[osg-users] Thoughts on Vulkan
Hey all, Now that the Khronos group has announced Vulkan/SPIR-V, are there any plans to integrate it into OSG as an alternative back end? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] osg+osm
On Tue, Sep 17, 2013 at 11:09 AM, Айдар Ильдарович e...@inbox.ru wrote: Hello! I use OpenSceneGraph 3.0.1 in my research. Please, help me, how to use and to load files of OpenStreetMap (extension o5m, pbf, osm) in osgviewer. You can't do this directly. The files have to be opened and parsed using another library and then you need to decide how to set up your scene graph based on your needs. How would OpenSceneGraph know what you want to do with an OpenStreetMap node or way and other such elements? ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Trouble loading OSG plugins
I wasn't able to build the static version of OpenSceneGraph using the provided guide, so I tried building a shared version directly using CMake and the Android NDK toolchain which seems to have worked. The problem I have right now is with osgPlugins-x.y.z being prefixed to the search path by OSG. As an experiment, I put some osgPlugins in the sdcard folder (external storage) /sdcardpath/osgPlugins-x.y.z/ and if I explicitly specify this path (/sdcardpath) to the osgDB::Registry, it loads the plugins just fine. This isn't okay for general app deployment though. I was able to pack shared object (plugin) files without a 'lib' prefix in the APK/lib directory. I'm unable to create any more directories in APK/lib. Unfortunately I can't load these because (I think) OSG is prefixing the path with osgPlugins-x.y.z.. I want to disable this behavior so I can package the app properly. On Tue, Sep 10, 2013 at 3:32 PM, Jorge Izquierdo Ciges jori...@gmail.comwrote: Android has a serie of troubles: 1rst every library has to begin with lib if has to be included in the APK 2nd every library has to be loaded at the first try or will never be loaded until you close your application 3rd If your library needs a dependence he will not load it even if its on the same directory So... 1rst - OSG searches plugins without lib at the begin and they are compiled with lib if my memory is well enough. 2nd - this is confirmed for versions pre 4.2 3rd - this is confirmed for versions pre 4.2 At least on a Nexus 4 the 2/3 don't happen anymore with the last version. I have OSG with shared libraries and plugins in Android but it is far from submission because we changed a lot of things and we are checking everything thrice, so don't expect it soon. But... you can make it work, my phone is proof. 2013/9/10 Preet prismatic.proj...@gmail.com m to create an osgPlugins directory where OSG ends up being deployed in the application package so I just want OSG to load plugins from the directory where the rest of the shared objects live on Android: /data/app-lib/org.some.app/--- where all the *.so libraries live, including osgdb_xx.so ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Possible bug with OpenSceneGraph for OpenGL ES 2
Hi, I've noticed on two separate platforms (PowerVR SGX 540 Adreno 320) that OSG will put out a warning when textures are used: Warning: detected OpenGL error 'invalid enumerant' after applying GLMode 0xde1 GLMode ox0de1 is GL_TEXTURE_2D. In include/osg/State, when the applyModeOnTexUnit function is invoked, glEnabled(GL_TEXTURE_2D) and glDisable(GL_TEXTURE_2D) may be called. On some OpenGL ES 2 platforms, this will return GL_INVALID_ENUMERANT because GL_TEXTURE_2D is always enabled for OpenGL ES 2. A possible fix is to wrap the glEnable and glDisable calls with an #ifndef OSG_GLES2_AVAILABLE, but there might be a better place to address this. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Trouble loading OSG plugins
I'm having issues trying to get OSG to load plugins on Android for a shared library build... though OSG seems to be working fine on Android as a shared lib without plugins which is great! I can't seem to create an osgPlugins directory where OSG ends up being deployed in the application package so I just want OSG to load plugins from the directory where the rest of the shared objects live on Android: /data/app-lib/org.some.app/--- where all the *.so libraries live, including osgdb_xx.so So I do this: osgDB::Registry::instance()-setLibraryFilePathList(/data/app-lib/org.some.app); But it seems like its prefixing 'osgPlugins-3.3.0' to its search path. This is the output I get from OSG: Warning: dynamic library 'osgPlugins-3.3.0/osgdb_png.so' does not exist (or isn't readable): Cannot load library: load_library(linker.cpp:745): library osgPlugins-3.3.0/osgdb_png.so not found DynamicLibrary::failed loading osgPlugins-3.3.0/osgdb_png.so Is there something else I need to do to specify the path where it should search? Is there a way to avoid creating an osgPlugins folder altogether as a build option and just have it look for the plugins without prefixing anything to their path? I ran across this problem when using OSG for another platform as well and worked around it using a few tricks but Android is much less tolerant in this regard. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Trouble loading OSG plugins
On Tue, Sep 10, 2013 at 3:32 PM, Jorge Izquierdo Ciges jori...@gmail.com wrote: So... 1rst - OSG searches plugins without lib at the begin and they are compiled with lib if my memory is well enough. You were right. I was able to build an APK with the plugin files (named osgdb_xx.so) but after deploying to the device itself all the plugin shared objects were no longer in /data/app-lib/org.some.app/ I have OSG with shared libraries and plugins in Android but it is far from submission because we changed a lot of things and we are checking everything thrice, so don't expect it soon. Do you have a public repo where you are testing this out? I suspect you're setting up a build similar to the static one where CMakeLists will do most of the magic. My build notes are here (still have to add some stuff, but its mostly complete: https://github.com/preet/scratch/wiki/OpenSceneGraph-on-Android) On Tue, Sep 10, 2013 at 6:10 PM, Raphael Grasset raphael.gras...@gmail.com wrote: Hi Preet, For the path prefix, not sure if it works, but look at src/osgDB/Registry.cpp, and the function createLibraryNameForExtension, you will see a: std::string prepend = std::string(osgPlugins-)+std::string(osgGetVersion())+std::string(/); so if you create an empty string it may address your path issue. This set me on the right path to try an alternative once I realized I couldn't pack the default osgdb_xx.so plugins in with the APK libs. I used the osgPlugins CMake and set the TARGET_DEFAULT_PREFIX to libosgdb_ (and modified createLibraryNameForExtension accordingly). Prefixing them with lib meant they actually showed up on the device when I deployed the APK... and I just ended up loading the plugins directly with the linker (ie i just explicitly linked against them when I built the app). Doing all that takes a lot of effort though... it's cleaner (from OSG's stand point) to save the plugins in internal storage and access them that way. The cost with that method is the plugins have to be packed and then extracted/installed during the application's first run. Yet, it will be preferable that you build everything in static mode: what was the reasons your static built didn't work? I was able to configure the build using CMake successfully, but calling make in the build directory fails with a bunch of 'undefined reference to' errors: I explained it briefly here: http://lists.openscenegraph.org/pipermail/osg-users-openscenegraph.org/2013-September/064587.html Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Android build fails with osg from git
Hiya, I think some of the recent code wrt osgPresentation is causing the Android build to fail. Here's my cmake config call: cmake \ -DOSG_BUILD_PLATFORM_ANDROID=ON \ -DANDROID_NDK=/home/preet/Dev/env/sys/android-ndk-r8e \ -DCMAKE_INSTALL_PREFIX=/home/preet/Dev/env/android/osg-git-debug \ -DOSG_GL1_AVAILABLE=OFF \ -DOSG_GL2_AVAILABLE=OFF \ -DOSG_GL3_AVAILABLE=OFF \ -DOSG_GLES1_AVAILABLE=OFF \ -DOSG_GLES2_AVAILABLE=ON \ -DOSG_GL_LIBRARY_STATIC=OFF \ -DOSG_GL_DISPLAYLISTS_AVAILABLE=OFF \ -DOSG_GL_MATRICES_AVAILABLE=OFF \ -DOSG_GL_VERTEX_FUNCS_AVAILABLE=OFF \ -DOSG_GL_VERTEX_ARRAY_FUNCS_AVAILABLE=OFF \ -DOSG_GL_FIXED_FUNCTION_AVAILABLE=OFF \ -DANDROID_ABI=armeabi-v7a \ -DANDROID_PLATFORM=8 \ -DANDROID_STL=gnustl_shared \ -DANDROID_DEBUG=ON \ ../osg-src Here's the resulting error: make[3]: *** No rule to make target `/home/preet/Downloads/build/osg/src/osgPresentation/Model.cpp', needed by `obj/local/armeabi-v7a/objs-debug/osgPresentation/Model.o'. Stop. Also... I get a bunch of warnings during compile (unrelated to the presentation error -- this is with older commits as well). The following line is output 100's of times. Android NDK: WARNING: Invalid LOCAL_CPP_EXTENSION values: cpp This doesn't seem normal... Should I be concerned about it? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Android build fails with osg from git
Hmm. Even with the osgPresentation issue fixed, I'm unable to build the libs I get hundreds of errors like: /home/preet/Dev/env/sys/android-ndk-r8e/toolchains/arm-linux-androideabi-4.6/prebuilt/linux-x86_64/bin/../lib/gcc/arm-linux-androideabi/4.6/../../../../arm-linux-androideabi/bin/ld: ./obj/local/armeabi-v7a/objs-debug/osgDB/SharedStateManager.o: in function construction vtable for osg::NodeVisitor-in-osgDB::SharedStateManager:SharedStateManager.cpp(. data.rel.ro+0x158): error: undefined reference to 'virtual thunk to osg::NodeVisitor::~NodeVisitor()' /home/preet/Dev/env/sys/android-ndk-r8e/toolchains/arm-linux-androideabi-4.6/prebuilt/linux-x86_64/bin/../lib/gcc/arm-linux-androideabi/4.6/../../../../arm-linux-androideabi/bin/ld: ./obj/local/armeabi-v7a/objs-debug/osgDB/SharedStateManager.o: in function construction vtable for osg::NodeVisitor-in-osgDB::SharedStateManager:SharedStateManager.cpp(. data.rel.ro+0x15c): error: undefined reference to 'virtual thunk to osg::NodeVisitor::~NodeVisitor()' /home/preet/Dev/env/sys/android-ndk-r8e/toolchains/arm-linux-androideabi-4.6/prebuilt/linux-x86_64/bin/../lib/gcc/arm-linux-androideabi/4.6/../../../../arm-linux-androideabi/bin/ld: ./obj/local/armeabi-v7a/objs-debug/osgDB/Version.o: in function osgDBGetVersion:/home/preet/Downloads/build/osg-3.2.0/src/osgDB/Version.cpp:20: error: undefined reference to 'osgGetVersion' collect2: ld returned 1 exit status make[3]: *** [obj/local/armeabi-v7a/libosgDB.so] Error 1 make[2]: *** [Android-OpenSceneGraph] Error 2 make[1]: *** [CMakeFiles/ndk.dir/all] Error 2 I thought this might be related to: http://lists.openscenegraph.org/pipermail/osg-users-openscenegraph.org/2013-July/063911.html But I'm having issues building the libs -- haven't gotten to the part where I try out an example yet... in any case, I tried adding a LOCAL_LDFLAGS := /home/preet/Dev/env/sys/android-ndk-r8e/sources/cxx-stl/gnu-libstdc++/4.6/libs/armeabi-v7a/libgnustl_static.a (or shared ... I tried to build both but they both fail at the same point) to the top level Android.mk file in the cmake build dir but that doesn't help. Any advice would be appreciated. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Solving forum.openscenegraph.org problems
I second the StackExchange suggestion if the mailing list and forums are going to be replaced by one system. StackExchange sites are much easier to search through than relying on Google to look up things in the mailing list or forums. I think it'd cut down on repeat questions a lot (how many users here have been told by a more experienced OSG user to go back and search through the lists because the question being asked has been discussed before? I know I have!) It's also much easier to integrate code snippets and images to provide better questions and answers. There are plenty of OSG questions on StackOverflow so one can get an idea of how that kind of site would look. Regarding user registration - I never really understood the point of having your full name being a mandatory thing for the forum sign up. Is it a courtesy thing? Is it to prevent spam? Preet On Wed, Jul 31, 2013 at 1:36 PM, Alexandre Vaillancourt alexandre.vaillancourt.l...@gmail.com wrote: Another idea could be to use a system like StackExchange. This probably does not have an option to offer support by email, though. And if there is an aim for 'growing the community', I'm not sure that it's the right path. However the 'QA' process is fantastic. -- Alexandre Vaillancourt ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] OpenSceneGraph transformation precision
It looks like a common precision issue in OpenGL -- you can't directly pass planetary scale values since you're limited to floats. To get around this, parent your Geodes with a Transform node as you did in your second example. See this mailing list thread for more info: http://lists.openscenegraph.org/pipermail/osg-users-openscenegraph.org/2012-February/056373.html ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Text with transparent background
Is blending enabled in the scene graph? ie rootNode-getOrCreateStateSet()-setMode(GL_BLEND,osg::StateAttribute::ON); On Fri, Jul 19, 2013 at 3:05 AM, Xiangfei Cui sophist...@gmail.com wrote: Hi, There are two osg::Drawables. One is a filled red polygon, the other is a white Text on the polygon. I wish as to the Text's background is transparent so I can see red background, but the result is wrong. as to the figure: Thank you! Cheers, Xiangfei -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=55288#55288 Attachments: http://forum.openscenegraph.org//files/qq20130719120253_166.bmp ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] load many and big bmp image into textures
1. If you use osg::ref_ptr for your Texture2D objects, they should be cleaned up when you remove any references to them (ie just remove the geode(s) the texture is attached to or any other references to the texture) 2. bad_alloc usually means you're running out of memory. On Fri, Jul 26, 2013 at 2:48 AM, Nat Doublet doub...@cgxaero.com wrote: Hi, I cut an image in 16*16 bitmap files (12M each). I want to put one by one in an osg::texture2d, after use, I want to unload it just after (to avoid trouble with memory). My code is : [code] std::pairint, int XY = std::make_pair(ix, iy); osg::StateSet* p_stateSet = geode-getOrCreateStateSet(); osg::Texture2D* texture = new osg::Texture2D(); texture-setWrap(osg::Texture2D::WRAP_S,osg::Texture2D::REPEAT); texture-setWrap(osg::Texture2D::WRAP_T,osg::Texture2D::REPEAT); std::string fileImg = p_mapXYImgFileName.at(XY); texture-setImage(osgDB::readImageFile(fileImg, p_option)); p_stateSet-setTextureAttributeAndModes(0,texture,osg::StateAttribute::ON); geode-setStateSet(p_stateSet); p_stateSet-releaseGLObjects(); [/code] 2 questions : 1. because I cannot call texure2d destructor (same for stateset), am I right to try to free memory with releaseGLObjects? (my computer don't want to manage with 2Gb of data loaded) 2.After iterations on X, Y (not always the same number), I go a bad allocation error on the texture-setimage()... [code] Exception Microsoft C++ : std::bad_alloc à l'emplacement mémoire 0x0019c54c. bad allocation [/code] Have you got an idea? some wayround? Thank you! Cheers, Nat -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=55454#55454 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How do I build OpenSceneGraph as a shared lib without sonames?
