Rotation matrices were replaced with quaternions but incorrect contructor was used. There was a call to Quat(angle, xAxis, yAxis, zAxis ) but there is no such constructor in Quat class to create quaternion for rotation. As a result we got this values being written into quaternion directly.
I've replaced Quat contructor calls with the ones creating rotation quaternions Quat( angle, Vec3( axis ) )."
unsigned int Image::computeNumComponents(GLenum pixelFormat)
so I added these types to the switch statement:
case(GL_RED_INTEGER_EXT): return 1;
case(GL_GREEN_INTEGER_EXT): return 1;
case(GL_BLUE_INTEGER_EXT): return 1;
case(GL_ALPHA_INTEGER_EXT): return 1;
case(GL_RGB_INTEGER_EXT): return 3;
case(GL_RGBA_INTEGER_EXT): return 4;
case(GL_BGR_INTEGER_EXT): return 3;
case(GL_BGRA_INTEGER_EXT): return 4;
case(GL_LUMINANCE_INTEGER_EXT): return 1;
case(GL_LUMINANCE_ALPHA_INTEGER_EXT): return 2;
That's all... now it computes the number of components and, thus, the image size
correctly."
Source and CMake files are:
CMakeLists.txt
ReaderWriterBVH.cpp
Also there are 3 example BVH files. The first two are captured from motions of human beings - maybe a kung-fu master here. PLEASE use command below to see the results:
# osgviewer example1.bvh -O solids
This will demonstrate the animating of a skeleton and render bones as solid boxes. Note that the motion assumes XOZ is the ground and has an offset from the center, so we should adjust our view to get best effects.
You may also use "-O contours" to render bones as lines. The viewer shows nothing if without any options because osgAnimation::Bone does not render itself. User may add customized models to each named bones as osganimationskinning does to make uses of this plugin in their own applications.
I was wondering to support a BvhNode in my osgModeling peoject before, but soon found it better be a plugin for animation. A problem is, how to bind real geometry models to the skeleton. Maybe we could have a bindingToNode() visitor in future to find geodes matching names of bones and add them as bones' children."
--------------------------------------------
Descripton:
The patch does provide a new class PixelDataBufferObject which is capable of allocating memory on the GPU side (PBO memory) of arbitrary size. The memory can then further be used to be enabled into read mode (GL_PIXEL_UNPACK_BUFFER_ARB) or in write mode (GL_PIXEL_PACK_BUFFER_ARB). Enabling the buffer into write mode will force the driver to write data from bounded textures into that buffer (i.e. glGetTexImage). Using buffer in read mode give you the possibility to read data from the buffer into a texture with e.g. glTexSubImage or other instuctions. Hence no data is copied over the CPU (host memory), all the operations are done in the GPU memory.
--------------------------------------------
Compatibility:
The new class require the unbindBuffer method from the base class BufferObject to be virtual, which shouldn't break any functionality of already existing classes. Except of this the new class is fully orthogonal to existing one, hence can be safely added into already existing osg system.
--------------------------------------------
Testing:
The new class was tested in the current svn version of osgPPU. I am using the new class to copy data from textures into the PBO and hence provide them to CUDA kernels. Also reading the results back from CUDA is implemented using the provided patch. The given patch gives a possibility of easy interoperability between CUDA and osg (osgPPU ;) )
--------------------------------------------
I think in general it is a better way to derive the PixelBufferObject class from PixelDataBufferObject, since the second one is a generalization of the first one. However this could break the current functionality, hence I haven't implemented it in such a way. However I would push that on a stack of wished osg 3.x features, since this will reflect the OpenGL PBO functionality through the classes better.
"
"To reproduce the bug:
1. Create a template osg::Sequence node (and underlying geometry) but do not attach the node to the current active scenegraph.
