Julien,
Yo may want to verify if you have CACHE_NODES enabled. If so, then you may
think there's a resource leak but in fact the scene graph leading to your
Geometry may be cached by OSG.
--
Read this topic online here:
Just to clarify that my suggestion to change the Drawable's desructor is only
if there is indeed a GL object resource leak, which I'm not totally convinced
yet (by reviewing the code).
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=74499#74499
I suggest that we remove Geometry::dirtyGLObjects() as it does nothing other
than call its base class, i.e. Drawable::dirtyGLObjects().
Code:
void Geometry::dirtyGLObjects()
{
Drawable::dirtyGLObjects();
}
I also noticed that the Drawable's destructor is calling a virtual
Got it. It's not necessarily a memory but a resource leak.
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=74497#74497
___
osg-users mailing list
osg-users@lists.openscenegraph.org
I'm a little confused. Is there or is there not a memory leak?
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=74494#74494
___
osg-users mailing list
osg-users@lists.openscenegraph.org
Hi,
Has anyone experienced issues with OpenThreads Mutext::lock()? Occasionally a
crash occurs when starting my application but only if the threading model is
not SingleThreaded. The crash occurs in the following functions of OpenThreads:
Code:
int Mutex::lock() {
Win32MutexPrivateData
Actually, if we want to adjust the threading affinity when cycling through the
threading model, then would`t be better to if we replaced:
Code:
if (isRealized() && _threadingModel!=SingleThreaded) startThreading();
by
Code:
if (isRealized() && _threadingModel!=SingleThreaded)
Hi,
I also looked at the osg-submissions archives and could not find any trace of
this change. I went back as far as June 2017.
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=74329#74329
___
Hi,
I looked at the commit but the description does not clearly provide a reason
for the change. I looked at the forum but could not find anything. I'll try the
mailing list next.
Thank you!
Cheers,
Guy
--
Read this topic online here:
Hi,
When I try to cycle through the threading models by pressing the 'm' key (when
the ThreadingHandler is registered) the stats would indicates the correct
threading model, but the engine would remain in SingleThreaded. So I started
investigating the issue and I noticed that, with OSG v3.6.2,
Hi,
This is not really a big issue, but I noticed that the OpenThreads PDB file is
not copied to the INSTALL directory when building OSG v3.6.2 on Windows with
VS2013.
I looked at .\src\OpenThreads\win32\cmake_install.cmake and the lines required
to copy the PDB are missings.
Thanks,
Excellent. I tested it and everything seems to work as expected. Thanks for the
quick response.
BTW: I'm also investigating a way to somehow "cache" the last drawable that we
intersected with (terrain only). If this works, it could prove to be a
significant performance gain in situations
Although this is an old post, I recently noticed the same issue in v3.4.0.
Guy
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=7#7
___
osg-users mailing list
If I have a chance, I'll try to look at it as well based on your suggestion
solution.
Cheers,
Guy
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=72217#72217
___
osg-users mailing list
Hi,
I'm having problem with the osg::ComputeBoundsVisitor and osg::Billboards. In
essence, the osg::ComputeBoundsVisitor does not consider the pivot position of
the facing billboard. For example, let say I have a simple model that contains
a single facing quad billboard (like a tree) located
Understood. I compared the Config file used for building OSG and the one that
is copied by the INSTALL project, which is used by my app, and they are both
the same.
Below is the content of my OpenThreads\Config header file.
Code:
#ifndef _OPENTHREADS_CONFIG
#define _OPENTHREADS_CONFIG
/*
Hi,
I was wondering if someone could explain why the
Optimizer::RemoveLoadedProxyNodesVisitor is not removing ProxyNodes with
NO_AUTOMATIC_LOADING set?
Here is the motivation behind my question... We often need to load relatively
large openflight terrains that contains a significant number
Hi,
I was wondering if anyone observed a similar problems. In essence, we recently
upgrade from v2.8.2 to 3.4.0 and we are now noticing random crashed when
launching many of our applications. Note that we also upgraded from VS2005 to
VS2013, so technically it can also be a compiler issue as
you caj also use the environment variable OSG_WINDOW, e.g.:
set OSG_WINDOW = 40 40 640 480
Note: space delimited.