Hi, I'm building for a platform where I can't use sym links when packaging an application and soname versioning complicates the deploy process. Is there a way to tell osg/cmake that I don't want to build versioned libraries with sonames? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How do I build OpenSceneGraph as a shared lib without sonames?
Easy fix; In the main CMakeLists.txt simply change # We want to build SONAMES shared librariess SET(OPENSCENEGRAPH_SONAMES TRUE) -- change to false SET(OPENTHREADS_SONAMES TRUE) --- change to false Would be cool to have a cmake option that controls this though ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Using osg::Camera without a view
Hiya, This is a bit of a strange question. I want to use an osg::Camera purely as a shell object to use some of its handy methods (ie. all the setters and getters for the various view/model/projection matrices). Can I do this without setting up a viewer or invoking any OpenGL code (assume inactive context)? If I do something like: // note there's no osgViewer here osg::Camera cam; cam.setViewport(...); cam.setViewMatrixAsLookAt(...); cam.setProjectionMatrixAsPerspective(...); osg::Matrixd viewMatrix = cam.getViewMatrix(); Will this 'just work'? Or is there internal stuff in osg::Camera that needs the object to belong to a viewer / have a proper view/context/etc setup? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] qmlosg (QtQuick + osg integration)
Hi Riccardo, I can't speak to the specific slow down you're encountering with context switches. I tried to solve a similar problem (osg in qt quick) but I don't use context switching -- I could never figure out how to get that to work correctly so I'll have to go through what you did carelfully :) Have you checked out the following docs: http://doc-snapshot.qt-project.org/qt5-dev/qtquick/qtquick-visualcanvas-scenegraph.html http://doc-snapshot.qt-project.org/qt5-dev/qtquick/quick-scenegraph-textureinsgnode.html http://doc-snapshot.qt-project.org/qt5-dev/qtquick/quick-scenegraph-textureinthread.html And here's my attempt [copy of the method used in the docs but with an osg embedded viewer and *no* context switching -- I just reset some specific state] https://github.com/preet/scratch/tree/master/qt5/qquickfboviewport ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How can I check which objects are frustum culled?
Is it possible to get a list of nodes that *do* pass the cull test then (outside of the obvious 'look at whats rendered')? Preet On Thu, Mar 21, 2013 at 5:02 AM, Robert Osfield robert.osfi...@gmail.comwrote: Hi Preet, The OSG cull traversal doesn't track the objects that are culled, only the ones that pass the cull test. Robert. On 21 March 2013 02:36, Preet prismatic.proj...@gmail.com wrote: Hiya, In my scene I'd like to check which nodes are being culled before rendering due to frustum culling. The culling mode I use my main (only) camera is osg::CullSettings::VIEW_FRUSTUM_SIDES_CULLING. Is it possible to do this? Can I get a list of objects that are culled, or something similar? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How can I check which objects are frustum culled?
Hi Robert,
Or... just use the Stats feature that
does just this. Have a look at the src/osgViewer/StatsHandler.cpp to
see how the Stats collection can be enabled and used.
Thanks for the quick reply, this worked great. I'm just posting a small
snippet in case anyone else searches the list for something similar since
it took me a little while to narrow down what I was looking for.
// enable stat collection for scene
myViewer-getStats()-collectStats(scene,true);
// enable stat collection for camera (do this after rendering/everything is
setup)
myViewer-getCamera()-getStats()-collectStats(scene,true);
// dump some stats in between frame calls
osg::Stats * sceneStats = m_osg_viewer-getCamera()-getStats();
osg::Stats::AttributeMap attMap =
sceneStats-getAttributeMap(sceneStats-getLatestFrameNumber());
size_t numObjects = 0;
size_t numTris = 0;
size_t numVx = 0;
osg::Stats::AttributeMap::iterator amIt;
for(amIt = attMap.begin();
amIt != attMap.end(); ++amIt)
{
std::cout amIt-first , amIt-second std::endl;
}
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How can I check which objects are frustum culled?
Hiya, In my scene I'd like to check which nodes are being culled before rendering due to frustum culling. The culling mode I use my main (only) camera is osg::CullSettings::VIEW_FRUSTUM_SIDES_CULLING. Is it possible to do this? Can I get a list of objects that are culled, or something similar? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] problem of rendering overlaped models.
When you say weird what do you mean? If you have polygons that are positioned/aligned very closely to one another you might be seeing z-fighting (http://www.zeuscmd.com/tutorials/opengl/15-PolygonOffset.php). In that case you can tell OpenSceneGraph to use polygon offset on one of the models. osg::ref_ptrosg::PolygonOffset polyOffset = new osg::PolygonOffset; polyOffset-setFactor(1.0f); polyOffset-setUnits(1.0f); ss = someNode-getOrCreateStateSet(); ss-setAttributeAndModes(polyOffset); On Sat, Feb 16, 2013 at 1:57 AM, wh_xiexing wh_xiex...@sina.com wrote: ** hi friends: i have 2 models to render, some part of which are overlaped . so the result is some kind of weird. how can i resolve this probem? do i need to split the model and align them? or set different render details for the 2 models? -- Shawl ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Mipmapping a dynamic texture
Hi I have a single osg::Image that I use as an atlas. I update the atlas image every so often using osg::Image::copySubImage. I'd like to use hardware mipmapping (ie. glGenerateMipMap) for the corresponding texture. When I modify the osg::Image, will OSG update the mipmaps as well? Or do I need to explicitly tell OSG to do this (if so, how?) Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Creating a billboard shader for OpenGL ES 2
Hiya, This is somewhat related to my last thread (maybe I should have just explained the situation from the beginning). I'm trying to create a shader that does billboarding for hundreds of objects at once (where using AutoTransform is too slow). I merge all these objects together to reduce the number of draw calls, and pass along specific vertex attribute data (specifically model-space relative vertex data when I want the billboard center to be the model's center) and uniforms (camera vectors) to do the required billboard transform in the shader. This kind of works, but the result is 'jittery' compared to AutoTransform. I don't know whether its the way I specify the uniform update callback or if its related to reduced shader precision (the coordinates I used are all close to the origin though, so I hope that's not the case) or something else entirely. I created a minimal example that reproduces the issue [1]... (just copy-past-compile). The example shows a square billboarded with AutoTransform working perfectly, and my not so great example directly behind it. [1] https://gist.github.com/4582261 * Does anyone know why my example billboard implementation causes that jitter/delay? * Or, has anyone created a better shader that can do billboarding for batched geometry (and if so, could they maybe share how)? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Why am I getting this bleed through effect when moving my camera around?
I have two geometries close together -- one is an osgText object and the other is just a simple set of planar triangles laying flat behind the text. The two are separated by a small distance. Both objects are parented with an AutoTransform. When I move my camera around in a certain way, the object behind the text bleeds through it. When the camera is not being moved, both objects appear fine. I made a small video because I had trouble capturing the issue with a screenshot. In the video, the text is purple. The object in question is the black rectangle with the purple border behind the text. http://www.youtube.com/watch?v=htj168dJHGw When I move the camera relatively fast, you can see the black rectangle's triangles bleed through the text momentarily. I tried messing around with the placement of the objects; I even tried to increase the distance between the text and the geometry behind it a substantial amount but this didn't seem to have an effect. Both objects are in the same render bin. I've never really seen something like this happen before where the effect is only present *while* the scene is moving around. If I move the camera around slow enough or keep it still, the effect isn't visible. Does anyone have any ideas as to what the issue is here? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Why am I getting this bleed through effect when moving my camera around?
Nevermind -- it turns out my idea of a 'substantial' amount of distance wasn't enough. On Sun, Jan 20, 2013 at 12:42 AM, Preet prismatic.proj...@gmail.com wrote: I have two geometries close together -- one is an osgText object and the other is just a simple set of planar triangles laying flat behind the text. The two are separated by a small distance. Both objects are parented with an AutoTransform. When I move my camera around in a certain way, the object behind the text bleeds through it. When the camera is not being moved, both objects appear fine. I made a small video because I had trouble capturing the issue with a screenshot. In the video, the text is purple. The object in question is the black rectangle with the purple border behind the text. http://www.youtube.com/watch?v=htj168dJHGw When I move the camera relatively fast, you can see the black rectangle's triangles bleed through the text momentarily. I tried messing around with the placement of the objects; I even tried to increase the distance between the text and the geometry behind it a substantial amount but this didn't seem to have an effect. Both objects are in the same render bin. I've never really seen something like this happen before where the effect is only present *while* the scene is moving around. If I move the camera around slow enough or keep it still, the effect isn't visible. Does anyone have any ideas as to what the issue is here? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Does OSG support ES2.0?
Hi Paul, On Tue, Jan 8, 2013 at 8:20 AM, Paul Heraty heratyp...@eircom.net wrote: Hi Preet, Do you have a copy of the shaders in that code that you can share please: shaders/NoShading_vert.glsl shaders/NoShading_frag.glsl They are in the same repo: https://github.com/preet/scratch/tree/master/openscenegraph/osg_text_geometry/shaders Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Does OSG support ES2.0?
To follow up on Jorge's reply:
Here's an example of how you can use shaders for osgText:
https://github.com/preet/scratch/blob/master/openscenegraph/osg_text_geometry/main_old.cpp
If you're still having issues, increase the OSG debugging output:
(osg::setDebugNotify(...)) and
osgViewer::getCamera()-getGraphicsContext()-getState()-
setCheckForGLErrors(osg::State::ONCE_PER_ATTRIBUTE)
to try and get a bit more info on the error. Something like APITrace (
https://github.com/apitrace/apitrace) might help too if you can get it
running on the target platform.
To get textures working properly on my platform I had to make a couple of
changes to the source. Unfortunately, for some stupid reason I accidentally
scrubbed the git history so I don't really have diffs of all the changes I
made. One specific thing I remember was having to set _glMaxTextureCoords =
1 in /src/osg/State.cpp to get textures working on my device.
Replace the if clause on line 900 with:
if ( osg::getGLVersionNumber() = 2.0 ||
osg::isGLExtensionSupported(_contextID,GL_ARB_vertex_shader) ||
OSG_GLES2_FEATURES)
{
glGetIntegerv(GL_MAX_COMBINED_TEXTURE_IMAGE_UNITS,_glMaxTextureUnits);
#ifdef OSG_GLES2_AVAILABLE
_glMaxTextureCoords = 1;
#else
glGetIntegerv(GL_MAX_TEXTURE_COORDS,_glMaxTextureCoords);
#endif
}
Preet
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How can I call a vector method on a generic array type?
Hi, I'm trying to call std::vector functions on generic array objects. So I'd like to be able to get an osg::Array and call something like size() or clear() on it, but I'm not sure how I can do this without casting it to one of the 'proper' array types. Is this possible? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Fitting my world into the viewport
Hi, On Sun, Dec 23, 2012 at 4:14 PM, Daniel Peraza danielper...@gmail.comwrote: Hi, I'm very new to OSG and OpenGL programming in general, but right now I'm working on my thesis and my application requires to load several models on the same scene. Once every model is done loading, I must adjust the viewer's settings so that an user is able to see the entire scene. I know this can be achieved by fitting the world's bounding sphere into the viewport, the problem is that I don't know OSG Classes well enough to do so. The default OpenSceneGraph viewer does this automatically for you. If you don't specify an initial camera, the viewer will calculate a view extents that tries to fit the entire scene. Unless your models are extremely far apart and beyond the default clipping planes, you should see all your models right away. My application is Qt-based and uses a Track Ball Manipulator. I wonder if I must use the manipulator methods, or adjust the camera's projection directly. OSG has its own set of manipulators, including a track ball. To integrate the two, check out the osgviewerQt example and osgQt libs. Unless there's something special about your scene, I don't see any need to adjust the camera's projection directly. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Render multiple osgText nodes with the same drawcall
It will probably be more efficient to create a scene graph from osg::Geometry/osg::Texture/osg::StateSet rather than using osgText. You could use the osgText::Font and associated Textures to help with the texturing side to the rendering. The osg::Geometry you create you could put as many quads in as you need. I have a couple of questions on using the texture directly rather than using osgText::Text... When grabbing the textures with all the glyphs in it (osgText::Font::getGlyphTextureList), can I safely assume that for one font resolution there will only be one texture for that font? That is will *all* characters which are requested with the same font be handled by one texture? How/when do individual glyphs get added to their collective texture? (I looked through the source code but got a little lost) -- in a small example I tried with just osg::Geometry/Texture as you suggested above, I looked up glyph texture coordinates for a bunch of characters without manually adding any glyphs and it 'just worked'. Do I never need to do anything to ensure the glyphs I need are in the texture or is this taken care of implicitly? Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] What can I do to improve rendering performance for a bunch of billboarded nodes?
I have a scene that contains 100s of small low-complexity flat geometries (triangles, squares, ngons, etc). Each of them is parented with an autotransform node so that they constantly face the camera. Color data is passed through vertex attributes and there's no texture data. Is there anything I can do to improve rendering performance in this scenario? This is on a mobile device (OpenGL ES 2) and I get a fair bit of slow down as the number of objects increases. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] osg plugin path naming issue?