2. At some point during the rendering loop (perhaps on a keystroke) clone the sequence node (I use the call:
dynamic_cast<osg::Node*>(templateNode -> clone( osg::CopyOp( (osg::CopyOp::CopyFlags)osg::CopyOp::DEEP_COPY_NODES ) ) )
3. Set the cloned sequence node duration to a value that makes the animation run slower (i.e. 2.0).
4. Start the cloned sequence (using setMode()).
5. Repeat steps 2 \u2013 4 and observe that the cloned sequences do not run slow but run as fast, appearing to ignore the duration that has been set on them.
Looking at the \u2018good documentation\u2019 (2.4 source code), I see that _start is being set to _now (osg::Sequence::setMode(), line 192). Should this not _start not be set to -1.0?"
The plug-in is a wrapper around open-exr (http://www.openexr.com) that consists of two projects, ilmbase-1.0.1 and openexr-1.6.1.
I have only tested it on windows XP 32 machine. So there might be some work making it work on other platforms.
The plug-in supports writing and reading EXR files. When writing it can use the data type GL_HALF_FLOAT_ARB(se ilmbase-1.0.1) and GL_FLOAT. When reading the data type always becomes GL_HALF_FLOAT_ARB. It supports textures with three and four channels.
When reading an exr file it automatically removes Alfa channel if it didn't store any information."
--
From Robert Osfield, started work on ported it to other platforms, but could fix some problems relating to error:
?Imf::OStream::OStream(const Imf::OStream&)? is private
I'm checking in now so that others can have a bash at completing the port.
Standard is to generate one stl file.
With an additional option it is possible to write one file per Geode. This option is not very "useful" for typical application, I use it for separating and conversion of geometric data. So it could be removed if considered to special."
* When used PDS RenderStage::runCameraSetUp sets flag that FBO has already stencil,depth buffer attached. Prevents adding next depth buffer.
* Sets correct traits for p-buffer if used PDS and something goes wrong with FBO setup or p-buffer is used directly.
* Adds warning to camera if user add depth/stencil already attached through PDS.
* Sets blitMask when use blit to resolve buffer.
There is also new example with using multisampled FBO."
I have not reverted added Compiler options. I assume that one may want to have warnings enabled for the application but may not want to see them while OSG libraries and examples compile.
Modified files:
osg/Export - now explicitly includes osg/Config to make sure OSG_DISABLE_MSVC_WARNINGS is read
osg/Config.in - declares OSG_DISABLE_MSVC_WARNINGS flag to be added to autogenerated osg/Config
CMakeLists.txt - declares OSG_DISABLE_MSVC_WARNINGS as option with default ON setting
"
New features
+Read and write of osg::LOD, osg::Switch, osgSim::Sequence, osgim::MultiSwitch and osgSim::DOFTransform data in <extra>
+Read and write of osg::Node description data in <extra>
+Plugin option "NoExtras" to prevent writing of <extra> data and only traverse the active children when saving
Changes/additions
+instanced_geometry and instanced_controller are now loaded in a single Geode with multiple Geometries instead of multiple geodes with a single Geometry
+Changed all calls to the deprecated createAndPlace() to the new add() methods
+All transformation elements <scale>, <rotate>, <translate>, <lookat>, <matrix>, <skew> are now concatenated properly in to a single MatrixTransform.
Previously this was not done in order as required by Collada and and not all elements were included.
+Complete skew matrix creation
+Automatically add GL_RESCALE_NORMAL if scale is non-identity
+Blinn shininess remapping to [0,128] when in range [0,1]
+Changes to CMake file to make it compile on Windows
+Coding style and code documentation
Bug fixes
+Transparent texture writing fixed
+Fixed bug in using osg node name as collada node ID
+Fixed usage of double sided faces in GOOGLEEARTH extra
+Not adding blendfunc and blendcolor when opaque
TODO/Wishlist
-solve differences in drawables, DAE reader should place multiple collation elements into multiple primitivesets in a single geometry where possible (only when same material)
-solve differences in matrices
-multitexture support
-skinned mesh and generic animations using osgAnimation
-profile_GLSL based on COLLADA OpenGL Effects Viewer http://ati.amd.com/developer/rendermonkey/downloads.html
-handling more <extra> to more closely mimic the intended lighting"
The set up for this to work is a bit more complex than most files, since the engine expects all files to be in a certain place, and it tends to mix case a lot. I tried to explain everything in the VBSP_README.txt file."
