[osg-users] create glass effect
Hi all, how can I create an effect that seems to the glass? thank you. Davide ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
R: RE: [osg-users] create glass effect
thank you, can you tell me how can I load theese shaders? I'm able to load only the shaders .frag and .vert used in osg examples. I must download the HLSL or the CgFX shader? Thank you Davide Messaggio originale Da: [EMAIL PROTECTED] Data: 18-lug-2007 15.30 A: osg usersosg-users@openscenegraph.net Ogg: RE: [osg-users] create glass effect Hello Davide, You can check out the Nvidia Shader Library for examples: http://developer.download.nvidia. com/shaderlibrary/webpages/shader_library.html bill -Original Message- From: [EMAIL PROTECTED] on behalf of [EMAIL PROTECTED] Sent: Wed 7/18/2007 9:03 AM To: [EMAIL PROTECTED] net Subject: [osg-users] create glass effect Hi all, how can I create an effect that seems to the glass? thank you. Davide ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph. net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph. net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
R: Re: [osg-users] problem applying textures to .flt models
Sorry, but I don't understand where I'm wrong.
I don't have a osg::
Geometry, I only have an osg::Node and I can't set the texture
coordiantes ...
Thank you
Davide
Messaggio originale
Da:
[EMAIL PROTECTED]
Data: 5-lug-2007 10.07
A: osg usersosg-
[EMAIL PROTECTED]
Ogg: Re: [osg-users] problem applying
textures to .flt models
Hi,
You need texture coordinates.
Regards,
Brede
On 7/4/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote:
Hi all, I can
apply textures to geometry, shapes and models .3ds, but
not to models
.flt.
The code I'm using is:
// set up the texture
state.
osg::Texture2D* texture = new osg::Texture2D;
texture-
setDataVariance(osg::Object::DYNAMIC); // protect from being
optimized away as static state.
osg::Image* textureImage =
osgDB::
readImageFile((LPCTSTR)texturePath);
if(!
textureImage){
AfxMessageBox(Impssibile caricare la texture);
return;
}
//
Assign the texture
to the image we read from file:
texture-setImage
(textureImage);
osg::StateSet* stateset = node-
getOrCreateStateSet();
stateset-setTextureAttributeAndModes
(0,
texture,osg::StateAttribute::ON);
stateset-setMode(
GL_LIGHTING, osg::
StateAttribute::OFF );
node-
setStateSet(stateset);
How can I do?
Thank you.
___
osg-users mailing
list
osg-users@openscenegraph.net
http://openscenegraph.
net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.
net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
[osg-users] problem applying textures to .flt models
Hi all, I can apply textures to geometry, shapes and models .3ds, but
not to models .flt.
The code I'm using is:
// set up the texture
state.
osg::Texture2D* texture = new osg::Texture2D;
texture-setDataVariance(osg::Object::DYNAMIC); // protect from being
optimized away as static state.
osg::Image* textureImage = osgDB::
readImageFile((LPCTSTR)texturePath);
if(!textureImage){
AfxMessageBox(Impssibile caricare la texture);
return;
}
//
Assign the texture to the image we read from file:
texture-setImage
(textureImage);
osg::StateSet* stateset = node-
getOrCreateStateSet();
stateset-setTextureAttributeAndModes(0,
texture,osg::StateAttribute::ON);
stateset-setMode( GL_LIGHTING, osg::
StateAttribute::OFF );
node-setStateSet(stateset);
How can I do?
Thank you.