Hiya, OSG couldn't find a few plugins on an embedded target where I was manually installing the libraries. I noticed that osgDB was searching for plugins named with a 'd' at the end (I *am* using debug libs) but my actual plugins don't have a 'd' appended at the end of their name. For example, osgDB was looking for osgdb_freetyped.so, when the plugin itself was built as osgdb_freetype.so Everything works as expected when I rename the files (append a 'd' to all plugins) before putting them on the device. I don't move their relative paths or anything (so they stay in osgPlugins-x.x.x). Is this a bug or did I do something weird when building OSG? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] OSG and Qt GLdouble typedef conflicts with OpenGL ES2
Hi Robert, Thanks for the quick reply. Just for anyone else who runs into this, another way to get aorund this without changing order of includes is just to add the osg defines before Qt is included in the application: #define GL_DOUBLE 0x140A #define GLdouble double It was a bit of a pain to change up order of includes everywhere in my application. Since both OSG and Qt have to check whether or not GLdouble is defined by the GL implementation, I think its safe to do this [and it seems to be working fine]. Regards, Preet On Wed, Dec 12, 2012 at 3:51 AM, Robert Osfield robert.osfi...@gmail.comwrote: Hi Preet, On 12 December 2012 05:59, Preet prismatic.proj...@gmail.com wrote: Hiya I'm using OSG with Qt to create an OpenGL ES 2 application. GL ES2 does not have GLdoubles. When building witHh against ES2, Qt typedefs GLdouble to GLfloat. OSG typedefs GLdouble to just 'double'. As a result I get an error when trying to compile my application. How do I get around this? It will be an order of include issue, so just be careful which order you include files, or keep the Qt GL code seperate from the OSG Qt code. Personally I'm surprised by the Qt downcasting GLdoube to GLfloat, this could break codes that rely on doubles. On the OSG side we generally use floats for all key data that gets passed to OpenGL, it tends to be just things like glu (which is now part of the OSG) that uses doubles. Robert. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Render multiple osgText nodes with the same drawcall
Hiya, I'm using OSG on a mobile target (OpenGL ES 2) and my scene runs really really slow. I'm trying to optimize it by reducing the number of draw calls. My scene has a bunch of labels that draw text along a path. Each label consists of several osgText nodes that are single characters in length, each with their own transform on top. So if a single label is OSG it will have three osgText nodes. I want to throw all of the label attributes (which I think is just vertex and tex coord arrays) into a single VBO, but I'm not really sure how I would do that (or even if I could). I'd appreciate any advice. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] OSG and Qt GLdouble typedef conflicts with OpenGL ES2
Hiya I'm using OSG with Qt to create an OpenGL ES 2 application. GL ES2 does not have GLdoubles. When building with against ES2, Qt typedefs GLdouble to GLfloat. OSG typedefs GLdouble to just 'double'. As a result I get an error when trying to compile my application. How do I get around this? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Rendering sets of objects in layers/passes
Hiya, I have multiple sets of depth sorted geometry that I'd like rendered in a specific order. So the depth sorted part itself only matters within a single set, if that makes any sense. For example, I might have three sets, each containing 100 osg::Geometry objects. I'd like each of the objects to render correctly with respect to the other objects in their own set, but I'd like all the objects in the first set to show up over all the objects in the second set and so on. I'm not sure how to do this with OpenSceneGraph. I tried giving each set of objects a separate DepthSorted render bin... but this doesn't do what I want. I'd appreciate any advice. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How can I get the viewer's camera to 'ignore' a node's geometry?
Hiya, A combination of: * setting the cull mode to myCam-setCullingMode(osg::CullSettings::VIEW_FRUSTUM_SIDES_CULLING) * manually specifying the near and far planes * letting osg compute the right bounding box for the earth geometry by itself seems to be working. Thanks to all for the help! Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How can I get the viewer's camera to 'ignore' a node's geometry?
Sergey's suggestion (invalidating the geometry's bounds with a custom compute bounds callback) sort of worked, but led into another issue. If I zoom out far enough, the Earth surface geometry disappears. Specifically, if the camera is pointing at the center of the Earth, the Earth geometry disappears when the camera eye is further than ~5.13E6m from the surface of the Earth. Changing the near and far planes don't have any effect here. However, if I disable culling I don't have this problem (but disabling culling kills the framerate): myCam-setCullingMode(osg::CullSettings::DO_NOT_COMPUTE_NEAR_FAR); So right now I'm: * Setting the near and far plane manually (myCam-setComputeNearFarMode is set to DO_NOT_COMPUTE) whenever the camera is updated to ensure my desired view volume is correct * Setting the Earth surface geometry compute bounds callback to disabled And that works fine up until I move the camera out far enough as I described previously. Something happens at that causes the geometry to disappear unless I disable culling altogether. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How can I get the viewer's camera to 'ignore' a node's geometry?
On Sat, Oct 6, 2012 at 1:02 AM, Preet prismatic.proj...@gmail.com wrote: Sergey's suggestion (invalidating the geometry's bounds with a custom compute bounds callback) sort of worked, but led into another issue. If I zoom out far enough, the Earth surface geometry disappears. Specifically, if the camera is pointing at the center of the Earth, the Earth geometry disappears when the camera eye is further than ~5.13E6m from the surface of the Earth. Changing the near and far planes don't have any effect here. However, if I disable culling I don't have this problem (but disabling culling kills the framerate): myCam-setCullingMode(osg::CullSettings::DO_NOT_COMPUTE_NEAR_FAR); Sorry, typo above, I meant to write: myCam-setCullingMode(osg::CullSettings::NO_CULLING); So right now I'm: * Setting the near and far plane manually (myCam-setComputeNearFarMode is set to DO_NOT_COMPUTE) whenever the camera is updated to ensure my desired view volume is correct * Setting the Earth surface geometry compute bounds callback to disabled And that works fine up until I move the camera out far enough as I described previously. Something happens at that causes the geometry to disappear unless I disable culling altogether. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How can I get the viewer's camera to 'ignore' a node's geometry?
Hey, By default some of the parameters of a camera are computed based on the contents of a scene (like the near and far plane). Camera manipulators will also behave differently based on the placement of objects in the scene. I have one node in my scene that I want the camera and any manipulators to 'ignore' in this respect, but I still want the node to be rendered. Is there anyway I can do this? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How can I get the viewer's camera to 'ignore' a node's geometry?
On Thu, Oct 4, 2012 at 3:52 PM, Chris Hanson xe...@alphapixel.com wrote: On Thu, Oct 4, 2012 at 12:48 PM, Preet prismatic.proj...@gmail.com wrote: By default some of the parameters of a camera are computed based on the contents of a scene (like the near and far plane). Camera manipulators will also behave differently based on the placement of objects in the scene. I have one node in my scene that I want the camera and any manipulators to 'ignore' in this respect, but I still want the node to be rendered. Is there anyway I can do this? Probably not. The nodes have to be Culled in order to be added to the Render List, and the scene bounds and such are computed during culling. The bigger question is what are you trying to do, and why? That answer will probably allow us to suggest something useful, like depth partitioning or something. I'm trying to visualize stuff like buildings and roads on the surface of the Earth. But for reference, I'd also like to display the surface of the Earth (just as a set of points). Including geometry for the Earth's surface makes the other stuff disappear (and reappear at certain angles). To me it looks like an issue with the near/far planes of the frustum. I wanted to avoid specifying the near/far planes manually since the behaviour without the Earth surface geometry is perfect as is. I don't care about the overhead of drawing the surface geometry since its so simple, and I don't want the scene graph to account for the surface geometry (outside of rendering it); the other objects in the scene should determine the camera frustum. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] How can I get the viewer's camera to 'ignore' a node's geometry?
On Thu, Oct 4, 2012 at 5:37 PM, Chris Hanson xe...@alphapixel.com wrote: I wanted to avoid specifying the near/far planes manually since the behaviour without the Earth surface geometry is perfect as is. I don't care about the overhead of drawing the surface geometry since its so simple, and I don't want the scene graph to account for the surface geometry (outside of rendering it); the other objects in the scene should determine the camera frustum. Due to the way OGL works with near/far, if you want it drawn, you need it taken into account when computing the near/far planes. Are you just worried about the far plane being set too far? Including geometry for the Earth's surface makes the other stuff disappear (and reappear at certain angles). It sounds to me like you're not trying to solve the right problem. You should first determine WHY these other items aren't displayed at times. Sorry if it wasn't clear; the items aren't being displayed because the near and far planes that osg sets creates a frustum that doesn't include them. I need to manually set the near and far planes to see everything. The near and far planes that osg sets without the earth surface geometry is fine. Once you add the surface geometry though, the near and far planes are set such that the items are no long in the view frustum. I was looking for a flag or something that basically tells osg hey, don't include this node when you do your view frustum calculations ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Manually updating a camera instead of using a manipulator
Figured this out; I was unprojecting coordinates but not updating them
in the right order. So that this thread isn't completely useless, here's my
correct code for converting screen coordinates to world coordinates.
void calcNormDeviceCoords(int xScreen, int yScreen, double xNorm,
double yNorm)
{
xNorm = 2*double(xScreen)/double(MY_SCREEN_WIDTH) - 1;
yNorm = 1 - 2*double(yScreen)/double(MY_SCREEN_HEIGHT);
}
// dNorm = 0.0 for near plane on view frustum, dNorm = 1.0 for far plane
void convCoordScreenToWorld(double xNorm,double yNorm,
double dNorm,osg::Vec3d worldPt)
{
osg::Camera * viewCam = MY_VIEWER-getCamera();
osg::Matrixd xfMVP(viewCam-getViewMatrix()*
viewCam-getProjectionMatrix());
osg::Matrixd xfInvMVP = osg::Matrixd::inverse(xfMVP);
worldPt = osg::Vec3d(xNorm,yNorm,dNorm) * xfInvMVP;
}
Sorry for the spam
Preet
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Window Event Queue not being fully processed?
Hi, I'm using GraphicsWindowEmbedded to draw a 3d viewport in a QtQuick application (osgQt is not applicable here). I pass mouse events from Qt over to OSG by adding them directly to the window's event queue (myGWE-getEventQueue()-mouseButton...). This seems to work as expected for the most part, but it seems like the last couple of mouse events aren't processed. So now, when I again use the mouse to interact with the viewport, the view will jump to a slightly different position before again tracking further mouse events. I'm using a TrackballManipulator, and I've disabled throwing. I also tried setting the viewer's runFrameScheme to ON_DEMAND, but this didn't make a difference. To be clear, I manually call the viewer's frame() method after each mouse event. I'd appreciate any advice. Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Using osgText with GraphicsWindowEmbedded and Qt causes problems
Hi Robert, On Mon, Jul 9, 2012 at 9:01 AM, Robert Osfield robert.osfi...@gmail.com wrote: HI Preet, You'll need to rest OpenGL into consistent state when moving between Qt and the OSG. I don't know anything about QGLFrameBufferObject so can't comment on what state it'll be managing. This was indeed the case. For anyone else that runs into issues integrating the two libs, I've found some OpenGL states that should be reset after drawing an OSG frame, before control is given back to Qt. More state settings might need to be changed depending on what exactly OpenSceneGraph is being used for, but the following seems to work with some basic textured geometry. // qpainter uses vertex attribs 3,4,5 so we need to explicitly disable them: glDisableVertexAttribArray(3); glDisableVertexAttribArray(4); glDisableVertexAttribArray(5); // qpainter requires depth testing to be off glDisable(GL_DEPTH_TEST) // not sure if this is required; osg changes this when using textures so I reset it to default glPixelStorei(GL_UNPACK_ALIGNMENT,4); What I can say is that the OSG has it's own native FBO support that will work just fine all by itself and normally there wouldn't be any need to get QT involved. Personally I'd try to keep Qt out for doing too much OpenGL work itself. Robert. On 5 July 2012 01:30, Preet prismatic.proj...@gmail.com wrote: ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Bug in revision 13092
Did you call realize() on your viewer?
ie... myViewer-realize();
On Fri, Jul 6, 2012 at 7:40 AM, Łukasz Lucek lukasz.lu...@gmail.com wrote:
Hi,
I think that I have found a bug in revision 13092.
I'm using osg as part of qt application, so i create an
osgViewer::GraphicsWindowEmbedded like this:
Code:
_gw = new osgViewer::GraphicsWindowEmbedded(x,y,width,height);
then later, i forward qt events to osg. Especially i forward the resize event
and mouse move events like this:
Code:
void QGraphicsEventAdapter::resizeEvent( QResizeEvent* event ){
_gw-getEventQueue()-windowResize(0, 0, event-size().width(),
event-size().height());
_gw-resized(0, 0, event-size().width(), event-size().height());
}
void QGraphicsEventAdapter::mouseMoveEvent( QGraphicsSceneMouseEvent* event ){
QGraphicsScene::mouseMoveEvent(event);
if( event-isAccepted() ) return;
QPoint widgetPos(0,0);
widgetPos = parentView-mapToGlobal(widgetPos);
_gw-getEventQueue()-mouseMotion(event-screenPos().x()-widgetPos.x(),
event-screenPos().y()-widgetPos.y());
}
And now, i have an osg camera maniupulator derived from
osgGA::StandardManipulator, with a method
Code:
virtual bool performMovementLeftMouseButton( const double eventTimeDelta,
const double dx, const double dy );
my manipulator expects, that the dx and dy values are normalized. It used to
work this way before revision 13092, but after this revision the values are
not normalized anymore.
I think it has something to do with the change made to the Viewer.cpp file,
however it is to complicated for me to dig into osg event handling.
...
Thank you!
Cheers,
Łukasz
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=48733#48733
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [vpb] Texture parameters for Android?
On Sat, Jul 7, 2012 at 10:54 AM, Massimo Tarantini subbi...@yahoo.it wrote:
Hello,
i have compiled OSG and VPB for Android, and i'm using Necessitas on an
Archos 80G9. All the stuff works (well, the touch events are lost...) but i
can't see the texture.
I have tryied some combinations for the options --image-ext and
[--compressed, --RGBA-compressed, --RGB-16, --RGB-24], but unsuccesfully.
I suppose that OpenGL compression can't work on Android. But if i use for
example --image-ext jpeg --RGB-16, why it does not work?
In the debug output i read:
[...]