This plugin has been integrated with the pre-exisiting bsp plugin.
After taking a look at the current state of the BMP loader I decided it might be worth a shot at reimplementing that part. For example: the current loader doesn't properly handle 1- and 4-bit files, incorrectly loads 16-bit files as intensity-alpha (they are RGB555), is full of dead code, and generally not in very good shape.
Attached is my re-implementation for review.
I've checked it against the test images from http://wvnvaxa.wvnet.edu/vmswww/bmp.html and models that use BMP files.
"
Before:
1. if the camera is not a slave camera
1.1 set the eventState's graphics context to the current context.
2. if the current master view is not the view which has the focus
2.1 set the current master view to be the view which has the focus
2.2 use the new master view's eventState instead of the old one
Now as you can see from this sequence, the graphics context is set on the eventState before switching to the view which has focus (and thus using another eventState). So the new eventState, in the case we need to switch views, will contain an old graphics context, not the correct one.
Just inversing these steps fixes the problem:
1. if the current master view is not the view which has the focus
1.1 set the current master view to be the view which has the focus
1.2 use the new master view's eventState instead of the old one
2. if the camera is not a slave camera
2.1 set the eventState's graphics context to the current context.
Now, the eventState will refer to the correct graphics context in both cases.
Attached is a fixed CompositeViewer.cpp (based on today's SVN) which does this. Note that some other things are done in the 1. and 2. cases, but they have no influence on each other so they can just be swapped without problems.
"
I suggest moving the definition of OSG_DEFAULT_LIBRARY_PATH (which is responsible for this global rebuild) into osgDB/CMakeLists.txt which is the only library in the code where this definition is (and is likely ever to be) used. This way if the user changes it, only osgDB will rebuild."
set to off. But could be activated/decativated via CMake as well as system
environment variable. I also modified src\osgViewer\CMakeLists.txt to turn
off this workaround by default as suggested."
parts of their scene graphs. Within these common part some nodes have
event handlers which use the action adapter argument to the event
handler to determinate which view received the event.
Here is the problem, osgViewer::CompositeViewer::eventTraversal sets
the action adapter field in the EventVisitor always to the last view
which received an event, instead of using the view which actually
received the event, so determination of the correct view does not work.
I looked at the code a bit, and moved the code for setting the action
adapter to a IMO better place"
facilitate the subclassing of Image providing interactive behaviours so as
used in the vnc interactive VncImage class.
osgViewer::InteractiveImageHandler provides an event handler that convertes osgGA
mouse and keyboard events into the coordinate frame of an image based on ray intersection with geometry in
the associated subgraph.
Changed the ordering of events processing in Viewer and CompositeViewer to allow
scene graph event handlers to take precidence over viewer event handlers and camera manipulators
---------------------------
function FltExportVisitor::writeExternalReference( const osg::ProxyNode& proxy ):
Line 423 in file expPrimaryRecords.cpp has to be changed from
const ParentPools* pp = static_cast<const ParentPools*>(proxy.getUserData() );
to
const ParentPools* pp = dynamic_cast<const ParentPools*>(proxy.getUserData() );
"
The semantic change that went into the ac loader with the past patch was
incorrect wrt the document describing the behaviour of ac files and
inconsistent with what ac3d itself displays for that files.
This attached change reverts the behaviour to the original one. The
infrastructure to change this is left in place.
The change is based on rev 9045."
"
and a later post the same osg-submissions thread:
"it's been a while since I have made the changes but I think it was due to problems with static builds of OpenThreads on windows. I was using
OpenThreads in a communication/synchronisation library (without
OpenSceneGraph). It seems I forgot to post a small change in the CMakeLists file of OpenThreads. If a user turns DYNAMIC_OPENTHREADS to OFF (static build) OT_LIBRARY_STATIC will be defined in the Config.