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
[osg-users] CEGUI
Hi, I tried to run the osgcegui example, and no GUI was rendered. All the GUI creation code was taken from CEGUI example (which works). I didn't get any errors from CEGUI or OSG. I also tried to set the renderGUI function as post callback of the camera with no better results. any ideas? is it something with the Zbuffer? or Z position of the GUI? thanks, Guy. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Loading shader .fx
Hi, how can I load shaders with .fx extension? Thank you Davide ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
R: Re: [osg-users] Loading shader .fx
I'm non 100% sure, but I think .fx shader language belongs to Microsoft. I have some shader .fx that are very realistic and wonderful and I would like to apply them to some parts of my scene... Messaggio originale Da: [EMAIL PROTECTED] Data: 2-lug-2007 12.27 A: osg usersosg-users@openscenegraph.net Ogg: Re: [osg-users] Loading shader .fx On 7/2/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi, how can I load shaders with .fx extension? The OSG core doesn't support .fx shaders. I'm not 100% sure what .fx shaders exactly entail so I did a quick search on .fx shaders on the net and got lots of links that aren't particularly useful. Is it part of Nidia's Cg, part of Direct3D? Robert. ___ osg- users mailing list osg-users@openscenegraph.net http://openscenegraph. net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] viewer on two different workstation
Hello, in my application I have a CompositeViewer with two osgViewer:: View, it is possible display one View on a workstation and the other View on another workstation connected in the same LAN? How can I do this? thank you and sorry for my bad English :) ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] camera in camera and COMPUTE_NEAR_FAR_USING_BOUNDING_VOLUMES Problem
hi,all I use camera in camera in osg2.0 under windowsxp. |cameraB root-- | |cameraC The cameraC uses relate cf. The main camera(viewer.getCamera())'s near_far_mode uses COMPUTE_NEAR_FAR_USING_BOUNDING_VOLUMES. The cameraC's near_far_mode and viewport are not assigned explicitly. Th cameraC's viewMatrix and projectMatrix is set to E. The Pseudo code: osgViewer::Viewer viewer; viewer.setSceneData(root); ... osg::NodePath np = cameraC-getParentalNodePaths()[0]; bool result = viewer.computeIntersections(x,y,np, ...); I found the result = false sometime. I debuged it found that because COMPUTE_NEAR_FAR_USING_BOUNDING_VOLUMES, the projection matrix was auto changed when the eyepoint was changed. How can I do right computeIntersections with camera in camera ? Or I missing somthing? Or I should use slave camera , not camera in camera? ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Reading from texture
- Original Message - From: Robert Osfield [EMAIL PROTECTED] To: osg users osg-users@openscenegraph.net Sent: Tuesday, June 26, 2007 3:28 PM Subject: Re: [osg-users] Reading from texture On 6/26/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Thanks for the advice, After implementing it, I encountered 3 problems: 1. There are minor differences of the data values between the two implementations (reading from the current texture, and attaching image) maybe images doesn't support half float precision? There is some half float support in the 2.0 version of osg::Image, but this might not coverage your usage. I'm still working with OSG1.2. currently I don't have time for the porting (even that I played a little with OSG2.0 on my free time :)). anyway, I will try to compile the OSG2.0 image code with OSG1.2 and see what happens. 2. The image was rescaled to power of 2. (300x300 --- 256x256) while textures sizes don't have to be power of two. (is there a way to define image that won't be a power of 2 and attach it to the FBO?) What makes you think the images are rescaled to a power of two when reading? There isn't any such code. I don't remember the exact words but OSG prompts in the console something like rescaling (300x300) image to (256x256). There only code for rescaling to power of two happens when applying on non power of two image to a power of two texture. I allocated both at the same code section using the same width and height. Note the OSG support non power of two texture where supported by OpenGL drivers, but by default this is switched off due to portability problems related to OpenGL driver saying non power of two is supported, but then implementing it in software... See osg::Texture::setResizeNonPowerOfTwoHint(bool); I'll look for it. 3. It run slower than binding the texture object and reading from current texture. It shouldn't run slower, if it is then there is possible some issue along the way such as pixel formats being inconsistent. Both are using GL_RGBA mode with GL_FLOAT as the type. The texture has internal format of half float, I don't seem to remember setting the image to half float but I did what I saw in one of the examples that show how to attach image to texture and having the texture automaticlly updated when the image is changed. Could that be the problem? is it possible to set image internalFormat? did you mean something else by pixel format? Robert. thanks, Guy. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Reading from texture
thanks I'll get on to it as soon as possible. btw, the setResizeNonPowerOfTwoHint(false) solved the rescaling. thanks, Guy. - Original Message - From: Robert Osfield [EMAIL PROTECTED] To: osg users osg-users@openscenegraph.net Sent: Thursday, June 28, 2007 9:55 AM Subject: Re: [osg-users] Reading from texture Hi, OSG-2.0 does support a wider range of pixel formats so rather than chase ghosts it would probably be best just upgrading. Porting to 2.0 should not be that difficult, if you use osgProducer there is the library still existing as its own separate project and works fine with 2.0. The core OSG has seen a number of changes, but most shouldn't effect backwards compatibility too much. Robert. On 6/28/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: - Original Message - From: Robert Osfield [EMAIL PROTECTED] To: osg users osg-users@openscenegraph.net Sent: Tuesday, June 26, 2007 3:28 PM Subject: Re: [osg-users] Reading from texture On 6/26/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Thanks for the advice, After implementing it, I encountered 3 problems: 1. There are minor differences of the data values between the two implementations (reading from the current texture, and attaching image) maybe images doesn't support half float precision? There is some half float support in the 2.0 version of osg::Image, but this might not coverage your usage. I'm still working with OSG1.2. currently I don't have time for the porting (even that I played a little with OSG2.0 on my free time :)). anyway, I will try to compile the OSG2.0 image code with OSG1.2 and see what happens. 