07-07 16:39:42.039: D/Qt(1611): uncompressed ive stream
[...]
07-07 16:39:42.343: D/Qt(1611): TreatBoundariesToValidDataAsDefaultValue=0
07-07 16:39:42.343: D/Qt(1611): 4
07-07 16:39:42.343: D/Qt(1611): Created new 0xde6e60 TextureObject,
_numOfTextureObjects 5
07-07 16:39:42.343: D/Qt(1611): 4
07-07 16:39:42.343: D/Qt(1611): Created new 0xde6e60 TextureObject,
_numOfTextureObjects 6
07-07 16:39:42.343: D/Qt(1611): 4
07-07 16:39:42.343: D/Qt(1611): Created new 0xde6e60 TextureObject,
_numOfTextureObjects 7
07-07 16:39:42.343: D/Qt(1611): 4
07-07 16:39:42.351: D/Qt(1611): Warning: detected OpenGL error 'invalid
enumerant' at after RenderBin::draw(..)
I ran into something like this with GL ES 2. You didn't specify what
version you compiled against, so this may or may not help you.
Note the error generated. I found a couple of fixes by searching
through the list and after some trial and error this is what worked on
my platform:
In the State header file (/include/osg/State i think) there's a
function called State::applyModeOnTexUnit() that calls
glEnable(GL_TEXTURE_2D). This isn't valid for GLES 2 so comment it
out... I did something like:
#ifndef OSG_GLES2_AVAILABLE
// with opengl es 2, GL_TEXTURE_2D is always enabled
// so trying to enable it explicitly causes errors
if (enabled) glEnable(mode);
else glDisable(mode);
#endif
Also in State.cpp (/src/osg/State.cpp) there's a function called
initializeExtensionProcs(). Line 900 [i pulled osg from git] has an
if() clause that checks for the max number of texture units available.
I replaced it with:
if ( osg::getGLVersionNumber() = 2.0 ||
osg::isGLExtensionSupported(_contextID,GL_ARB_vertex_shader) ||
OSG_GLES2_FEATURES)
{
glGetIntegerv(GL_MAX_COMBINED_TEXTURE_IMAGE_UNITS,_glMaxTextureUnits);
#ifdef OSG_GLES2_AVAILABLE
_glMaxTextureCoords = 1;
#else
glGetIntegerv(GL_MAX_TEXTURE_COORDS,_glMaxTextureCoords);
#endif
}
After those changes, textures started working for me.
However, I noticed you're using Qt/Necessitas... not sure exactly what
you're using Qt for, but just as a heads up, I ran into more problems
using textures with QML that as of now I've been unable to solve (ie
it makes QML unusable with OSG).
07-07 16:39:42.351: D/Qt(1611): 2
07-07 16:39:42.351: D/Qt(1611): end cull_draw() 0xdbf838
07-07 16:39:42.351: D/Qt(1611): 5
Some advice?
Thank you!
Cheers,
Massimo
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=48741#48741
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Saving and restoring GL state with OpenSceneGraph
Hi Frank,
Thanks for the reply! I'll try setting up something similar in my application.
On Fri, Jul 6, 2012 at 3:52 PM, Frank Kane fk...@sundog-soft.com wrote:
From what I've read, captureCurrentState just returns the current state
attributes in OSG - it's not actually capturing the current GL state.
But, you could flip the problem on its head and capture the actual OpenGL
state just prior to calling your library, and then restoring it when your
library is done with its drawing. Here is some code to start from that
catches most of the stuff that's likely to change; for completeness, you'd
want to capture all of the state documented at
http://www.khronos.org/opengles/sdk/1.1/docs/man/glGet.xml as well.
Code:
typedef struct glState_S {
GLboolean depthTest, blend, cullFace;
GLboolean dither, colorLogicOp, polygonOffsetLine, polygonOffsetFill;
GLboolean polygonOffsetPoint, polygonSmooth, scissorTest, stencilTest;
} glState;
static std::stackglState stateStack;
SILVERLININGDLL_API bool PushAllState(void)
{
glState state;
state.blend = glIsEnabled(GL_BLEND);
state.depthTest = glIsEnabled(GL_DEPTH_TEST);
state.cullFace = glIsEnabled(GL_CULL_FACE);
state.dither = glIsEnabled(GL_DITHER);
state.polygonOffsetFill = glIsEnabled(GL_POLYGON_OFFSET_FILL);
state.scissorTest = glIsEnabled(GL_SCISSOR_TEST);
state.stencilTest = glIsEnabled(GL_STENCIL_TEST);
stateStack.push(state);
CheckError(__LINE__);
return true;
}
SILVERLININGDLL_API bool PopAllState(void)
{
if (!stateStack.empty()) {
glState state = stateStack.top();
if (state.blend) glEnable(GL_BLEND); else glDisable(GL_BLEND);
if (state.depthTest) glEnable(GL_DEPTH_TEST); else
glDisable(GL_DEPTH_TEST);
if (state.cullFace) glEnable(GL_CULL_FACE); else
glDisable(GL_CULL_FACE);
if (state.dither) glEnable(GL_DITHER); else glDisable(GL_DITHER);
if (state.polygonOffsetFill) glEnable(GL_POLYGON_OFFSET_FILL); else
glDisable(GL_POLYGON_OFFSET_FILL);
if (state.scissorTest) glEnable(GL_SCISSOR_TEST); else
glDisable(GL_SCISSOR_TEST);
if (state.stencilTest) glEnable(GL_STENCIL_TEST); else
glDisable(GL_STENCIL_TEST);
stateStack.pop();
}
CheckError(__LINE__);
return true;
}
Best regards,
Frank Kane
Founder, Sundog Software LLC
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=48737#48737
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Saving and restoring GL state with OpenSceneGraph
Hi, I'm trying to find out a bit more about how to get OpenSceneGraph to share an OpenGL context with another library... and to do that properly i'd like to isolate both by saving and restoring state as appropriate. I know this isn't the 'right' way to use OSG, and I know that this kills performance, especially since OSG is designed to use lazy state updating... but sometimes its a necessary evil. I'm targeting OpenGL ES2. Looking through some other posts, I've gathered: You can get/save the current stateset by calling something like: myViewer-getCamera()-getGraphicsContext()-getState()-captureCurrentState(myStateSet); You can reset the current state by calling reset: myViewer-getCamera()-getGraphicsContext()-getState()-reset(); You can apply a saved state by calling apply: myViewer-getCamera()-getGraphicsContext()-getState()-apply(myStateSet); But I'm not sure about a few things: When should the osg::State be first saved? So basically, when is the scene graph first traversed and the osg::State built up so that it's appropriate for OSG to go ahead and render something? Does the initial OpenGL state when OSG first builds up osg::State matter (I'm guessing no)? After saving this valid state, i'd do something like this to restore it and render my scene... // draw call myViewer-getCamera()-getGraphicsContext()-getState()-reset(); // is this line necessary? myViewer-getCamera()-getGraphicsContext()-getState()-apply(myStateSet); myViewer-frame(); // save state again if I modified the scene Does that seem reasonable? Any advice/input would be appreciated. Thanks Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] osgText not working with programmable pipeline?
I compiled OSG against OpenGL 3 and disabled all the fixed function stuff, and I can do basic things like build up geometry, apply shaders and so on. Trying to draw text using osgText doesn't work out so well though. I'm not sure what's going on... but basically the output looks like font character quads, just not textured correctly (wild guess). Here's a screenshot from the results of running osgtext in the example folder: http://i.imgur.com/NX6dv.png?1 Do I need to specify my own shader for the text geode I create? If so, are there any examples on how to do this? Also here's the debug output for a minimal osg text example (just create some text, attach it to a geode, and render the scene): http://pastie.org/4200461 I see ShaderComposer is called but I don't know if it's doing anything. Also this line doesn't seem right: Glyph::subload(): texture sub-image width and/or height of 0, ignoring operation. I'd appreciate any advice. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] osgText not working with programmable pipeline?
After some trial and error, osgText renders fine when explicitly
setting up shaders. Here's a bare minimum working example in case
anyone else runs into this.
Scene Graph:
// text geometry
std::string textStr(Hello World);
osg::ref_ptrosgText::Text myText = new osgText::Text;
myText-setFont(DroidSans-Bold.ttf);
myText-setCharacterSize(20.0f);
myText-setText(textStr);
// text node
osg::ref_ptrosg::Geode textGeode = new osg::Geode;
textGeode-addDrawable(myText);
// text shaders
osg::ref_ptrosg::Program shProgram = new osg::Program;
shProgram-setName(TextShader);
std::string vShader = std::string(#version 120\n) +
readFileAsString(vertexshader.glsl);
shProgram-addShader(new osg::Shader(osg::Shader::VERTEX,vShader));
std::string fShader = std::string(#version 120\n) +
readFileAsString(fragmentshader.glsl);
shProgram-addShader(new osg::Shader(osg::Shader::FRAGMENT,fShader));
osg::ref_ptrosg::Uniform textColor = new
osg::Uniform(MaterialColor,osg::Vec4(0,1,1,1));
osg::ref_ptrosg::Uniform textTexture = new osg::Uniform(GlyphTexture,0);
osg::StateSet *ss = textGeode-getOrCreateStateSet();
ss-addUniform(textColor);
ss-addUniform(textTexture);
ss-setAttributeAndModes(shProgram,osg::StateAttribute::ON);
osg::ref_ptrosg::Group nodeRoot = new osg::Group;
nodeRoot-addChild(textGeode);
Vertex Shader:
#version 120
varying mediump vec4 VertexColor;
varying mediump vec2 TexCoord0;
uniform vec4 MaterialColor;
void main()
{
gl_Position = gl_ModelViewProjectionMatrix * gl_Vertex;
VertexColor = MaterialColor;
TexCoord0 = gl_MultiTexCoord0.xy;
}
Fragment Shader:
varying vec4 VertexColor;
varying vec2 TexCoord0;
// set to zero by default
uniform sampler2D GlyphTexture;
void main()
{
gl_FragColor = VertexColor * texture2D(GlyphTexture,TexCoord0).;
}
Preet
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Issue adding transparency to a simple model
Thanks Paul and Sergey I'll try both methods. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Using osgText with GraphicsWindowEmbedded and Qt causes problems
Hi. I have a scene consisting of just an osgText object and its shaders. The scene works fine in a default application where OpenSceneGraph is the only library being used, but won't render correctly when I try to draw to a QGLFrameBufferObject. Drawing to the FBO works fine if I don't use osgText (other than text, I've just tried simple models so far). Using osgText 'messes' the Qt UI up, by changing the order of what (Qt items) are painted or changing their size or stretching them... nothing from OSG seems to be painted at all. I'm guessing there's a conflict with OpenGL calls between Qt and OSG. OSG is set up using GraphicsWindowEmbedded, and in Qt, I bind my framebuffer before calling osgViewer's frame(), and release it after. I'm lost as to why using osgText of all things would cause this issue. Just the inclusion of osgText in my scene tree causes the problem. Does anyone have any suggestions as to how I can force Qt and OSG to play nice? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Issue adding transparency to a simple model
Hi.
I have a simple model I wanted to apply transparency to. I load the
model data into a geometry node, and set my scene up. To get a
transparent effect, I apply an alpha value to gl_FragColor in my
fragment shader.
The shaders are trivial:
Vertex:
void main()
{
gl_Position = gl_ModelViewProjectionMatrix * gl_Vertex;
}
Fragment:
void main()
{
gl_FragColor = vec4(1,0,0,0.5);
}
I get a strange output that sort of works, but renders a different
result based on the camera position. It's hard to describe so I
attached a short youtube clip. It seems like certain triangles are
only rendered at certain camera angles.
http://youtu.be/fQUFSff-uuU
I played around with some of the settings I thought might be relevant
but I still get the effect that the video shows.
osg::StateSet *ss = geodeMesh-getOrCreateStateSet();
ss-setAttributeAndModes(shProgram,osg::StateAttribute::ON);
ss-setMode(GL_BLEND,osg::StateAttribute::ON);
// options i tried changing around
//ss-setMode(GL_DEPTH_TEST,osg::StateAttribute::ON);
//ss-setMode(GL_DEPTH_WRITEMASK,osg::StateAttribute::OFF);
//ss-setMode(GL_CULL_FACE,osg::StateAttribute::OFF);
//ss-setRenderingHint(osg::StateSet::TRANSPARENT_BIN);
Any advice would be appreciated.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Issue adding transparency to a simple model
On Tue, Jul 3, 2012 at 4:45 PM, Paul Martz pma...@skew-matrix.com wrote: What are you doing to ensure that the triangles are rendered in back-to-front order? -Paul Nothing! How can I preprocess the data to achieve that? ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Issue using TrackballManipulator with GraphicsWindowEmbedded
Hey all. I'm having some trouble using TrackballManipulator with GraphicsWindowEmbedded. The two main problems are: * I have to pass normalized x,y coordinates (ie, x/width(), y/height()) to the graphics window's event queue. Looking at other examples, this doesn't seem right. * Zoom and Pan seem to be working, but Rotation just rotates around the plane that the window itself represents (so it rotates the view in 2D on the plane that has a normal in the camera's lookAt direction) I've set up the graphics window and viewer as follows: // setup viewer and window m_osg_viewer = new osgViewer::Viewer; m_osg_window = m_osg_viewer-setUpViewerAsEmbeddedInWindow(0,0,this-width(),this-height()); m_osg_window-getState()-setUseModelViewAndProjectionUniforms(true); m_osg_window-getState()-setUseVertexAttributeAliasing(true); m_osg_viewer-setCameraManipulator(new osgGA::TrackballManipulator); m_osg_viewer-setThreadingModel(osgViewer::Viewer::SingleThreaded); m_osg_viewer-setSceneData(groupRoot); In addition I've tried: * Setting an initial position and perspective matrix for the camera * Calling setMouseInputRange and setUseFixedMouseInputRange on the graphic window's event queue But they didn't have any effect. Am I just missing something obvious? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Issue using TrackballManipulator with GraphicsWindowEmbedded
Ah, forgot to call viewer-realize() Sorry for the spam :) Preet On Sun, Jul 1, 2012 at 4:42 PM, Preet prismatic.proj...@gmail.com wrote: Hey all. I'm having some trouble using TrackballManipulator with GraphicsWindowEmbedded. The two main problems are: * I have to pass normalized x,y coordinates (ie, x/width(), y/height()) to the graphics window's event queue. Looking at other examples, this doesn't seem right. * Zoom and Pan seem to be working, but Rotation just rotates around the plane that the window itself represents (so it rotates the view in 2D on the plane that has a normal in the camera's lookAt direction) I've set up the graphics window and viewer as follows: // setup viewer and window m_osg_viewer = new osgViewer::Viewer; m_osg_window = m_osg_viewer-setUpViewerAsEmbeddedInWindow(0,0,this-width(),this-height()); m_osg_window-getState()-setUseModelViewAndProjectionUniforms(true); m_osg_window-getState()-setUseVertexAttributeAliasing(true); m_osg_viewer-setCameraManipulator(new osgGA::TrackballManipulator); m_osg_viewer-setThreadingModel(osgViewer::Viewer::SingleThreaded); m_osg_viewer-setSceneData(groupRoot); In addition I've tried: * Setting an initial position and perspective matrix for the camera * Calling setMouseInputRange and setUseFixedMouseInputRange on the graphic window's event queue But they didn't have any effect. Am I just missing something obvious? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Problems using shader attributes
- small edit: the correctly rendered scene with straight opengl: http://i.imgur.com/gLLdM.png ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Building a slimmed down static version of OSG
Hi,
I got a basic scene graph running (manually created a simple mesh and
was able to view it). I ran into problems trying to run the osg file
loader plugins and didn't troubleshoot that part any further. Also, I
was building from git.