Without these changes a windows user will always end up with a "__declspec(dllexport)" or "__declspec(dllimport)" which is a problem for static builds."
And another post from Blasius on this topic:
"I tested with VS2005 and VS2008. For 32 bit everything works as expected. For x64 and VS2008 I could successfully do the cmake-configure and then the compilation but I had occasional crashes of cmTryCompileExec.exe (during the cmake-configure phase) which seems to be a cmake bug. With VS2005 and 64bit cmake does not set _OPENTHREADS_ATOMIC_USE_WIN32_INTERLOCKED although the interlocked functionality should be there. If I place the source snippet from the CHECK_CXX_SOURCE_RUNS macro to a separate sourcefile I can compile and run the resulting executable successfully. Forcing OPENTHREADS_ATOMIC_USE_WIN32_INTERLOCKED (on VS2005/x64) reveals a bug in "intrin.h" which seems to be fixed in VS2008 but not in VS2005.
In case anyone is interested the lines:
__MACHINEI(unsigned char _interlockedbittestandset(long *a, long b))
__MACHINEI(unsigned char _interlockedbittestandreset(long *a, long b))
__MACHINEX64(unsigned char _interlockedbittestandset64(__int64 *a, __int64 b))
__MACHINEX64(unsigned char _interlockedbittestandreset64(__int64 *a, __int64 b))
should be changed to:
__MACHINEI(unsigned char _interlockedbittestandset(long volatile *a, long b))
__MACHINEI(unsigned char _interlockedbittestandreset(long volatile *a, long b))
__MACHINEX64(unsigned char _interlockedbittestandset64(__int64 volatile *a, __int64 b))
__MACHINEX64(unsigned char _interlockedbittestandreset64(__int64 volatile *a, __int64 b))
The worst thing that can happen is that interlocked funtionality is not detected during cmake-configure and the mutex fallback is used.
Which reminds me another small glitch in the Atomic header so I attached a corrected version.
Why is the OT_LIBRARY_STATIC added to the config file? It is not needed anywhere.
OT_LIBRARY_STATIC is needed if you are doing static-builds on Windows. See my previous post on that.
"
- Matrix transform nodes were stripped out, and the
vertices and normals of each node were
transposed by the modelling matrix. My change preserves
the matrix transform nodes, so that
models can still be articulated by changing the matrices.
- Lights were copied from the COIN scenegraph to the OSG
scenegraph, but they were not associated
with a LightSource node. My change for this creates a
Group and adds a LightSource for each Light.
- If VRML textures have names, the name is now copied to
--
Form Robert Osfild, changed SbString to std::string usage
M Inventor/ConvertFromInventor.h
M Inventor/ConvertFromInventor.cpp
a capping the number of PagedLOD to a sepcified maximum, with pruning of inactive
PagedLOD when the total number of inactive and active PagedLOD goes above the maximum.
To enable the mode set the env var OSG_MAX_PAGEDLOD to a value something like 1000.
support. While testing the pluggin we found a small bug and we are
submitting a patch.
The first attachment is a small program that creates a scene with two
pyramids, transformed by two instances of
osg::PositionAttitudeTransform. One of them is rotated 90 degrees in
the X axis, and the scene is exported both in the native OSG (.osg)
and COLLADA (.dae) formats. In the first case the rotated pyramid is
displayed correctly whereas in the second the pyramid seems not to be
rotated.
In the COLLADA 1.4.1 specification (found at
http://www.khronos.org/collada/) it is specified that
"The <rotate> element contains a list of four floating-point values
[...] followed by an angle in degrees" but the plugin seems to write
the value in radians.
The problem seems to be in the method daeWriter::apply() that seems to
be writing the angle value in radians to the COLLADA file. The patch
can be found in the second attachment and is simply a call to
RadiansToDegrees wrapped around the angle.
"
It assumes that inverse(_inverseMatrix * _offset) = _offset * _matrix
This is only true when _offset=identity, so I think it should be inverse(_offset) * _matrix
This inconsistency can cause problems when switching from UFO to other manipulators, because the UFO:Manipulator::getMatrix function is not necessarily returning a correct inverse of the currently set ModelViewTransform.