2. The image was rescaled to power of 2. (300x300 --- 256x256) while textures sizes don't have to be power of two. (is there a way to define image that won't be a power of 2 and attach it to the FBO?) What makes you think the images are rescaled to a power of two when reading? There isn't any such code. I don't remember the exact words but OSG prompts in the console something like rescaling (300x300) image to (256x256). There only code for rescaling to power of two happens when applying on non power of two image to a power of two texture. I allocated both at the same code section using the same width and height. Note the OSG support non power of two texture where supported by OpenGL drivers, but by default this is switched off due to portability problems related to OpenGL driver saying non power of two is supported, but then implementing it in software... See osg::Texture::setResizeNonPowerOfTwoHint(bool); I'll look for it. 3. It run slower than binding the texture object and reading from current texture. It shouldn't run slower, if it is then there is possible some issue along the way such as pixel formats being inconsistent. Both are using GL_RGBA mode with GL_FLOAT as the type. The texture has internal format of half float, I don't seem to remember setting the image to half float but I did what I saw in one of the examples that show how to attach image to texture and having the texture automaticlly updated when the image is changed. Could that be the problem? is it possible to set image internalFormat? did you mean something else by pixel format? Robert. thanks, Guy. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Reading from texture
Hello all, I've asked this before but didn't got replies so I bring the issue up again. I'm having in my application more than one FBO and after the frame was rendered I want selectively read one of them. I plan to use Image::readFromCurrentTexture to read the data from the texture attached to the FBO. The question is, if there are two textures attached to two different FBOs, how do I know which one is bound in the context? how do I read it, or the other image? or both? I want to prevent expilcitly calling to bind since it appears to be time consuming, and people here say it's unnecessary. any information on the subject would be helpfull, thanks, Guy. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Reading from texture
Thanks for the advice, After implementing it, I encountered 3 problems: 1. There are minor differences of the data values between the two implementations (reading from the current texture, and attaching image) maybe images doesn't support half float precision? 2. The image was rescaled to power of 2. (300x300 --- 256x256) while textures sizes don't have to be power of two. (is there a way to define image that won't be a power of 2 and attach it to the FBO?) 3. It run slower than binding the texture object and reading from current texture. if there are other ideas, I'll be glad to hear. thanks, Guy. - Original Message - From: Robert Osfield [EMAIL PROTECTED] To: osg users osg-users@openscenegraph.net Sent: Tuesday, June 26, 2007 11:34 AM Subject: Re: [osg-users] Reading from texture Why not just attach osg::Image to the FBO Camera in question and let the OSG automatically copy the image from the FBO after its written to. On 6/26/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hello all, I've asked this before but didn't got replies so I bring the issue up again. I'm having in my application more than one FBO and after the frame was rendered I want selectively read one of them. I plan to use Image::readFromCurrentTexture to read the data from the texture attached to the FBO. The question is, if there are two textures attached to two different FBOs, how do I know which one is bound in the context? how do I read it, or the other image? or both? I want to prevent expilcitly calling to bind since it appears to be time consuming, and people here say it's unnecessary. any information on the subject would be helpfull, thanks, Guy. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: Re: [osg-users] FBO Multipass shading
Hi Robert, using the camera RTT mechanism was one of my last thoughts. Until yet i haven't any idea how i could control the render order for my passes. You have pushed me in the right way ;o) Thanks for the hint! Best regards, Christian - Original Message - From: [EMAIL PROTECTED] Robert Osfield To: osg-users@openscenegraph.net Date: 25.06.2007 10:30:44 Subject: Re: [osg-users] FBO Multipass shading Hi Christain, The way I would do is is have a set of osg::Camera set up to do FBO RTT, each outputting a separate texture. The rendering order (Camera::setRenderOrder) should be set to draw the first Camera in your sequence first and so on. A later tweak might be to try and use the same FBO in each Camera, but we will probably have to add some extra set methods into osg::Camera to do this. Robert. On 6/25/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hello, after i've implemented and bind some simple single pass shader effects like DOT3 per pixel lighting i want to bind some multi pass shader effects like DOF or Motionblurr to my scene. To use this multi pass shader effects it's required to use the frame buffer object (FBO) as render to texture target, where the result of a pass is the input for the next pass. I've searched the mailing lists and examples to find hints how i can realize this, but i can't find anything... One of my thoughts is to solve this problem using a self implemented draw cullback mechanism, where i set the FBO rendering mechanismen via OpenGL commands for each geometry... Is this the right idea? Or can i realize the effect in a better way? Thanks for any help, best regards Christian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] a bug in toggleFullscreen
hi , i found a bug in toggleFullscreen in viewerEventHandlers.cpp. I pressed 's' frist to show the stats, then I pressed 'f' to toggle to no fullscreen. I found that the 'Frame rate...' was disappeared. if we move the window a little, the 'frame rate...' appears again. This bug also affect the hud picking. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Blending for dummies :)