Do you run into the same crash with 32 and 64 bit builds? What does
the stack trace from the crash say? Also the reimplementation of
GraphicsWindowEmbedded::requestRedraw should be calling
GraphicsWindow::requestRedraw():
void requestRedraw()
{
GraphicsWindow::requestRedraw();
}
... but that doesn't seem to be your issue here.
Preet
On Thu, Jun 21, 2012 at 11:13 PM, David Smith mo...@mokon.net wrote:
Hi Preet,
Did you make any progress on this? I'm getting a similar error and I am
stumped. I have a fedora 16 32bit machine with osg 3.0.1 where my code
compiles fine and then I have a fedora 17 64bit machine also with osg 3.0.1
where I get this undefined reference.
I added a virtual void requestRedraw(){} ; to the GraphicsWindowEmbedded
header file just to see what would happen and it compiles but I get a crash
right away on an addChild in Qt land.
Thanks,
David
--
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=48421#48421
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] 139.com Spam
me three... gmail does a decent job of filtering it out once you mark it as spam though On Thu, Jun 7, 2012 at 4:49 PM, Jordi Torres jtorresfa...@gmail.com wrote: Me too. El 07/06/2012 21:43, rocco martino martinoro...@gmail.com escribió: Idem 2012/6/7 Kim Bale kcb...@googlemail.com Is anyone else getting these Chinese emails quoting the subject of a post made to osg-users? I've had a few now and I have no idea where they're coming from. K. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] OpenSceneGraph on the Playbook
On Wed, Jun 6, 2012 at 5:30 AM, Sergey Polischuk pol...@yandex.ru wrote: Hi gles 2 dont support fixed pipeline lighting stuff like GL_LIGHT# and such, which osg uses to manage lighting. To calculate proper lighting you should supply shaders with light source parameters(position\direction, color etc) through uniforms and use those. Cheers, Sergey. In my previous email I noted that I'm using the shader provided with the Android OpenGL ES2 example. I started to see lighting after I stopped using Qt as a window manager -- I'll deal with that why using Qt changes the output later. I switched to using EGL directly through blackberry's utils. The result is much better, but still flawed. Right now, I'm seeing this weird effect where it seems like faces are either being culled incorrectly or showing the wrong front/back face at certain angles. I don't know how to describe this further so I uploaded a short video: http://youtu.be/_yxLrGJD8HM I'd appreciate it someone could take a look. I also tried explicitly setting cull face front and back state attributes to off, but it doesn't make any difference. The shaders from that example are here: http://pastie.org/4040310: Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] OpenSceneGraph on the Playbook
On Wed, Jun 6, 2012 at 5:48 PM, Chris Hanson xe...@alphapixel.com wrote: It looks like backface culling is off (GLES2 does have back face culling). Turn this on. It also behaves as if you ended up with a context that perhaps has no Z buffer. This can happen if you choose your context poorly via EGL. On some Adreno GLES2 devices I've used, I had to drop down to a 5,6,5 RGB channel depth in order to get a Z buffer. The Z buffer wasn't explicitly set in the blackberry utils... I played around with the screen modes and EGL attributes (used RGBA and had to specify a 24bit depth buffer with 8 bit stencil) and now it works! It's a bit slow on the device with higher poly models (dont know if the screen mode has anything to do with that) but it works! Woo playbook! Thank you so much to all for the help :) Here's a sample project (you'll need to adjust the paths to point to your own osg libs): https://github.com/canurabus/scratch-blackberry/tree/master/template_osg Regards, Preet -- Chris 'Xenon' Hanson, omo sanza lettere. xe...@alphapixel.com http://www.alphapixel.com/ Training • Consulting • Contracting 3D • Scene Graphs (Open Scene Graph/OSG) • OpenGL 2 • OpenGL 3 • OpenGL 4 • GLSL • OpenGL ES 1 • OpenGL ES 2 • OpenCL Digital Imaging • GIS • GPS • Telemetry • Cryptography • Digital Audio • LIDAR • Kinect • Embedded • Mobile • iPhone/iPad/iOS • Android ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] OpenSceneGraph on the Playbook
On Tue, Jun 5, 2012 at 12:17 PM, Chris Hanson xe...@alphapixel.com wrote: You either need to write your own GLES2 shader (that's GL2 you have above) or use a shader gen tool that will generate one for you on the fly. Where did the shader, above, come from? It was automatically genereated by osg... I didn't explicitly add a shader. Following your (and Sergey's) advice, I tried using the vertex and fragment shaders from the Android OpenGL ES2 example. I don't get any errors from OSG, but I still don't get any output except for an empty viewport. I've made sure my object is in front of the camera, etc. osg output: http://pastie.org/4034304 source (same as above, just using shaders now) http://pastie.org/4034318 Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] OpenSceneGraph on the Playbook
Hiya,
I managed to compile and deploy a small OSG test application on the
Playbook, but have run into some trouble getting it to work correctly.
I'm not creating my own context or windowing system; instead I'm using
osg's GraphicsWindowEmbedded with Qt. As a test, I set up a scene with
a rotating cube. The application seems to display the viewport
(there's the telltale purple-blue background in OSG), but fails to
render the cube.
The output seems to indicate that creating a shader for the cube
fails, but I can't discern any more than that. It feels like this has
something to do with OpenGL ES2... I compiled OSG for ES2 based on the
instructions I found on the mailing list and the OSG website. I've
pasted the output I think is relevant below, with the full output
here: (http://pastie.org/4027605), and the OSG scene set up here:
(http://pastie.org/4027631). The application works fine and renders as
expected using OpenGL on my desktop. Could anyone suggest something to
get this working?
Preet
glVersion=2, isGlslSupported=YES, glslLanguageVersion=1.016
State::convertShaderSourceToOsgBuiltIns()
++Before Converted source
void main()
{
gl_Position = ftransform();
gl_FrontColor = gl_Color;
}
Converted source
uniform mat4 osg_ModelViewProjectionMatrix;
attribute vec4 osg_Color;
attribute vec4 osg_Vertex;
void main()
{
gl_Position = osg_ModelViewProjectionMatrix * osg_Vertex;
gl_FrontColor = osg_Color;
}
Compiling VERTEX source:
1: uniform mat4 osg_ModelViewProjectionMatrix;
2: attribute vec4 osg_Color;
3: attribute vec4 osg_Vertex;
4:
5: void main()
6: {
7: gl_Position = osg_ModelViewProjectionMatrix * oVERTEX
glCompileShader FAILED
VERTEX Shader infolog:
0:8(16): error: `gl_FrontColor' undeclared
0:8(16): error: type mismatch
FRAGMENT glCompileShader FAILED
FRAGMENT Shader infolog:
0:6(19): error: `gl_Color' undeclared
0:6(19): error: Operands to arithmetic operators must be numeric
0:6(19): error: type mismatch
glLinkProgram FAILED
Program infolog:
linking with uncompiled shaderlinking with uncompiled shader
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Enabling notifications from within program?
Thanks to all for the replies... I wasn't getting any output because
of some unrelated linking errors. After fixing them, debugging is
working as expected :)
On Fri, Jun 1, 2012 at 2:46 AM, Sergey Polischuk pol...@yandex.ru wrote:
Hi
You can set your own notify handler like this to check if you ever get
messages from osg:
class NH : public osg::NotifyHandler
{
public:
NH()
: m_out(osg_log.txt)
{
}
~NH()
{
m_out.close();
}
void notify (osg::NotifySeverity severity, const char *message)
{
m_out message;
}
private:
std::ofstream m_out;
};
osg::setNotifyHandler(new NH);
Cheers,
Sergey.
31.05.2012, 22:25, Preet prismatic.proj...@gmail.com:
Hi Robert,
That's what I'm currently doing (ie using stack traces to try and
figure stuff out), but I was wondering why osg doesn't give me *any*
output at all. For instance, I'm apparently able to do a bunch of
stuff: Load an *.osg model, setup an animation path, create nodes,
etc, pretty much all of the scene setup goes fine up until I try to
create the viewer. Doesn't doing any of those things before the viewer
provide output when I've set the notify level to debug? I made sure I
compiled with OSG_DISABLE_NOTIFY (or whatever the exact wording for
that flag was) to allow notifications.
On Thu, May 31, 2012 at 4:57 AM, Robert Osfield
robert.osfi...@gmail.com wrote:
Hi Preet,
The osg::notify system isn't related to handling of std exceptions
except where some specific code might catch an exception and report
the output to osg::notify. This means that upping the notify level
won't effect how exceptions are handled.
The best thing to do is run a debugger and the look at the stack trace
where the application crashes.
Robert.
On 31 May 2012 07:57, Preet prismatic.proj...@gmail.com wrote:
Hiya,
I'm trying to debug a std::bad_alloc() exception from osg that doesn't
product any other output. I can't set environment variables, so I
tried:
osg::setNotifyLevel(osg::DEBUG_INFO);
This still doesn't give me any output. The system I'm on dumps stdout
and stderr to a log file in a specific directory and since Notify.cpp
dumps to both of those it seems like I should be seeing something, but
I'm not. However, I do get statements like std::cout Hello
std::endl outputted as expected when used in my application (just not
from osg). Am I missing something obvious?
Preet
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Enabling notifications from within program?
Hiya, I'm trying to debug a std::bad_alloc() exception from osg that doesn't product any other output. I can't set environment variables, so I tried: osg::setNotifyLevel(osg::DEBUG_INFO); This still doesn't give me any output. The system I'm on dumps stdout and stderr to a log file in a specific directory and since Notify.cpp dumps to both of those it seems like I should be seeing something, but I'm not. However, I do get statements like std::cout Hello std::endl outputted as expected when used in my application (just not from osg). Am I missing something obvious? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Explanation of macros in staticviewer example?
Hiya, In the osgstaticviewer.cpp example, there are three sets of macros: USE_OSGPLUGIN()... USE_DOTOSGWRAPPER_LIBRARY()... USE_SERIALIZER_WRAPPER_LIBRARY()... When compiling a static application, how do we know which ones we need to include? It seems like they're all related to osgDB with USE_OSGPLUGIN corresponding to plugins. I'm not sure what the other two are for. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Building a slimmed down static version of OSG
Quick follow up: The error seems to occur when linking using GCC 4.4.2, which is part of the tool chain I use. When I use my distro's GCC 4.7 to do a static build with the same CMake settings, I don't get the error. To get around it in 4.4.2, I explicitly defined requestRedraw() in GraphicsWindowEmbedded and had it call GraphicsWindow::requestRedraw(). I don't know if this creates any new problems since I still haven't gotten osg working, but it looks like it shouldn't do anything evil and may be worth changing in source if someone else can confirm its a bug with older compilers. Status so far: I can deploy an app with osg to the PlayBook but it instantly crashes. Right now I'm basically sifting through core dumps and seeing what else I can do to try and get it to work. Preet On Thu, May 31, 2012 at 12:07 PM, Thomas Hogarth thomas.hoga...@gmail.com wrote: Hi Preet Not too sure what exactly your issue is but the error (.rodata._ZTVN9osgViewer22GraphicsWindowEmbeddedE[_ZTVN9osgViewer22GraphicsWindowEmbeddedE]+0xcc):-1: error: undefined reference to `non-virtual thunk to osgViewer::GraphicsWindow::requestRedraw()' Suggests that your osgViewer lib isn't linked or GraphicsWindow.cpp wasn't compiled into your lib. Setting the windowing system to None is correct but it may be that it needs to be set before you run cmake. Look at osgViewers CMakeLists.txt and see how it's done on Android. Cheers Tom -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=47975#47975 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Enabling notifications from within program?