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=44341#44341
___
osg-users mailing
I think that the first approach of creating your own file stream for encryption
is an excellent one, but there is a problem (I think). What would happen if the
stream's seekg()/seekp() is called. For example, if you wanted to seek to the
end of the file, you would need to decrypt the entire
Now I understand! That makes perfect sense to me now.
I tried calling GraphicContext::setClearMask(GL_COLOR_BUFFER_BIT) in my app and
now it works as expected.
Thanks... your the best.
Guy
--
Read this topic online here:
Hi,
I have a stituation where the viewport is smaller than the windows, causing the
region outside the viewport to not be cleared properly. Does anyone know how to
clear the background window?
I tied several things including the following:
Code:
void ClearClientBackground( HWND hWnd )
{
I guess I could have been a little clearer in my description. So here's an
example that should help:
1) I have a osg::camera of size 640x480. So the window's client area is
640x480, or GraphicsContext::Trats::(x,y,width,height) = (0, 0, 640, 480).
2) The GL viewport of this camera is 320x240,
Thanks for all your replies. That confirms what I imagined.
I forgot to mention that in addition of requiring multiple cameras, we also
need multiple windows. Each window presents a different view of the scene.
Because of the complexity of the scene I wanted to gain a little performance by
I have a newby question to ask... so please be gentil
Let's say I have 2 cameras running in different draw threads. Each camera is
also associated with a different graphic context. If I load a 2D texture then
each context will have its own opengl texture object (created in the
Hi,
I noticed that the alpha component in the BMP file (when RGBA or BGRA) is not
being preserved in the image. I would like to proposed the following small
change to correct this problem. Instead of ignoring the alpha component, I
first check if the bytesPerPixel == 4 bytes. If so, then I
Here is my system configuration:
Laptop XPS M1730
Windows 7
4 GB RAM
NVidia 8800M GTX, 512 MB
NVidia Driver v257.71
I hope this help. I've spent all weekend trying to deciphere what's going on.
Still don't know why it's not working.
--
Read this topic online here:
Sorry for the confusion - I ment to say the SharedStateManager.
I should have mentioned (in my original post) that I already merged the
SharedStateManager from the trunk into v2.8.2. Fortunately, the changes were
easy to merge but did not resolve the problem.
I'm trying to figure out a way to
Yes. Everything works when I don't use the compile context (i.e the scene is
rendered correctly), except that I still detect a problem with the number of
texture objects when using the database pager.
--
Read this topic online here:
Hi,
I think there are a few problems with compiling GL objects with the
DatabasePager (in OSG 2.8.2). The following topic sums it up pretty well:
http://forum.openscenegraph.org/viewtopic.php?t=1032start=0postdays=0postorder=aschighlight=compilecontext
So far, I was able to compile all the
I noticed a huge discrepency between loading an FLT with and without the
DatabasePager.
Before loading any database model, the number of TextureObject created is 16.
When I load a model via readNode(), the number of TextureObjects created
increases to 112 (for 96 textures objects for the
I noticed a huge discrepency between loading an FLT with and without the
DatabasePager.
Before loading any database model, the number of TextureObject created is 16.
When I load a model via readNode(), the number of TextureObjects created
increases to 112 (for 96 textures objects for the
I've been looking activating the SharedStateManager within the DatabasePager
and I was curious to know why the share() function is called from within the
main thread (i.e in DatabasePager::addLoadedDataToSceneGraph()) as opposed to
the DatabasePager thread (like the GL compile post-process).
I
After thinking about it, shouldn't the share() function be called before the
compile GL objects process which is invoked from the DatabasePager thread.
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=31655#31655
Hi,
Is there a plugin (or mechanism) that would allow me to load SE Core DVED
terrain databases which contains many pageable tiles but there's no master
file (see https://www.se-core.org/Default.aspx?tabid=175).
Here is an example of an SE Core DVED terrain database directory structure:
While analyzing the TXNode class, I noticed that it makes a suttle assumption
that the coordinate system is Z up. Almost everyhere else in OSG it make no
such assumption (or it provides a CoordinateFrame class to convert from one
coordinate system to another such as from Z up to Y up).