Thanks to everyone that helped on the issue, I also had some more readings about it, but the problem I have is that the objects in the transparent bin are sorted by depth against each other but I need the vetices to be sorted. I'm having a drawable which contains many vertices and is transparent, from some direction I see through the front of the object it's back correctly and from the other direction it doesn't look correct (even when disabling depth test, or setting depth mask to true/false), the problem is that the vertices aren't drawn in the correct order and so the source alpha and destination alpha are swapped. Is their a way to tell osg to sort the vertices of the drawable depending on the viewing angle? is it VERY slow? any ideas? thanks, Guy. - Original Message - From: Paul Martz To: 'osg users' Sent: Tuesday, June 19, 2007 7:22 PM Subject: RE: [osg-users] Blending for dummies :) Any geometries that blend against each other should be in separate Drawables, and their StateSet should have the render bin hint set to TRANSPARENT_BIN. OSG will then draw these Drawables after all opaque Drawables and in back to front order. There are other strategies too, but this is the simple common one. For more information, there are many OSG examples that demonstrate how to set up OSG for correct rendering of translucent objects. Also, check the archives, as this has been covered multiple times recently. Hope that helps, Paul Martz Skew Matrix Software LLC http://www.skew-matrix.com 303 859 9466 From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Tuesday, June 19, 2007 12:15 PM To: osg users Subject: [osg-users] Blending for dummies :) Hello, I'm trying to activate blending in some node. I set in the stateset GL_BLEND on, and added osg::BlendFunc but it seems to work only from backward views. I tried to disable GL_DEPTH and GL_CULL with no better results. any ideas? thanks, Guy. -- ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Blending for dummies :)
I think the technique of seperating the object to CW CCW direction applies for convex geometry only. It would work if back faces also have multiple layers along the viewing angle. But I guess I'll live with it. suppose I could afford the awefull time consumption of sorting the vertices each frame, would the best place to put it is in the cull traversal as cull callback? thanks anyway for the replies I've learned much, Guy. - Original Message - From: Robert Osfield [EMAIL PROTECTED] To: osg users osg-users@openscenegraph.net Sent: Sunday, June 24, 2007 11:40 AM Subject: Re: [osg-users] Blending for dummies :) Hi Guy, Sorting Geometry by depth on each frame would be very expensive for any other than trivial geometry. You don't want to do, and the OSG doesn't support it either. Approaches you can take include: 1) Break you geometry into smaller separate osg::Geometry so that the coarse grained depth sorting can work more efficiently. 2) Use a double pass approach where you draw the back faces first, then draw the front faces. You can achieve this by duplicating all your primitives, swapping the primitives binding and then ordering them so the reverse binding are first then the original bindings. You also need to enable back face culling. Robert. On 6/24/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Thanks to everyone that helped on the issue, I also had some more readings about it, but the problem I have is that the objects in the transparent bin are sorted by depth against each other but I need the vetices to be sorted. I'm having a drawable which contains many vertices and is transparent, from some direction I see through the front of the object it's back correctly and from the other direction it doesn't look correct (even when disabling depth test, or setting depth mask to true/false), the problem is that the vertices aren't drawn in the correct order and so the source alpha and destination alpha are swapped. Is their a way to tell osg to sort the vertices of the drawable depending on the viewing angle? is it VERY slow? any ideas? thanks, Guy. - Original Message - From: Paul Martz To: 'osg users' Sent: Tuesday, June 19, 2007 7:22 PM Subject: RE: [osg-users] Blending for dummies :) Any geometries that blend against each other should be in separate Drawables, and their StateSet should have the render bin hint set to TRANSPARENT_BIN. OSG will then draw these Drawables after all opaque Drawables and in back to front order. There are other strategies too, but this is the simple common one. For more information, there are many OSG examples that demonstrate how to set up OSG for correct rendering of translucent objects. Also, check the archives, as this has been covered multiple times recently. Hope that helps, Paul Martz Skew Matrix Software LLC http://www.skew-matrix.com 303 859 9466 From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Tuesday, June 19, 2007 12:15 PM To: osg users Subject: [osg-users] Blending for dummies :) Hello, I'm trying to activate blending in some node. I set in the stateset GL_BLEND on, and added osg::BlendFunc but it seems to work only from backward views. I tried to disable GL_DEPTH and GL_CULL with no better results. any ideas? thanks, Guy. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Re: Re: Asking again about states in shaders, and texturing issues (Ulrich Hertlein)
Robert, thanks for the reply, but I don't undestand something. I'm probably don't understand something about OSG or OpenGL, so I'll describe the way I believe things works, and I'll be glad for corrections. suppose I'm having several textures, than there possibly could be several texture objects on the gpu, and after the frame is rendered, at most only one of them could be the active/current/bound texture. and readFromCurrentTexture will read only that texture. suppose I want to read another texture, I do have to bind it, and only then call readFromCurrentTexture. No? thanks, Guy. - Original Message - From: Robert Osfield [EMAIL PROTECTED] To: osg users osg-users@openscenegraph.net Sent: Monday, June 18, 2007 11:37 AM Subject: Re: [osg-users] Re: Re: Asking again about states in shaders,and texturing issues (Ulrich Hertlein) When using osg::Camrea with FBO there is no need to do any expicit binding of texture objects, it will all done for your. This applies to the rest of the OSG as well, you won't need to bind texture objects manually, this will all be done for you. Only in very niche case where you are creating your own drawables that you might need to start manually binding textures. On 6/18/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hello, I have another small question regarding textures and texture objects. is there a possibility to know what is the current texture object that is bound? I need it since I render to FBO and I want to read it using to Image::readFromCurrentTexture. In order to be sure that the Current Texture is the one attached to the FBO as the color buffer, I call the bind of the osg::TextureObject that correspond to the ContextID. something like: texture.getTextureObject(contextID)-bind(); The problem is that the call to the bind function, really slows down the rendering (from ~200Hz down to ~110Hz). I thought to check if the texture objects needs binding and only if so, call the bind. thanks, Guy. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Blending for dummies :)