It was tempting to change the name of the _offset member to _inverseOffset, or to maintain both variables, but in the end went for the minimal change."
If a 400 level error occurs, a FILE_NOT_FOUND ReadResult is appropriate.
If a 500 level error occurs (such a 503, Service unavailable), the application might want to try to load the file again in a few seconds/minutes. This submission returns ERROR_IN_READING_FILE if a 500 level error occurs so that clients can easily distinguish between the errors.
The actual error code is also added to the "message" of the ReadResult so if a client needs more information, they can just parse the message to retrieve the error code."
The graph is displayed "under" (behind) the normal bar chart you get when you press 's' twice. It doesn't hide the normal stats, you can still read them without any trouble, and that way, it doesn't take any more screen space. It starts from the left, and will scroll left when there is enough data to fill the screen width. The graph lines have the same colors we're used to (except I made the event color a bit bluer, so it's not exactly the same as the update color). A screen shot is attached.
The lines get a bit confused when they're all overlapping at the bottom of the graph, but I think that's the least of our concerns (if they're all at the bottom of the graph - except FPS of course - then great!).
The only thing I'm not very keen about is that to make things simple, I clamp the values to a given maximum. Right now, the maximums I have set are:
* Frame rate: 100 fps (people have 60, 75, 85Hz refresh rates, so there's no one right value, but I think 100 is OK)
* Stats: 0.016 seconds (what you need to get 60Hz minimum)
This could be changed so that the scale of the graph changes according to the maximum value in the last screenful of the graph instead of clamping values. We would then need to display the scale for each value on the side of the graph, because if the scale changes, you need to know what it is at this moment.
I tried to make things easy to change, so for example if you don't like that the graph is in the same space as the normal stats bars, it's easy to move it anywhere else, and make it have other dimensions. The maximums and colors are also easy to change.
The impact on performance should be minimal, since it's one vertex per graph line that's added per frame, and vertices are removed when they scroll off the screen, so you'll never have more than say 1280 * (3 + ncameras) vertices on the screen at one time. No polygons, I used line strips. The scrolling is done with a MatrixTransform."
Quicktime supports only files with 3/4-channels rgba-files and not 1/2-channels rgb-files.
This submission is from Tatsuhiro Nishioka, here's his original quote:
When FlightGear crashes, the error message
"GraphicsImportGetNaturalBounds failed" shows up. By adding printf
debug, I found the error was -8969: codecBadDataErr when loading a
gray-scaled (2 channels) rgba files even though the file can be loaded
with Gimp and osgViewer properly.
So I made an investigation on this problem and found an interesting
thing. This error occurs only when non-rgb files are loaded before rgb
files. The reason is that rgba files can be handled by both
osgdb_rgb.so and osgdb_qt.so, but the error happens only when
osgdb_qt.so try to load a gray-scaled rgba file.
When a program is about to load an rgba file, osgdb_rgb.so is loaded
and it handles the rgba file properly. In contrast, when a gray-scaled
rgb file is being loaded after a non-rgb file (say png) is already
loaded by osgdb_qt.so, osgdb_qt.so tries to load the file instead of
osgdb_rgb, which causes the error above.
Anyway, the bad thing is that QuickTime cannot handle gray-scaled rgb
files properly. The solution for this is not to let osgdb_qt handle
rgb files since osgdb_rgb can handle these properly.
"
This code will add two extra statistics options:
-Camera scene statistics, stats for the scene after culling (updated at 10 Hz)
-View scene statistics, stats for the complete scene (updated at 5 Hz)
Each camera and each view will expand the statistics to the right.
I also added the requests and objects to compile of the databasepager to the databasepager statistics.""
Attached is a change that is able to provide shared textures for the clamp and
the repeat case.
So this appears to be the best fix I guess ...
Also it additionaly shares the TexEnv StateAttribute in a whole ac3d model."