Hi Robert, That's what I'm currently doing (ie using stack traces to try and figure stuff out), but I was wondering why osg doesn't give me *any* output at all. For instance, I'm apparently able to do a bunch of stuff: Load an *.osg model, setup an animation path, create nodes, etc, pretty much all of the scene setup goes fine up until I try to create the viewer. Doesn't doing any of those things before the viewer provide output when I've set the notify level to debug? I made sure I compiled with OSG_DISABLE_NOTIFY (or whatever the exact wording for that flag was) to allow notifications. On Thu, May 31, 2012 at 4:57 AM, Robert Osfield robert.osfi...@gmail.com wrote: Hi Preet, The osg::notify system isn't related to handling of std exceptions except where some specific code might catch an exception and report the output to osg::notify. This means that upping the notify level won't effect how exceptions are handled. The best thing to do is run a debugger and the look at the stack trace where the application crashes. Robert. On 31 May 2012 07:57, Preet prismatic.proj...@gmail.com wrote: Hiya, I'm trying to debug a std::bad_alloc() exception from osg that doesn't product any other output. I can't set environment variables, so I tried: osg::setNotifyLevel(osg::DEBUG_INFO); This still doesn't give me any output. The system I'm on dumps stdout and stderr to a log file in a specific directory and since Notify.cpp dumps to both of those it seems like I should be seeing something, but I'm not. However, I do get statements like std::cout Hello std::endl outputted as expected when used in my application (just not from osg). Am I missing something obvious? Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Building a slimmed down static version of OSG
After a bit more playing around I'm able to get the linking down to one error (the previous errors were because of linking in the wrong order): (.rodata._ZTVN9osgViewer22GraphicsWindowEmbeddedE[_ZTVN9osgViewer22GraphicsWindowEmbeddedE]+0xcc):-1: error: undefined reference to `non-virtual thunk to osgViewer::GraphicsWindow::requestRedraw()' I wanted to make sure I wasn't screwing up static compilation, so I tried shared libs as well and got the same error message. I have no idea what this error message means and google wasn't much help (at least it feels like I'm getting closer!) * I'm using the latest source from git since 3.0.1 and 3.1.2 failed to compile against OpenGL ES 2 (for me anyway) * I set the windowing system to None in CMake before compiling (is that what's causing this issue?) * Qt is available on the device and I was planning to use GraphicsWindowEmbedded with Qt's OpenGL context I'd appreciate some help in figuring out where to go from here. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Building a slimmed down static version of OSG
Quick follow up, After playing around with the CMake gui for awhile, I managed to get it to compile a set of static libs. Unfortunately it seems that OpenThreads is missing from the built libs, and trying to compile a project that links with osg results in a bunch of errors regarding missing OpenThreads symbols. The Playbook runs QNX, which has pthreads. The CMake Configure still seems to fail the OpenThreads tests (is there anything I can point CMakeLists to so it'll pass the tests?). I'd appreciate any advice. Regards, Preet On Mon, May 28, 2012 at 6:00 PM, Preet prismatic.proj...@gmail.com wrote: Hi, I'm trying to build OSG for another platform (the BlackBerry Playbook). The NDK for the Playbook includes the tool chains required to compile libs/apps for the device. It also has OpenGL ES 1.1/2.0 libs, EGL, freetype, etc. I feel like it should be possible to compile OSG for the Playbook but I'm a little lost dealing with the included CMakeLists. I tried looking at this for reference: http://www.openscenegraph.com/index.php/documentation/platform-specifics/ios/23-configuring-cmake-for-ios, but I don't understand how to enable/disable certain plugins for OSG when configuring with CMake or how to know what's really required in a minimal build. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Building a slimmed down static version of OSG
Thanks for the reply. After messing with the CMakeLists in OpenThreads (forced pthreads), I got it to compile and build me a static lib archive... unfortutely I have pretty much the same problem, about a bajillion of these errors originating from various osg sources. Program.cpp:(.text+0x5158): undefined reference to `OpenThreads::Mutex::Mutex(OpenThreads::Mutex::MutexType)' Program.cpp:(.text+0x5160): undefined reference to `OpenThreads::Mutex::~Mutex()' On Tue, May 29, 2012 at 5:02 PM, Thomas Hogarth thomas.hoga...@gmail.com wrote: Hi Preet Good to hear you're working on a Blackberry port. In regard to the libs you need your main aim is to get osgViewer compiling. If that compiles then you can setup a viewer a render your scenegraph. You'll have two options for this, you can either create your own implementation of osgViewer::GraphicsWindow to handle setting up an OpenGL context, window etc. Or you can just use the OpenGL example you have and setup the viewer using the setupViewerAsEmbeded function which will allow you to draw the osg scenegraph into your example OpenGL view. Probably easiest to try setupViewerAsEmbeded first to test osg itself is working, then you can try and create your own viewer. With regard to OpenThreads it's probably related to the actual operating system you are using to build the libs. For example if you are on windows then CMake is probably selecting the windows threading stuff because it's not aware your trying to build for blackberry. If you take a look at src/OpenThreads/CMakeLists.txt near the bottom you'll find this Code: IF(NOT ANDROID) # Use our modified version of FindThreads.cmake which has Sproc hacks. FIND_PACKAGE(Threads) ENDIF() # Do we have sproc? IF(CMAKE_SYSTEM MATCHES IRIX) IF(CMAKE_USE_SPROC_INIT) # In this case, only Sproc exists, so no option. SET(OPENTHREADS_USE_SPROC_INSTEAD_OF_PTHREADS 1) ELSE() IF(CMAKE_HAVE_SPROC_H) OPTION(OPENTHREADS_USE_SPROC_INSTEAD_OF_PTHREADS Set to ON to build OpenThreads against sproc instead of pthreads OFF) ENDIF() ENDIF() ENDIF(CMAKE_SYSTEM MATCHES IRIX) # Add support for using Qt threads IF(QT4_FOUND) OPTION(BUILD_OPENTHREADS_WITH_QT Build OpenThreads with Qt threading support. OFF) ENDIF() # Maybe we should be using the FindThreads.cmake module? IF(ANDROID) SUBDIRS(pthreads) ELSEIF(QT4_FOUND AND BUILD_OPENTHREADS_WITH_QT) SUBDIRS(qt) ELSEIF(WIN32) # So I think Cygwin wants to use pthreads IF(CYGWIN) SUBDIRS(pthreads) ELSE() # Everybody else including Msys should probably go here SUBDIRS(win32) # examples) ENDIF() ELSE() IF(UNIX) IF(OPENTHREADS_USE_SPROC_INSTEAD_OF_PTHREADS) SUBDIRS(sproc) ELSE() SUBDIRS(pthreads) ENDIF() ELSE() MESSAGE(Sorry, OpenThreads may not support your platform) ENDIF() ENDIF() In there it is selecting the correct threading module for the platform. For a temporary fix you can just force the use of pthreads. Regarding plugins, cmake will only generate build targets for pulgins for which it has found the dependancies. Be careful though that it doesn't find your operating system versions. For good asset support you'll need at least libpng for loading of png images, and freetypes for fonts. The osg format plugins themselves have no external dependancies so should build as long as osg does. Cheers Tom -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=47924#47924 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Building a slimmed down static version of OSG
Hi, I'm trying to build OSG for another platform (the BlackBerry Playbook). The NDK for the Playbook includes the tool chains required to compile libs/apps for the device. It also has OpenGL ES 1.1/2.0 libs, EGL, freetype, etc. I feel like it should be possible to compile OSG for the Playbook but I'm a little lost dealing with the included CMakeLists. I tried looking at this for reference: http://www.openscenegraph.com/index.php/documentation/platform-specifics/ios/23-configuring-cmake-for-ios, but I don't understand how to enable/disable certain plugins for OSG when configuring with CMake or how to know what's really required in a minimal build. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Current state of GLES2
On Tue, Apr 24, 2012 at 4:44 AM, Robert Osfield robert.osfi...@gmail.com wrote: However, you can use the OSG with OpenGL 2.0 with a purely shader based state and if you do this the actual shaders you use for OpenGL or OpenGL ES 2.0 will be very similar. This similarity means you can prototype an OpenGL ES application in OpenGL just be using the shaders and avoiding all built in state. I'm fuzzy on the shader stuff. For all my 3d dev, I just jumped into using osg, and whatever I've built has been derived from tutorials/examples given online. With the level of abstraction osg provides, I haven't had to make any OpenGL calls directly, so when you say using the shaders and avoiding all built in state... I don't understand what avoiding all built in state is in an osg context. Does it mean I shouldn't be using osg::StateSet? You can also build the OSG so that it doesn't compile in the fixed function state, something that you do when you want to compile against the core profile of OpenGL 3.x. I'm using http://www.openscenegraph.org/projects/osg/wiki/Community/OpenGL-ES to set the environment up... and since I'd like to develop for OpenGL ES 2, I'll be using Cmake settings for Emulating OpenGL ES 2.0 when you only have standard OpenGL available to build osg to create my app. When it comes time to actually deploy, I'm guessing I'd rebuild osg against Cmake settings for OpenGL ES 2.0? I also downloaded an emulator/sdk for OpenGL ES 2 from PowerVR, and ran a bunch of default examples they provided. Does that factor in at all? They provide libEGL.so and libGLESv2.so libs for linking as well. On Mon, Apr 23, 2012 at 6:19 PM, Chris Hanson xe...@alphapixel.com wrote: Yes, with ShaderGen. What exactly is involved in using ShaderGen? From the osgshadergen example, it looks like all that's done it is attaching it as a node visitor to your scene graph. Does it just automagically go through your scene and convert fixed function pipeline stuff to shader pipeline stuff? Which OS? OSG really only has ES2/EGL windowing interfaces for Xwindows -- Win32 is not implemented. I'm on Linux Sorry for all the questions and thanks for the valuable feedback/support Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Current state of GLES2
Hiya, Thanks Robert and Chris for your insight and advice. I think I'll try deploying using OpenGL ES 1.1 with what I have for now, and rebuild the OSG stuff using shaders to eventually build with ES 2.0. As Robert mentioned, this way I'll get a decent amount of exposure to GLSL and the new pipeline before getting in way over my head. Well, if you got the PowerVR emulator working, you're ahead of me. I was surprised that it 'just worked'... that usually never happens with Linux. Downloaded it, unzipped it and all the examples ran flawlessly. It's even more surprising considering I have a really old machine with spotty driver support (integrated ATi chipset). If it helps, I'm running Arch Linux with all the latest updates. The SDK I got was the ES 2.0 Emulation for PC (Linux) off this page: http://www.imgtec.com/powervr/insider/sdkdownloads/. Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Current state of GLES2
Hiya, Not to hijack this thread, but I'm also interested in porting my osg application to an OpenGL ES2 device. I'm new to OSG and OpenGL but I understand that ES2 has some pretty significant changes in how things are drawn. What sort of steps should I be taking to accomplish this? I'm not sure what level of 'continuity' OSG offers between the different OpenGL versions, and I didn't design the original application with OpenGLES in mind (no shaders, just simple geometry, transforms and materials). Based on this thread I get the feeling I shouldn't expect my application to 'just work'... so what's the right way to proceed? Does 'most' of OSG's basic functionality still work with ES2? Will I need to rewrite much of my code? I downloaded an OpenGLES2 emulator from PowerVR... I'll try to compile OSG for ES2 and see if I can get any examples working, but beyond that I really don't know how to approach this and would appreciate any advice. Regards, Preet On Mon, Apr 23, 2012 at 4:51 PM, Eduardo Poyart poy...@gmail.com wrote: Very clarifying, thank you. I did experiment with ShaderGen and I even patched it to generate GLES2 code. However I'm going in a different direction. My models are simple. I wrote my own shader and added it to the root node. My shader currently has a fixed light position for testing purposes. Now I want to render it with textures. If OSG would detect GLES2 and appropriately write standard-named uniforms for textures, I could just use those names in the shader. Maybe that needs to be done in State::applyModeOnTexUnit, since there is code there for fixed-function Opengl (glEnable(GL_TEXTURE_2D)). If that seems like a good direction, I'll try to do it for textures. The same thing can be done for light positions later. This way I leave ShaderGen out of the way, and it can be improved later. Eduardo On Mon, Apr 23, 2012 at 7:51 AM, Robert Osfield robert.osfi...@gmail.com wrote: Hi Thomas et. al, I won't have a chance to complete the shader composition functionality in the near term so an improved ShaderGen would be welcome. Handling items like lights, texgen and clipplanes will require some handling elsewhere other than with just the standard functionality - we'll need to have the osg::Light, osg::TexGen and osg::ClipPlane all be written as appropriate osg_* uniforms with the values being passed to OpenGL using the current modelview matrix as is done by OpenGL with the fixed function pipeline. This might require amendments to osg::State/osgUtil::RenderStage and the osg::Light, osg::TexGen and osg::ClipPlane, but it might also be possible with just amendments to state attributes if they do a little compute work prior to passing their values to OpenGL as uniforms. I would suggest finding a workable solution for lights then when this works roll it out for the other positional state. Thoughts? Robert. On 23 April 2012 15:16, Thomas Hogarth thomas.hoga...@gmail.com wrote: Yes I have a GLES2 version of shader gen. At the moment it only supports a fixed hardcoded light position, which isn't great but it does support, texturing, lighting etc. I've attached it and we can look at including into core osg, but I think support for at least the initial position of one light source would be good. Also we need to either test for backward compatibility with GL2 or have two versions of ShaderGen in the core. I know Roberts long term goal is to introduce a Shader Composer system (bits and pieces already exist), but this can work as an aid to new developers until then. Let me know how you get on with it. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] How do I fill the framebuffer with an alpha channel?
Hey, I have some geometry which has a single primitive set that's a tri-strip. Some of the triangles in the primitive overlap, so when I add a material to the geometry with an alpha value I see the overlap as shown here: http://i.stack.imgur.com/s5yqV.png I don't want the object's opacity to be affected by self interferences. I read that one way to solve this in generic OpenGL is to fill the framebuffer with your desired alpha channel, switch the blending mode to glBlendFunc(GL_ONE_MINUS_DST_ALPHA, GL_DST_ALPHA) and draw the geometry. I can use osg::BlendFunc properly, but... * I don't know how to preemptively fill the framebuffer with your desired alpha channel with osg. Can anyone provide any pointers as to how that can be done? Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [build] How can I obtain the osgQt library?
Hiya, Same thing happened to me. You didn't mention which OS you were on, but if you can't find osgQt separately for your distro, you can always grab from source and compile it. Preet On Tue, Apr 17, 2012 at 3:31 AM, Mark Green hlgamead...@hotmail.com wrote: Hi, This is probably a stupid question, but how do I obtain the osgQt lib file? Unlike the other libraries I needed, It did not seem to come with my original install, and I have been unable to locate the lib file in any of the downloads on this site or anywhere else on the net. Thank you! Cheers, Mark -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=47088#47088 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] [build] How can I obtain the osgQt library?