Hello, I'm trying to activate blending in some node. I set in the stateset GL_BLEND on, and added osg::BlendFunc but it seems to work only from backward views. I tried to disable GL_DEPTH and GL_CULL with no better results. any ideas? thanks, Guy. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Re: Re: Asking again about states in shaders, and texturing issues (Ulrich Hertlein)
Hello, I have another small question regarding textures and texture objects. is there a possibility to know what is the current texture object that is bound? I need it since I render to FBO and I want to read it using to Image::readFromCurrentTexture. In order to be sure that the Current Texture is the one attached to the FBO as the color buffer, I call the bind of the osg::TextureObject that correspond to the ContextID. something like: texture.getTextureObject(contextID)-bind(); The problem is that the call to the bind function, really slows down the rendering (from ~200Hz down to ~110Hz). I thought to check if the texture objects needs binding and only if so, call the bind. thanks, Guy. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Pick
Ok...very clearnow I know the reason Thanks so much for all your replies!! - Original Message - From: Andreas Goebel [EMAIL PROTECTED] To: osg users osg-users@openscenegraph.net Sent: Tuesday, June 12, 2007 8:20 PM Subject: Re: [osg-users] Pick Robert Nitti schrieb: It's a silly little thing not even worth posting. I build a flat list of drawables and geodes while looping through the list of hits. If the address of a geode or drawable is the same as any previously inserted in the list, I reject it. I'm CERTAIN there is a more elegant way of doing this... but I didn't look much further than my nose, so... Hi, you could either use a std::set instead of a list, or use the std - algorithm unique which deletes doublettes from a list. The std::set is unsuitable in case you want to preserve the order of the hits (i.e. want to know which object comes first). Regards, Andreas -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Jeremy L. Moles Sent: Tuesday, June 12, 2007 1:33 PM To: osg users Subject: RE: [osg-users] Pick On Tue, 2007-06-12 at 11:56 -0400, Robert Nitti wrote: Hi -- 'Picking' creates an osg::LineSegment that will intersect your scene... all the way through. So if you have (for example) a box shape, and you are picking one side, it's very possible (in fact, almost certain, unless you've done something special to prevent it) that the list of pick items will return the opposite side to you also; so you are correct Jeremy in saying that you can pick the same object twice. I've created my own post-filter to clean-up the list of drawables and geodes so that I don't treat them twice. Is this code you would have permission to post here or attach to a response? :) I'd be very interested to see optimized versions of such an operation... I hope this helps. Thanks -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Jeremy L. Moles Sent: Tuesday, June 12, 2007 10:59 AM To: osg users Subject: Re: [osg-users] Pick On Tue, 2007-06-12 at 16:52 +0200, [EMAIL PROTECTED] wrote: Dear All I implemented a Picking function... In order to debug the picking function I simply print in monitor the id of the picked drawable. The problem is that the picking function is called two time for each mouse click. I don't think your picking function is getting called twice--I think you're seeing the same object twice in your pick results, which is something I'm running into myself. I just started learning picking this weekend, so I'm not 100% sure what's going on, but perhaps someone else will be able to chime in and help us both. I originally suspected this had something to do with having a 2D HUD Camera in my scene, but I don't think that's it. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ -- Leggi GRATIS le tue mail con il telefonino i-mode di Wind http://i-mode.wind.it/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] transparency
Dear All I have a questions about transparency. I set the transparency of some nodes by these commands: osg::StateSet* state = MyNode-getOrCreateStateSet(); osg::ref_ptrosg::Material mat = new osg::Material; mat-setAlpha( osg::Material::FRONT_AND_BACK, trasparence );//Potrei risparmiarlo(tanto ci pensa il trasformer) state-setMode( GL_BLEND, osg::StateAttribute::OVERRIDE | osg::StateAttribute::ON ); state-setMode( GL_CULL_FACE, osg::StateAttribute::OVERRIDE | osg::StateAttribute::ON ); state-setAttribute( mat.get() ); What happen is that I can see the transparency if I overlap the drawable with some objects...but overlapping the same drawable with other objects it doesnt work...it seems that there is an order...does it depend by the Bin concept?...how can avoid it? -- Passa a Infostrada. ADSL e Telefono senza limiti e senza canone Telecom http://click.libero.it/infostrada ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Light
Dear All I added to my scene two light nodes. Now, when i reset the camera position by the space bar, it obviously take in account also the light nodes. How can avoid it?...how can i exclude the lights when i reset the position? Thanks -- Passa a Infostrada. ADSL e Telefono senza limiti e senza canone Telecom http://click.libero.it/infostrada ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Re:Light
Hi Paul Sorry...My mail was not clear... I dont want to move the light nodesbut instead reset the camera position. Yes I'm using osgViewer... I have some drawables in the middle of my scene. Then i have some lights far then these drawables. When i press the spacebar my camera go far away then my drawable...becouse it also include the lights in the windows. How can I automatically reset the camera position in order to have only the drawable inside my windows? Hope its better now Thanks Paul Do you mean you're using osgviewer? If so, spacebar the camera position -- it doesn't affect the transformation of individual objects in your scene. If you want your light to move independently of the rest of your scene, based on some key press, you'll need to add your own event handler to capture your key press and modify the appropriate transformation matrices accordingly. Hope that helps, -Paul -- Passa a Infostrada. ADSL e Telefono senza limiti e senza canone Telecom http://click.libero.it/infostrada ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Re: Re: Asking again about states in shaders, andtexturing issues (Ulrich Hertlein)
Hi again,
thanks, now with this information it changes all my perspective about the
textures (I'm less concerned :) ).