Here is
I am aware of the ReadFileCallback - but I guess the problem is how to
distinguish between my paging terrain files from other database files loaded
during runtime. All my files are FLTs (with the exception of a few TXP files
for the terrain). I can't use the file extension as a filter since
Hi,
Is there a mechanism that would allow me to get notified when a subgraph gets
loaded by the DatabasePager. I essentially need to process the subgraph before
it gets added to scene.
I guess this would be valid for TXPNode and PagedLOD nodes as well?
Thanks,
--
Read this
Hi,
The ComputeBoundVisitor does not seem to support TXPNode nodes. Is there
another way of getting the boundingbox (not the boundingsphere) of a TXP
database?
I know that the TXPNode contains the extents of the terrain (which is what I
want) but the interface is not exposed to the osg lib -
I'm having the same problem on my laptop (single screen). When I launch the
osgViewer application, the screen gets rendered for 1 frame and then the
entire window becomes black. If I use ALT-TAB to cycle back and forth, then the
scene gets rendered correctly.
As you mentioned, it's very
I'm hoping that someone can help me with a weird problem I'm having with an
openflight terrain I have.
It seems that when I mesh my terrain (with Multigen Creator 2.5.1) I can't see
it anyore. Note that the problem only occurs when I disable lighting (i.e
GL_LIGHTING). However, when I
As promised, I modified the osgprerender example to include a single instance
of the precipitation effect and I was able to reproduce the problem.
Guy
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=25686#25686
Attachments:
I will modify the precipitation example to include a render-target
implementation and see if I can reproduce the problem. If I am successful, I
will post the example on the forum so that someone can review the
implementation. Like you said Rob, I may have a bad setup and I just don't know
it.
Is your viewing camera also using the local coordinate system (i.e
geocentric)?
Guy
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=25621#25621
___
osg-users mailing list
If I recall correctly, even the osgParticle supports LODs, but I think they are
called details instead (or something rather). However, osg does not
implicitely use it - you need to manage the details yourself.
Guy
--
Read this topic online here:
Hi,
I am experiencing VERY bad performance (i.e 1 FPS) when I enable precipitation
on my Dell laptop (XPS 1710) that uses a GeForce 7950 graphic card (dell driver
version 179.xx).
But here's the kicker... the performance is fine when running on any other
desktop systems that has a more
Hi,
Many developers at my locations are experiencing odd frame rate readings (from
the StatsHandler), and we would like to know if others are having the same
issue.
We are running in a multiple window/camera enviroment. When I minimize a
windows, then the frame rate drops! I would expect
Hi,
I'm little confused about when it's appropriate or even necessary to call
setDataVariance(DYNAMIC) - assuming we are not running in SingleThreaded mode.
For example:
1) If I have a switch node and want to change its state in the update or event
sequence (i.e during runtime), does the
Ok. I have a little more information for you guys.
Just to recap the situation: When I run using Bilinear, I get 30 Hz. When I run
using Trilinear, I get 60Hz.
During my investigation, I noticed that their are actually 2 bilinear modes:
LINEAR and LINEAR_MIPMAP_NEAREST. LINEAR does not
Thanks for your quick response. This forum is fantastic.
Although I haven't tried it yet, I think that Render::setCompileOnNextDraw() is
exactly what I am looking for.
Paul, I can't seem to find the osgUtil::IncrementalCompileOperator class you
mentioned in OSG v2.8.2. Is that a new class in
I've bee experiencing momentary frame drops when I rotate my camera in my
scene. After look at the osg code, I noticed that many of the opengl objects
are created during runtime and not when the model is loaded. For example,
opengl texture objects are only created when the Texture2d::apply() is
Hi,
I was wondering was experiencing the same odd behavior I am. I created a
visitor that essentially iterates through all the textures of a terrain scene
graph to change their texture filtering from bilinear to trilinear.
When I replace the filtering from bilinear to trilinear, I get a boost
This might sound like an odd question, but is there a way to disable rendering
(no culling and no draw) while still allowing intersections. The reason why I
am asking is because I have a distributed system - i.e. some applications only
performs intersections with the scene without rendering
Paul. Thanks for the advice.
