, called getBounds once on the
parent and then removed the callback after loading and before attaching
the node to
gwaldron wrote:
No, you need to apply it on the drawables, otherwise they will still get
culled.
I created some utilities to support DrawInstanced usage in osgEarth,
including a visitor that will apply the static bounding box to drawables.
Perhaps it would be useful to you:
hybr wrote:
there are problem: cullvisitor grabbing drawables for later drawing, and if
you dont go deeper, nothing would be drawed
Cheers.
The point is: At the group node it can be already decided, if the underlying
nodes are visible or not. Thus no need to compare against each single
Hello,
I need an advice regarding the culling setup.
In my application I use a vertex shader to transform 3d geometry from the
flat-world scene graph to the related position on a sphere (globe).
Having a normal camera, which hovers over the transformed spheric world breaks
the culling stage,
Hello Robert,
tank you for the fast replay!
Grep-ing the includes reported, that two classes have the setInitialBound()
interface: Drawable and Node.
I would like to avoid setting this initial-bound for each single peace of
geometry (~ Drawable).
Can I just put the initial-bounds on the
robertosfield wrote:
Hi Alexej,
You don't need to change the render-target to re-render a tile. I
can't work out exactly what you mean in your scheme but it sounds a bit odd.
The problem is not the re-rendering of a single tile. The job is to re-render a
tile subset from a collection of
robertosfield wrote:
Hi Alexej,
The main question I have, and I'm sure others have is, what are you
trying to do and why. Providing low level advice when one doesn't know
the high level context is likely to be mis-directed and mis-informed.
Sorry for confusing, I thought the job of
Hi folks,
// OSG 3.0.1
I currently observe, that my userData instances (derived from osg::Referenced)
are not dereferenced, if I delete the carrying node instance. The ref counter
in myUserData instance gets incremented by calling
node-setUserData(myUserData). But if delete the node (by
It was not sarcastic, I really thought it is a common case (looks like at least
there is a special class: OverlayNode, which tries to address similar
functionality ;) ).
I will try to discover, what is osgSim::OverlayNode all about. But here we
again face the documentation problem. The only
Hello Robert,
the reason for changing the FBO attachment is generating dynamic ground tiles
on the fly in a separate ground-scene-graph. Each ground-tile is a separate,
location bound texture (with allocated image). If the tile location gets dirty
I want to re-render the ground tile (the
Hello,
unfortunately this does not work for me.
Maybe I have to do something special in the FirstPass_ or setup the camera in a
special way -- I don't know.
Can someone clarify is here something special missed or what is right way to
exchange the FBO render target the light-way (without
Hello together,
Thanks for the help! I will try the cheaper approach next.
But in general, the current behavior of ignoring the color_buffer binding looks
more like a bug to me. Cause having such essential changes on the camera
settings, I would expect the renderer to be aware of that!?
Hi folks,
additional observation:
In case, where for each frame in the update-callback the color-buffer
attachment is bound to new Image instance, and all outcoming pictures are all
gray, except the first one -- I now store the reference to the initial
color-buffer bound image, and write it's
Hi hybr,
yes, it works now!
I was already on finding out, how can I change
RenderStage::_cameraRequiresSetUp per frame. You've shorten this a lot!
Many thanks for your help!!!
Cheers,
Alexej
--
Read this topic online here:
Hello folks,
the main intention is to use an off screen RTT scene setting to generate
textures for the main scene asynchronously. The RTT base seems to work, but
rebinding the color_buffer at each frame to a new target (image or texture)
does not work. (During the development osg::Image
Hi,
finally I found the bug.
The build framework was using the /MTd libc parameter per default. Adding OSG
(which seems to force the use of /MDd libc) caused the heap corruption.
Cheers,
lexar
--
Read this topic online here:
litllechicken wrote:
Hi, thanks for triing to help me, i found i was not due to OSG but to my
visual config witch failed.
Hi,
seems that I have a similar problem: crashing in Intersection destructor with
corrupted heap.
Can you please explain, what was wrong with your VS config?
// I use
Hi folks,
Unfortunately this is not an answer, but a similar question:
I'm also playing with Qt integrated OSG (as the osgViewQt suggested) and ask
me, how to modify the scene-graph in safe way? Cause there is no explicit
rendering loop (as mentioned in many examples):
Code:
while(
Hi Robert,
thanks for fast replay!
The qt-example comes with osg-2.8.2 for windows (and does not contain a real
rendering loop, just the frame() call).
It's not about a game project, it's about a GUI tool.
The ViewerBase::frame(..) function seems to contain the whole rendering
sequence
22 matches
Mail list logo