Hiya, I'm not too sure how to use CMake with Windows. There was an older message on this list you might find helpful: http://comments.gmane.org/gmane.comp.graphics.openscenegraph.user/67825 Its a bit overkill because it compiles all of osg if you follow it step for step. When you run CMake on the top level directory of OpenSceneGraph, it'll generate the files necessary to build all the OpenSceneGraph components. Depending on your IDE/build setup once you run CMake, you can probably figure out how to compile just osgQt. Again, I'm not too familiar with building on Windows, so maybe someone else from the list can chime in. Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Rendering osgText::Text using osg::Material?
Hey all, I'd like to render osgText using a material instead of the default osgText::setColor() method. This is because I instance the same text geometry, but would like to paint it different colors without duplicating anything. To do this, I add a material as an attribute to the text node. The material has colorMode set to OFF. Setting the diffuse color to my target color doesn't really work (alpha value works, but colors dont), but setting the emission color works fine. Why is this? Lighting is on. osg::Material myMat = new osg::Material; myMaterial-setColorMode(osg::Material::OFF); myMaterial-setDiffuse(osg::Material::FRONT,osg::Vec4(1,0,0,1)); ... osgText::Text textChar = new osgText::Text; textChar-getOrCreateStateSet()-setAttribute(myMaterial.get()); // black text, but alpha value works myMaterial-setEmission(osg::Material::FRONT,osg::Vec4(1,0,0,1)); // red text, alpha works Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Quick questions about assigning render bins
I need to render a bunch of different objects using layer data and I'm doing it by disabling depth test and specifying a render bin to control render order... something like myObject-getOrCreateStateSet()-setMode(GL_DEPTH_TEST,osg::StateAttribute::OFF); myObject-getOrCreateStateSet()-setRenderBinDetails(11,RenderBin); * Should I be doing this on a per-object basis, or would I see better performance if I created group nodes with the right renderbin setup and then added the objects to the group nodes (assume hundreds of objects, but 10-20 layers)? (I don't know how OSG sorts/orders the scene graph nodes before rendering) * Some of my objects have transparency, but again, I'm specifying render order. So is it okay to put those objects in the RenderBin renderbin as opposed to the DepthSortedBin? If Objects A and B are both transparent, both in RenderBin (10) and RenderBin (11) respectively, and A is behind B, I should see object A through B correct? This is what some preliminary test code seems to indicate (and in that code switching from DepthSortedBin to RenderBin doesn't result in any changes). Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Unexpected osgText::TextBase::computeBound() behaviour
Hi all Calling computeBound() on an osg::Text object seems to return height values as if the text were aligned using the baseline of the first line of text only. So if I had a text object with a total height of 50, using an alignment of osgText::TextBase::CENTER_CENTER positioned at (0,0,0), I'd expect yMin and yMax from the bounding box to be at -25 and 25. Instead it'll be according to what I specified previously. This happens with or without multiple lines in the text. I've linked a minimal example and a screenshot illustrating the issue. The screenshot shows the bounding box with respect to the text. Ignore the glitchy character glyphs -- that's an issue with the drivers on my system. screenshot: http://imgur.com/YJasY code: http://pastie.org/3754296 Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Can I have instances (vs. clones) with different colors?
Hey all, My scene makes use of a *lot* of repeating geometry. The geometry is either simple primitives (sphere, pyramids, etc) or osgText::Text. These objects are duplicated many times and are scaled/rotated/translated all over the scene. The objects need to have differing colors and transforms, but other than that they have identical geometry. Is there any way I can use instances of the same geometry instead of cloning them over and over again? I can parent the geometry nodes with their corresponding transforms using the same instance, but I can't set a different color. Just to try it out, I used instancing (with the same color for everything) as opposed to cloning, and the time to render the scene was cut by half (which is important in my application since I'm targeting mobile). I'd appreciate any advice! Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] osgUtil::Tessellator not working with double precision
Hiya, I need Vec3d precision for my geometry. I spent a long time trying to figure out why tessellator wouldn't work at all on my data, until I changed all of my Vec3d's to 'Vec3' on a whim. Now it works fine... but the issue still remains. How can I use double precision with tessellator? Regards Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Efficient usage of Tessellator
Hiya, I have a bunch of shapes that can be kind of complex that I want to draw in osg, and I only have their border points available. I can draw them fine if I use the tessellator util. What I'd like to know is what underlying primitives are sent to OpenGL when I do tessellation. For example, here's some code: osg::ref_ptrosg::Geometry geom = new osg::Geometry; geom-setVertexArray(vertices.get()); geom-addPrimitiveSet(new osg::DrawArrays(GL_POLYGON,0,numOuterPolyPoints)); geom-addPrimitiveSet(new osg::DrawArrays(GL_POLYGON,numOuterPolyPoints, numInnerPolyPoints)); osgUtil::Tessellator geomTess; geomTess.setTessellationType(osgUtil::Tessellator::TESS_TYPE_GEOMETRY); geomTess.retessellatePolygons(*geom); osg::ref_ptrosg::Geode nodeGeom = new osg::Geode; nodeArea-addDrawable(geom); myRootNode-addChild(nodeGeom); I've read that GL_POLYGON is slow, unavailable with OpenGL ES and even deprecated after 3.x. Does the Tessellator change the initial primitive type? Should I be doing the above code differently to maximize compatibility and/or performance (ie use a triangle primitive)? If so, how? Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Creating an 'outline' effect in osgText?
Hey all, I'm trying to create an outline effect with osgText. Here's an example of what I'm trying to accomplish: http://gimpology.com/uploads/3_8_8.png How can I get a similar effect? Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Creating an 'outline' effect in osgText?
Hey, On Thu, Mar 22, 2012 at 7:03 PM, Hamm, Brandon brandon.h...@qinetiq-na.com wrote: Preet, Take a look at the osgtext example. There's a setBackdropType() member function of osgText::Text that you can call to set a backdrop type of osgText::Text::OUTLINE. You can specify the color of the outline using the member function setBackdropColor. Thanks. This seems to be exactly what I'm looking for. Brandon Hamm I have a general follow up question about osgText. I'd like to draw the text along a path. I thought a good way to do this would have been to rotate the text character by character. So I'd create an osgText object for each character of text, and position/align them along the path... but this doesn't seem to preserve spacing between characters (this is expected, but I don't really know how to get around it), and the space character itself ( ) as an individual osgText object seems to have a width of zero (calculated with osg::Drawable::getBound()). * Is the approach I'm taking a 'good' one? If not, what would be a better methodology? ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Maintaining a reference to an osg node outside of osg
Hiya,
Typically I use osg::ref_ptr to let osg handle reference counting.
Right now I'm trying to create a 'rendering engine' for another
library with osg. The idea for the other library is to maintain an API
that allows different rendering engines -- osg, ogre, vtk, etc. The
other library has virtual methods that are called when objects need to
be added and removed from the scene, ie:
void RemoveObjectFromScene(SomeStruct myStruct)
The objects can be thought of as having a 1:1 relation with osg nodes.
I'd like to maintain a reference to the osg node corresponding to the
object so I don't have to traverse the scene graph to find a given
node (there are a lot of objects and this would get way too expensive
I think). I can modify SomeStruct as long as its library-agnostic --
so no osg specific stuff goes in there. Could I use a void * or
something similar to store a reference?
SomeStruct
{
void * mysteryPtr;
}
{
osg::ref_ptrosg::Node myNode = new osg::node; // 1 ref to myNode
m_persistant_root_node-addChild(myNode); // 2 refs to myNode
mysteryPtr = myNode.get(); // 2 refs to myNode (no change!)
}
// 1 refs to myNode since ref_ptr has gone out of scope as long as
m_persistent_root_node is alive
// mysteryPtr has NO bearing to any ref counting, except for the
fact that's its a valid pointer to osg::node as
// long as we have 0 refs to it
I'd be very grateful if someone could verify the above logic. If I'm
in the wrong here, I'd appreciate any advice pointing me in the right
direction.
Regards,
Preet
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Maintaining a reference to an osg node outside of osg
Hi, On Fri, Mar 16, 2012 at 4:43 PM, John Kaniarz john.kani...@us.army.mil wrote: What you're trying to do is the PIMPL design pattern. I'll leave the details to Google :) You may also want to brush up on the Factory design pattern for generating your opaque handles. In a direct answer to your question, no you can't use a void* to store a ref_ptr (without some trickery). You can, however, store a pointer to a ref_ptr as a void* . Then when you need to use it, convert it like this to avoid having to double dereference it every time you use it. If I do this, then I need to maintain the underlying ref_ptr object no? This isn't particularly convenient. What was wrong with assigning the osg::Node's pointer to mysteryPtr instead? Basically with ... osg::ref_ptrosg::Node someNode; mysteryPtr = someNode.Get(); As long as the pointer returned by the Get() method is valid for the lifetime of the object I'm safe right? Then if I want to do anything with object: osg::ref_ptrosg::Node someNode = some_cast_callosg::Node*(mysteryPtr); someNode-partyAllDay(); Code: ref_ptr Node node = *reinterpret_cast ref_ptr Node *(mysteryPtr); And don't forget to delete the pointer to the ref_ptr when you're done. if you leak the ref_ptr the node will never get collected. -- Read this topic online here: http://forum.openscenegraph.org/viewtopic.php?p=46372#46372 ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Stitching together sphere geometry in OpenSceneGraph
Thanks, setting the texture's wrap mode to CLAMP_TO_EDGE worked perfectly! Regards, Preet On Thu, Mar 1, 2012 at 10:45 PM, Jean-Sébastien Guay jean_...@videotron.ca wrote: Hello Preet, * The issue I've run into involves where my 2D texture wraps around and meets itself on the sphere: http://i.imgur.com/ftVH2.png I don't think there's anything wrong with your geometry and texcoords. Good work! The problem looks like you may be using the default texture wrap mode, which is CLAMP. This mode can also be called clamp to border, which when texture coordinates run outside the range of 0-1 will give a border color, which defaults to black. The problem is that when the texture coordinates are exactly 0 or 1, when you have linear interpolation enabled as filter mode (again the default), then the border color will still get sampled (since it's trying to filter using that texel and the one next to it, which is outside the 0-1 range). So you get those black lines around your textures. Try to specify the CLAMP_TO_EDGE wrap mode instead. This will simply sample the same texel at the border and should eliminate the black line. Hope this helps, J-S -- __ Jean-Sebastien Guay jean_...@videotron.ca http://whitestar02.dyndns-web.com/ ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Objects with (very) high mesh resolution not showing up in viewer?
Hiya, I used OpenSceneGraph to generate some straightforward geometry; an ellipsoid representing the Earth. I wrote a function which calculated vertices, normals, texture coords and indices for the geometry. The function takes the number of 'sectors' and 'rings' as arguments to generate spherical geometry. It works fine until I hit some limit as I increase the resolution of the mesh (by passing a larger number of sectors and rings to the function), after which point osgviewer doesn't display the data. There aren't any errors; the geometry just doesn't show up in the viewer. If I go back beyond some threshold the geometry shows up again. What's going on? Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Objects with (very) high mesh resolution not showing up in viewer?