What I ended-up doing is overriding the void ViewerBase::frame(double
simulationTime) function. Depending on the state of a rendering flag, I either
call or don't call the renderingTraversals() function.
Hopefully, that won't cause any other issues.
Cheers,
Guy
Unfortunately, my graphic crad does not support NPOT texture which is why I
can't use osg::TextureRectangle.
I also tried ensuring that my Texture2D is always a POT when I change the
windows size (see code below).
Code:
int32 nWitdh = viewport-width();
int32 nHeight =
Ahhh.. I was not aware of that small detail. Thanks for the info - I will try
it immediately.
Guy
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=20898#20898
___
osg-users mailing list
Dude you the greatest. Using osg::TextureRectangle worked!
Thanks for your great help.
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=20926#20926
___
osg-users mailing list
Hi,
I was having problems with a resizable window that required a render target
implementation (like FRAME_BUFFER). The problem only occurred on hardware that
did not support NPOT textures (what can I say - I have an old graphic card!).
The problem was isolated in Texture2D::copyTexImage2D()
Forgot to mentioned that I was using OSG 2.8.2.
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=20834#20834
___
osg-users mailing list
osg-users@lists.openscenegraph.org
Hi,
Is there a way to disable culling for a geode. Essentially, I want the geode
render every frame - usefull for skyboxes and skydomes. I tried calling
setCullingActive(false) but that does not work.
The only way I was able to make it work, so far, is by setting the
Hi,
I think I found a wrong display label when when the camera scene stats is
enabled. There is a label called Matrices. Shouldn't it be called Materials
instead?
Guy
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=1#1
Hi,
We are observing a few problems when displaying the statistics on the screen.
Essencially, when we enable stats the State::getInitialViewMatrix() return
identity matrix. If the stats are NOT enabled, then it return the view matrix
of my slave camera. Is this the correct behavior? If so,
Jan, I looked at GLEW after I wrote my original post. I was about to download
and use when I figured out an easier way. I essentially followed robert's
advise (thanks rob!).
NOTE: In my case, I would also need to recompile GLEW to use multiple
threads/context (which also require installing
NPOT = non-power-of-two
POT = power-of-two
Since my card does not support NPOT textures, can I create a POT texture but
command the renderTargetImplemtation to only render to a specific region of the
texture.
For example, if my texture if 512x512 (i.e POT texture) and my viewport is
320x240
I was able to use a POT texture after all with the renderTargetImplementation
of camera (as I mentioned in my last post).
All is well with the world once again.
Thanks everyone.
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=17307#17307
Hi,
I am having a weird problem which I hope you can help. I am running on Win2000
with a relatively old nvidia card.
1) When I run then osgtexturerectangle example and I supply a non-power-of-two
image as a command line argument, the example still runs as expected. So that
tell me that my
Hi,
I have several cameras that need to be configured slightly differently
depending on the opengl extensions supported by the card. For example, I have a
card that does not support p-buffers in which case I need to devise a different
way to perform RTT. Same goes for vertex and fragment
You're the man!!!
Thanks.
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=14697#14697
___
osg-users mailing list
osg-users@lists.openscenegraph.org
Hi,
I've implemented the osg::GraphicsContext::ResizedCallback to get notified when
the user resizes the window via the mouse. That work perfectly.
However, I also need to change the window size programatically (when the user
presses a button, for example). I tried calling
As you may have noticed from my scene graph layout, I have a switch node above
the RTT camera. This switch node allows me to enable/disable the RTT process.
By default, this switch node has the RTT process disabled. So, the RTT process
is not executed.
I also found that (by default) the update
Hi,
Please disregard my last 2 postings. I am such an idiot (*sometimes*). I just
noticed a big flaw in my scen graph layout that was causing the problem.
Sorry for the confusion.
Guy
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=14112#14112
Hi,
I was wondering is someone can help me with a specific problem I'm having with
render-to-texture. In essence, I have a pre-render camera (i.e m_pRttCamera)
that renders the entire scene to a rectangular texture (i.e m_pRttTexture).