the example I was talking about was GLSL example, and I think it worked only
because probably the default value for the uniform sampler is zero and the
texture unit in the stateset was zero.
thanks a lot,
Guy.
- Original Message -
From: Ulrich Hertlein [EMAIL PROTECTED]
To: osg users osg-users@openscenegraph.net
Sent: Thursday, June 07, 2007 11:53 AM
Subject: Re: [osg-users] Re: Re: Asking again about states in shaders,
andtexturing issues (Ulrich Hertlein)
Hi Guy,
Guy Lifchitz wrote:
... Thanks for answering. I looked at the samples. Please help me verify
that I understand
correctly the texturing process: 1. The samplers uniforms are only
integers for the texture unit.
Yes.
2. When creating osg::Texture{nD} object the actual glGenTexture
function is called and the
object is created on the graphics card.
Exactly. The osg::Texture objects contain the OpenGL texture IDs.
(Actually a list of IDs, one for each context.)
3. When binding texture to stateset you only state the texture unit.
You state the texture unit that the osg::Texture should be bound to.
4. When rendering, the stateset binds the texture object using the id
generated with the
osg::Texture object, and bind it to the texture unit set with for that
object in the stateset.
am I right?
Yes.
To use the texture in the shader I need to add uniform with the sampler
name I'm using in the
shader and set it's value to the texture unit I gave the stateset with
the osg::Texture?
Yes.
now if all the above is true I still have two questions: 1. to bind 2
texture units to the same
texture object I only need to pass the same pointer of osg::Texture{nD}
with two different calls
to StateSet::setTextureAttribute that has different texture units?
Yes. (Boy, you're systematic. ;-)
2. I saw in some examples usage of sampler texUnit0 as the sampler of
the texture, but I haven't
seen in the C code any reference to that name, so I thought it might be
a built in uniform but I
also haven't seen it in the quick_ref. How does it work?
I'm not aware of any built-in uniforms concerning textures in OSG.
Do you mean an OSG example (if so, which) or a GLSL example.
Cheers!
/Ulrich
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
Re: [osg-users] Re: Re: Asking again about states in shaders, andtexturing issues (Ulrich Hertlein)
I forgot another little question, is the binding of texture unit to texture object is time consuming and should be avoided as much as possible, or is it a simple operation? thanks again, Guy. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: Re: [osg-users] Trouble with osgProducer::Camera setting camera position
Hi Robert, Charles,
thanks for the hints. OSG 1.9.x is not an option, because we want to use only a
stable version in our project, so i can't follow Charles introduction.
Robert, correct me if i am wrong. When i use
osgProducer::Viewer::setViewByMatrix(..) i am forced to make this call between
the update() and frame() call of the viewer to overwrite the Producer settings.
This makes it impossible to use the Producer camera features because i set a
static value. So i've to reimplement all Producer::Camera features like moving
ect..
Is there no other option to use the camera with all Producer features? It will
me drives nuts, if i must reimplement all given features of an existing
component...
Best regards,
Christian
- Original Message -
From: [EMAIL PROTECTED] Robert Osfield
To: osg-users@openscenegraph.net
Date: 01.06.2007 20:38:24
Subject: Re: [osg-users] Trouble with osgProducer::Camera setting camera
position
Hi Christian,
You are probably coming up against the problem of the
osgProducer::Viewer using camera manipulators to set the view matrix
on each frame, this overwrites the Producer setting.
One thing you could try is to use
osgProducer::Viewer::setViewByMatrix(..); and pass in a Matrix set up
as look at.