I think it might be a scale issue wrt to the camera. I just found out about the OSG_NOTIFY_LEVEL env var, and I get a bunch of messages about the viewer's drawing traversal: cull_draw() 0x988f788 _clampProjectionMatrix not applied, invalid depth range, znear = 3.40282e+38 zfar = -3.40282e+38 end cull_draw() 0x988f788 I'm using large scaled units (ie, millions) in the scene. Maybe the geometry isn't within the camera's frustum, or something like that? I don't get why change the number of vertices that make up the geometry would cause a change like that though. I construct a spheroid using arguments for parameters that correspond to latitude / longitude style divisions. The latitude argument defines how many 'rings' go down one axis of the sphere, and the longitude argument defines how many 'sectors' each 'ring' is divided into. So if I specify 36 rings and 72 sectors for each ring, I expect my sphere to have (36*72) vertices. This works fine. If I up the ante to 48 rings and 72 sectors, osgViewer won't show the geometry. The function I call isn't doing anything special... and seems to complete fine before being sent to the viewer. http://pastie.org/3497231 - line 184 is where I defined the function. I'm going to Preet On Thu, Mar 1, 2012 at 7:00 AM, Sebastian Messerschmidt sebastian.messerschm...@gmx.de wrote: Hi Preet, can you give us some more details on the limit you are hitting here? Also I don't see any hints how your geometry is organized? As a single drawable? If the latter is the case, I guess you are simply hitting some OpenGL/Driver limits regarding the maximum size of a draw array. cheers Sebastian Hiya, I used OpenSceneGraph to generate some straightforward geometry; an ellipsoid representing the Earth. I wrote a function which calculated vertices, normals, texture coords and indices for the geometry. The function takes the number of 'sectors' and 'rings' as arguments to generate spherical geometry. It works fine until I hit some limit as I increase the resolution of the mesh (by passing a larger number of sectors and rings to the function), after which point osgviewer doesn't display the data. There aren't any errors; the geometry just doesn't show up in the viewer. If I go back beyond some threshold the geometry shows up again. What's going on? Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Stitching together sphere geometry in OpenSceneGraph
Hey all, I'm trying to draw a simple sphere (yes, I know there's a drawable type available) to learn how to build up simple geometry, by specifying vertices, normals, texture coordinates and indices. I eventually want to create a function that will give me vertices for spheroids that aren't perfect spheres, but follow the same principles. I've been doing okay so far, and used a spherical coordinate transform to get cartesian coordinates. I have vertices, normals, texture coordinates and indices. * The issue I've run into involves where my 2D texture wraps around and meets itself on the sphere: http://i.imgur.com/ftVH2.png Like shown in the picture, there's a visible seam there. There are also shadows where multiple identical vertices are present at the poles. I'd like to know how to get rid of the seam and shadows! I am manually mapping the texture using the simplest 2d texture-sphere, shown here: http://paulbourke.net/texture_colour/texturemap/texture.gif If I use the bare minimum vertices (no duplicates) to define the sphere then the seaming and shadows disappear but I'm unable have the 2D texture wrap all the way around, and see extreme warping at the poles (which makes sense): http://i.stack.imgur.com/l9baA.png Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] A few questions on GraphicsWindowEmbedded
Hiya, I was able to display osg content on Qt's Declarative interface (QtQuick, QML, etc) using the osg::GraphicsWindowEmbedded class. I followed the examples for the FLTK and GLUT toolkits to get this (sort of) working. There are a few of things I don't understand, and would appreciate some clarification on. * What do the parameters passed to the osgViewer::GraphicsWindowEmbedded constructor (x,y,w,h) actually do? I played around with the values and it had no visible effect on my displayed result * What does GraphicsWindowEmbedded do behind the scenes? Since you don't explicitly specify a GraphicsContext for it, does that mean it uses the OpenGL context of the framework you're embedding it into? * I noticed that the viewport is rendered onto the window in absolute coordinates. With QtQuick, when you create an item to display on the screen, it has a child/parent relationship and coordinates are relative to the parent. Moving the parent will also move the child as you'd expect. This doesn't happen with GraphicsWindowEmbedded though, probably because with QtQuick, all the elements are painted onto one 'canvas' (I think), and GraphicsWindowEmbedded just paints straight onto that. * If, instead of having GraphicsWindowEmbedded draw on my screen whenever I call the viewer's frame() method, is it possible to render the viewer camera's output to a texture that I could manually draw onto a quad or something? This way I don't have to specify absolute coordinates for the viewport like I mentioned above. I already tried this by rendering to an osg::Image, then painting it on the screen with Qt, but I have a relatively old computer and osg::Image-readPixels() takes too much time. My results with the osgscreencapture example were also pretty terrible (15-16 fps @ 640x480). For anyone else who needs to do something similar, here's some scrap code you can use to define a QDeclarativeItem that'll show an osg view. http://pastie.org/3478592 (header) http://pastie.org/3478602 (source) Regards, Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] A few questions on GraphicsWindowEmbedded
Hi Robert, Thanks for the thorough response. Simply put, QtQuick is a different way to build up GUIs. It uses a hybrid CSS/JS language called QML to design a UI instead of the widget / forms approach that GUI toolkits usually employ. It is geared specially towards mobile and tablet devices. I'm working on a couple of projects for systems which have a small touch display where typical UI design isn't appropriate. A very specific example is an in car infotainment system. I'd like to use OSG to render map data and display it in on my interface which I've built through QtQuick. Here's an example of what my interface looks like: http://prismaticproject.weebly.com/uploads/8/1/9/7/8197704/1660451_orig.png Unlike normal QGLWidgets, which the GraphicsWindowQt class uses, creating an item in QtQuick to display stuff is slightly different; there are no 'widgets'. You instead create an item, where you get access to a single paint() method where the only OpenGL functionality you have is making direct calls. All QtQuick items are rendered onto a single QGLWidget canvas. For example, I can do stuff like: //glBegin(GL_QUADS); //glColor3ub(0,0,255); //glVertex2d(boundingRect().left(), boundingRect().top()); //... And the item would render in its own assigned space on the screen, maintaining all of its positional parent child relationships. Regards, Preet On Tue, Feb 28, 2012 at 4:54 AM, Robert Osfield robert.osfi...@gmail.com wrote: Hi Preet, On 28 February 2012 09:16, Preet prismatic.proj...@gmail.com wrote: * What do the parameters passed to the osgViewer::GraphicsWindowEmbedded constructor (x,y,w,h) actually do? I played around with the values and it had no visible effect on my displayed result They tell the GraphicsWidnowEmedded object where about the window it's representing is in desktop coordinates. * What does GraphicsWindowEmbedded do behind the scenes? Since you don't explicitly specify a GraphicsContext for it, does that mean it uses the OpenGL context of the framework you're embedding it into? GraphicsWindowEmbedded actually does nothing behind the scenes at all - it's a non op class designed to make it possible to reuse the osgViewer::Viewer and CompositeViewer classes in places that already provide the graphics context and are strictly single threaded and have a single window. It's really tricking these classes that a designed to work with multi-threading and multi-context capable systems where makeCurrent() and swapBuffers() implementations are accessible to work without doing this operations so when calls to makeCurrent() and swapBuffers() are made by the viewer classes nothing happens. Since nothing happens it's entirely up to the calling app to do it for the viewer because it can't do itself. Since GraphicsWindowEmbedded is design to avoid properly implementing the key context management functions on the OSG or a subclass side it does restrict how you can use the viewer classes. It's very much a crude approach to provide viewer functionality, easy to get working yes, fully capable no. You can't add the full capabilities back in when using GraphicsWindowEmbedded, if you don't need multi-threading or multi-context this will be fine, but if you do look elsewhere. Actually implementing a GraphicsWindow properly is the most powerful way to adapt 3rd party graphics contexts with the osgViewer, and this is done by the osgQt::GraphicsWindowQt class, as was as the GraphicsWindowX11, Win32, Cocoa and Carbon implementations found natively in osgViewer. * I noticed that the viewport is rendered onto the window in absolute coordinates. With QtQuick, when you create an item to display on the screen, it has a child/parent relationship and coordinates are relative to the parent. Moving the parent will also move the child as you'd expect. This doesn't happen with GraphicsWindowEmbedded though, probably because with QtQuick, all the elements are painted onto one 'canvas' (I think), and GraphicsWindowEmbedded just paints straight onto that. I know absolutely nothing about QtQuick or how you've implemented the integration with GraphicsWindowEmbedded so I can't say anything about this. * If, instead of having GraphicsWindowEmbedded draw on my screen whenever I call the viewer's frame() method, is it possible to render the viewer camera's output to a texture that I could manually draw onto a quad or something? This way I don't have to specify absolute coordinates for the viewport like I mentioned above. I already tried this by rendering to an osg::Image, then painting it on the screen with Qt, but I have a relatively old computer and osg::Image-readPixels() takes too much time. My results with the osgscreencapture example were also pretty terrible (15-16 fps @ 640x480). If you want to do off screen rendering then this is very different thing, GraphicsWindowEmedded is not the best thing to do this, what you'll want is a pbuffer
Re: [osg-users] Displaying World Coordinate Frame in Relation to Local Coordinate Frame
Hi Scott, It wasn't quite clear to me what you wanted represent with your rocket's second set of axes. Your WCS origin is at the center of the earth (ie, ECEF coordinates). Your rocket also has its own local geometrical axes as your screenshots show. As your rocket orbits the Earth, its ECEF coordinates (XYZ relative to Earth's center) will change. I have no idea if this ASCII art will make any sense, but here goes: timepoint 1 ECEFlocal z z |__x |__x o -- earth || rocket timepoint 2 ECEF z |__x o -- earth local z /__x // rocket at some new position/orientation If you translate the rocket to the center of the earth at any point (so now the COG of the rocket is at (0,0,0)) only the orientation of the rocket can change. Are you looking for the 'difference' between the ECEF's X,Y and Z axes and your rocket's local X,Y,Z axes at any given point in time? If so... then I think you can use linear algebra to solve for the transform that gets you from ECEF to your local axes. ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Rendering an OpenSceneGraph scene in Qt QuickRendering an OpenSceneGraph scene in QML Rendering an OpenSceneGraph scene in QML
Hiya, I looked at the osgQt/GraphicsWindowQt class implementation. I think what it does is render to Qt's QGLWidget's OpenGL context using swapBuffers(). Again... being new to OpenGL, a lot of this stuff doesn't really make too much sense to me, but the swapBuffers() call is QGLWidget specific. There's no equivalent in the QDeclarative API, so I couldn't find a way of exposing QGLWidget to QtQuick directly as a QDeclarative item. From what I've read online, I think that the best that can be done is creating a custom QDeclarativeItem, and overriding its paint() method, where direct OpenGL calls can be made. I tested this out and it works... I can, for instance, paint geometry using glBegin() and glEnd() exactly as you'd expect, and it'll show up in a QML object. So I think I need to use OSG to render my scene as a texture to an FBO, then draw that using straight OpenGL calls. This is where I'm totally lost though. So if I had to rephrase the problem: Assuming I had a viewport I could only paint to using native OpenGL calls, how could I render an OpenSceneGraph scene in it? Regards, Preet On Fri, Feb 17, 2012 at 4:27 AM, Philipp Moeller philipp.moel...@geometryfactory.com wrote: Preet prismatic.proj...@gmail.com writes: Hey all. I have an OpenSceneGraph tree I'd like to have displayed in QtQuick. I'd like the user to be able to interact with the scene where the mouse can zoom/pan/rotate the camera, just like osgViewer allows. I saw that there were a couple of examples (osgQtWidgets and osgQtBrowser), but they're a little over my head. I think the right way to start trying this is by starting with the QtDeclarative model (http://doc.qt.nokia.com/4.7-snapshot/qtdeclarative.html)... but I'm not too sure where to go from here. Does anyone have any suggestions? Regards, Pris Hi Preet, I've looked into integrating OpenSceneGraph into QML. You might want to start with the osgviewerQt to get started with a viewer integrated into Qt. Then you should have a look at [1] to see how you can expose your osg widget to QML. If I ever get something reusable up and running, I'll let the list know. If you go down the route of exposing scene graph functionality in your Qt application you will need to roll some implementation of QAbstractItemModel, which is a real pain. It get's easier if you restrain yourself to graphs where each element only has a single parent. If you can throw something together I would be really interested in it, so let me know. [1] : http://developer.qt.nokia.com/doc/qt-4.8/gettingstartedqml.html#exposing-c-classes-to-qml -- Philipp Moeller GeometryFactory ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Rendering an OpenSceneGraph scene in Qt QuickRendering an OpenSceneGraph scene in QML Rendering an OpenSceneGraph scene in QML
Hey all. I have an OpenSceneGraph tree I'd like to have displayed in QtQuick. I'd like the user to be able to interact with the scene where the mouse can zoom/pan/rotate the camera, just like osgViewer allows. I saw that there were a couple of examples (osgQtWidgets and osgQtBrowser), but they're a little over my head. I think the right way to start trying this is by starting with the QtDeclarative model (http://doc.qt.nokia.com/4.7-snapshot/qtdeclarative.html)... but I'm not too sure where to go from here. Does anyone have any suggestions? Regards, Pris ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Unable to get osgViewerQt to work
Hiya, I've been trying to run the osgViewerQt example. I have osg (including osgQt) built from git. I'm able to use osgviewer without any issues. When I try to run osgViewerQt, here's the output: [preet@icarus osgviewerQt]$ ./osgViewerQt cessna.osg [xcb] Unknown request in queue while dequeuing [xcb] Most likely this is a multi-threaded client and XInitThreads has not been called [xcb] Aborting, sorry about that. osgViewerQt: xcb_io.c:178: dequeue_pending_request: Assertion `!xcb_xlib_unknown_req_in_deq' failed. Aborted I'm using libxcb-1.8-2 with Qt 4.8 on Arch Linux (kernel v 3.2.5). The Qt install should be fine as I can run other Qt applications without any issues. I'd appreciate any advice... Preet ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
Re: [osg-users] Drawing poly line, jitter/vibration issue when moving camera
Hey Robert,
Thanks for the detailed response. I had a few follow up questions.
On Tue, Feb 14, 2012 at 7:39 AM, Robert Osfield
robert.osfi...@gmail.com wrote:
Hi Preet,
Rui was a little cryptic is his explanation so I'll expand on it
further. What you are seeing is numerical precisions issues down on
the graphics card as internally the graphics card uses floats for
vertices and floats for the modelview matrix.
The solution to this problem is make sure all geometry data has a
local origin and use a transform node (such as an
osg::MatrixTransform) that decorates your geometry and places it in
world coordinates.
I didn't understand what was meant by local origin here. Let's say I
have a model of the Sphinx I want to show on the surface of the Earth
(coincidentally I'm trying to display geospatial data).
Are you suggesting that I draw the model with coordinates relative to
the origin of of the scene first ('local'), then parent the
corresponding geode with a matrix transform that scales/moves/rotates
it out to the correct position on the Earth's surface ('world
coordinates')?
I found a few references online to a somewhat similar solution that
suggested transforming the entire scene so it's centered at the origin
before passing it to OpenGL. Is this a worse or better approach than
the suggested fix?
The OSG by default uses double's for the Camera's
View matrix and double's for the transform nodes and when accumulating
all the matrices during the cull traversal uses doubles as well so
that by the time the OSG maintains precision all the way till it
passes the modelview matrix to OpenGL. Typically when you are close
to your geometry the translation component of Camera's View matrix
will cancel out with the translation component of your scene graph
transform node, so that you end up with a final modelview passed to
OpenGL with modest translation component that doesn't present any
numerical precision issues, solving the precision.
I understand that numerical precision decreases when rendering a scene
further from the origin, because there are a finite set of numbers you
can represent in floating point. Using doubles for the transform and
carrying doubles throughout the scene graph traversals increases
precision right up until the data is passed to OpenGL. I didn't quite
follow your explanation about the camera and geometry translations
cancelling out though... could you expand on this a bit?
This approach is commonly used when tackling whole earth geospatial
data, and you'll see it in action in the paged databases that
VirtualPlanetBuilder creates, 3rd party NodeKit's like osgEarth use
this approach as well.
I'll take a look at the source for both of these projects, thanks :)
Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
[osg-users] Drawing poly line, jitter/vibration issue when moving camera
Hiya, I'm trying to draw a simple poly line from a vertex array with the GL_LINE_STRIP primitive. When I view the object using osgviewer, it looks like everything is alright. Once I move the camera around with the mouse however, the lines begin to jitter, like the polyline was made of jello (honestly can't find better description). The effect is especially visible at higher zooms. I'm using large (values in the hundred-millions), high precision numbers (doubles) for the vertices of the line strip, and I think that's what's causing the problem (I couldn't recreate the issue using smaller, more sane numbers from 1-500 for example, even though the same double datatype is used to store both numbers). I've posted my code here: http://pastie.org/3377862 The code only relies on STL and osg, so it should 'just compile' if you have an osg environment setup. The relevant calls to osg are on lines 121-141. I'm using OSG 3.0.1 on Arch Linux, kernel 3.2.x if it's relevant. Does anyone have any ideas as to what's going on? Regards, Pris ___ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