This texture is the mapped to a simple quad that fills the
Hi,
One of our developers noticed that the Image::getColor(const Vec3) can
potentially crash if the UV coordinates are negative. Effectively, the
following function convert UV coordinates into S,T pixel coordinates and then
calls getColor(unsigned, unsigned, unsigned). Since the latter uses
Hi,
I was wondering if you can provide me with some guidance regarding RTT. BTW: I
read and understood the osgMultipleRenderTargets example very well (hopefully
8)). However, I was wondering if it would be possible to simply it a little for
my app.
My application has multiple slave cameras
Hi,
What's the best way to pass user data to all nodes and drawables. For example,
lets assumes I have 2 cameras (left right). Each camera points to the same
scene. When rendering the left camera, I want the the nodes and drawables to
use userdata1, and when rendering the right camera I want
Yep That worked. Thanks you guys.
Guy
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=13061#13061
___
osg-users mailing list
osg-users@lists.openscenegraph.org
Hi,
I integrated an overlays system into my code that essentially supports multiple
overlays, each having multiple hud items. Each overlay uses an osg::Camera at
its root. All this works very well, with the exception of the following issue:
when I remove the overlay (i.e the root camera) from
Hi,
Here is the call stack for the crash and the code:
[img]CallStack.bmp[/img]
Code:
void RenderStage::draw(osg::RenderInfo renderInfo,RenderLeaf* previous)
{
if (_stageDrawnThisFrame) return;
// push the stages camera so that drawing code can query it
if (_camera)
Hi,
ok. I registered a osgDB::Registry::ReadFileCallback and I overidden the
readNode() function. As Robert suggested, I added a node visitor to scan for
all statesets (in nodes and drawables) for textures.
Code:
void CGeTrilinearVisitor::apply( osg::Node rNode )
{
osg::StateSet*
Oooops. I forgot the set the visitor traversal mode to TRAVERSE_ALL_CHILDREN.
That was tupid of me!!?!?
I'll let you know if it works.
Guy
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=12986#12986
Hi,
I have several openflight terrain database that use bilinear mipmaping (which
look ugly). Is there a easy way to force all terrains to use trilinear
mipmapping upon loading the terrain? Obviously I can change the openflight
plugin to check if a trilinear flag is set, but I would like to
Hi,
I have 2 windows, each with its own camera/viewport. I need to display a
different hud on each camera. What's the best solution for this. I looked at
the hud example, but there is only one hud and it is added to the scene, and
therefore gets displayed on each camera.
Any suggestion would
Hi,
Anyone has a good example on how to render decals on a terrain (for scorch
mark, tracks, etc). For example, I want to put a scorch mark on a wall of a
building, but I don't want it to potrude over the edges of the wall; it should
be clipped to the edge of the wall.
Any idea or (better
Hi Gordan,
Multitexturing would apply if the scorch marks are permanent. In my situation I
want to fire several gun rounds and only display a scorch mark where the wall
was hit. Then they would fade out over time.
I think this is usually done by adding a extra quad slightly in front of the
Hi again,
I tried J-S 's suggestion about attaching the hud camera to a slave camera, but
that did not work (i.e the scene is rendered, but there is not hud).
viewer.getSlave( 0 )._camera-addChild( hudCamera );
Note that adding the hud camera to the scene (like in the hud example) did work
Hi,
Is there a way to get access to the CullingSet::pixelSize() function while in a
drawable callback? I can't find a way to get access to it via the RenderInfo
that is that passed as a parameter to the drawable callback.
Thank you!
Cheers,
Guy
--
Read this topic online here:
Hi,
I want to change the perspective of a slave camera but I can't seem to make it
work (so far). I tried changing the master camera but that affects ALL slave
cameras which is not what I want to do. Is there a way to change the
perspective view of a single slave camera? Do I need to switch
I should have looked at the osg::view code before posting the message.
Now that I set the camera as ABSOLTE_REF, the camera manipulators don't work
(as I feared). I tried deciphering the code, but I can't locate were the mouse
movements (for example in trackballmanipulator) is being applied to
Hi,
I was wondering what’s the best way to simulate thermal signatures for targets
and the terrain (i.e heat vision).
One simple and easy way is to load 2 models - one for day view and another for
thermal view. However, this poses a big problem for the terrain. I really don't
want to load
89 matches
Mail list logo