If you are new to the OSG and just starting out, then I'd recommend
using the osgViewer library found in the 1.9.x and SVN versions of the
OSG. Integration with core OSG is much cleaner in osgViewer than it
is with osgProducer.
Robert.
On 6/1/07, [EMAIL PROTECTED]
[EMAIL PROTECTED] wrote:
Hello,
I have a question about how to set the position of a camera. I do it the
following way:
Producer::CameraConfig* createCameraConfig(void)
{
// create a new render surface
Producer::RenderSurface *rsWindow = new Producer::RenderSurface;
// set the number of screens
rsWindow-setScreenNum(0);
// set the window title bar caption
rsWindow-setWindowName(CameraTest);
// set the window client size
rsWindow-setWindowRectangle(0,0, 1024, 768);
// create a new camera
Producer::Camera *camera = new Producer::Camera;
// set the render surface of the camera to the window that was just
created
camera-setRenderSurface(rsWindow);
// set the size of the camera viewport to the same size as the
window
camera-setProjectionRectangle(0,0,1024,768;
// set the camera position
camera-setViewByLookat(Producer::Vec3(0.0f,0.0f,-6.0f),
Producer::Vec3(0.0f,0.0f,0.0f),
Producer::Vec3(0.0f,1.0f,0.0f));
// create a new camera configuration container
Producer::CameraConfig *cfg = new Producer::CameraConfig;
// add the camera that was just created to the configuration
cfg-addCamera(Camera One,camera1);
return(cfg);
}
void createView()
{
// Create View with ProducerCamera Config
m_viewer = new osgProducer::Viewer(createCameraConfig());
// Create Scene root
m_viewerRoot= new osg::Group();
m_viewerRoot-setName(Root);
// Sett default Settings
m_viewer-setUpViewer(osgProducer::Viewer::STANDARD_SETTINGS);
/* Test to check position set up on View
m_viewer-setViewByLookat(Producer::Vec3(0.0f, 0.0f,-10.0f),
Producer::Vec3(0.0f, 0.0f, 0.0f),
Producer::Vec3(0.0f, 1.0f, 0.0f));
*/
m_viewer-setSceneData(m_viewerRoot.get());
m_viewer-realize();
}
I can set the position with setViewByLookat() to whatever I want, but
nothing changes to the initial camera position. It is always the default
position or something unspecific. Because Producer::Camera is an external
component which is used by the osgProducerViewer i tried to set the
position directly on the viewer by: setViewByLookat()
But this has no effect too. If i call later in the code const double* pos =
m_viewer-getPosition(); i got always some position values i've never set
...
Is there a way to predefine the Producer::Camera or Viewer Position?
Thanks for any help.
Best regards
Christian
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
Re: Re: Re: [osg-users] Trouble with osgProducer::Camera setting camera position
Hi Robert, thanks for the information about OSG 2.0. I will wait for this release and check the new viewer features. Perhaps it will make my life easier as you described, if not, i will follow your introduction. Thanks for help. Best regards, Christian - Original Message - From: [EMAIL PROTECTED] Robert Osfield To: osg-users@openscenegraph.net Date: 02.06.2007 14:00:03 Subject: Re: Re: [osg-users] Trouble with osgProducer::Camera setting camera position Hi Cristian, On 6/2/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi Robert, Charles, thanks for the hints. OSG 1.9.x is not an option, because we want to use only a stable version in our project, so i can't follow Charles introduction. Well 2.0 is coming out next week. I presume you are able to at least trail the dev versions. Refusing to consider testing out the dev versions is holding your back. osgViewer exists to make *your* life easier. Robert, correct me if i am wrong. When i use osgProducer::Viewer::setViewByMatrix(..) i am forced to make this call between the update() and frame() call of the viewer to overwrite the Producer settings. This makes it impossible to use the Producer camera features because i set a static value. So i've to reimplement all Producer::Camera features like moving ect.. Is there no other option to use the camera with all Producer features? It will me drives nuts, if i must reimplement all given features of an existing component... If you wish you can swtich off the osgProducer::Viewer's camera manipulator and manage Producer's view matrix directly, when you call Viewer::setUpViewer(mask) pass in the appropriate mask. You could also just reset the camera manipulator directly, something like viewer.getKeySwitchManipulator()-setHomePosition(...); If things are driving you nuts then just go a use 1.9.x series. Robert. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Linking problems with 1.9.4, wrappers and MinGW
Hi, I am trying to build 1.9.4 with MinGW. However when it gets to creating the wrapper DLLs, the linker, ld, takes a very large amount of memory (200MB) and after about an hour it aborts with a memory allocation failure. Everything else seems to compile and work correctly and I certainly have enough memory. Even trying to link with just Geometry wrapper obj into my application, it doesn't get past the linking phase. I can create a DLL from a very small subset of the osg wrapper objects (excluding Geometry). Has anyone encountered this before? What is special about the wrappers that leads to this problem (I am no C/C++/MinGW/DLL expert). Would a particular ld flag get around the problem? Yes, I am 90% certain I have to use MinGW. I am integrating with Haskell which uses MinGW as part of its backend. Cheers Mark ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] when add a skybox to scene, it brings z
Robert, David Thanks. The way I adopted was using two cameras to do this. The first renderes only the sky box. The second renderes other models. It's ugly but I could not find a better way the past week. Your advice was a great help to me. could you post the sample code to my email([EMAIL PROTECTED]) or osg-users mailing? Best regards, lilinx ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] when add a skybox to scene, it brings z flighting
osg-users，hi！ I use a EyeTransform(see osgVertexProgram) to Make Skybox. The Code: // osg::EyeTransform* skybox = new osg::EyeTransform; ... osg::Node*loadmodel = osgDB::readNodeFile(...); osg::Group* root = new osg::Group; root-addChild(skybox); root-addChild(loadmodel); viewer.setSceneData(root); //*** the cullingseting is COMPUTE_NEAR_FAR_USING_BOUNDING_VOLUMES when the eye near the scene center, it works fine. when the eye far from the scene center, it brings z flighting. if only setSceneData( loadmodel), it works fine both near and far from scene center. lilinx [EMAIL PROTECTED] 2007-04-30 ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Have question about the SceneView's cameraNode
Hi, I was poking around the OSG code and ran across the osg::CameraNode that is created in the SceneView class. I'm curious as to how I can benefit from having access to this node. I notice that it has a node visitor associated with it, but what isn't exactly clear is how the CameraNode is actually used. What is its role in OSG? Hypothetically, if I were to obtain a pointer to this CameraNode, could I modify it to render to texture instead of its default rendering method (whatever that may be)? Typically, I have always created separate CameraNode objects for use with RTT which works very well, but someone pointed out that a CameraNode is already obtainable from the SceneView class. From there I started wondering. Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Window creation parameters.
Hi, When using Producer to set up a default window, is it possible to specify parameters that the window should have upon creation? I know that I can specify my own window to use, but unfortunately, the 3rd party tool that I am using in conjunction with OSG isn't set up to do that easily. Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Window creation parameters.
Hi RJ, Thanks for the reply and links. However, I am just not seeing how I can alter the style of the window prior to its creation. I've tried creating a parent window and passing that via DeltaWin::GetInstance(0)-GetRenderSurface()-setParentWindow(win); However, all this does is to create a child window without the WS_VISIBLE style. I discovered that I will need to provide a parent window due to a couple of program specific reasons in my application. Looking at the RenderSurface_Win32.cpp file, there don't seem to be any notifications that the child window of the supplied parent has been created. At this point, a code snippet would be most helpful. Thanks, Brian On Mon, 02 Apr 2007 23:11:32 +0530, RJ [EMAIL PROTECTED] wrote : Hi Brian, Yes you can do that by using RenderSurface.Get the Producer:: RenderSurface object from the Producer::Camera and then you can call different function of RenderSurface class to set the window properties. For more information look at the documentation on this site http://andesengineering.com/Producer/Doxygen/html/classProducer_1_1RenderSur face.html There is also a tutorial on Producer on the link http://andesengineering.com/Producer/Tutorial/index.html I hope this answer your query.If not please let me know i will send you a code snippet on the same. best regards RJ On Mon, 2007-04-02 at 17:29 +, [EMAIL PROTECTED] wrote: Hi, When using Producer to set up a default window, is it possible to specify parameters that the window should have upon creation? I know that I can specify my own window to use, but unfortunately, the 3rd party tool that I am using in conjunction with OSG isn't set up to do that easily. Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Window creation parameters.
Hi RJ, Just thought I'd let you know that I got this to work (in a fashion.) I created a separate thread for the parent window to get/dispatch messages, and the child, or OSG window, displayed fine in that scenario. I coded some additional logic to determine when the child window becomes visible. Thanks for your help. I don't think I need the code snippet now. Brian On 2 Apr 2007 18:40:19 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote : Hi RJ, Thanks for the reply and links. However, I am just not seeing how I can alter the style of the window prior to its creation. I've tried creating a parent window and passing that via DeltaWin::GetInstance(0)-GetRenderSurface()-setParentWindow(win); However, all this does is to create a child window without the WS_VISIBLE style. I discovered that I will need to provide a parent window due to a couple of program specific reasons in my application. Looking at the RenderSurface_Win32.cpp file, there don't seem to be any notifications that the child window of the supplied parent has been created. At this point, a code snippet would be most helpful. Thanks, Brian On Mon, 02 Apr 2007 23:11:32 +0530, RJ [EMAIL PROTECTED] wrote : Hi Brian, Yes you can do that by using RenderSurface.Get the Producer:: RenderSurface object from the Producer::Camera and then you can call different function of RenderSurface class to set the window properties. For more information look at the documentation on this site http://andesengineering.com/Producer/Doxygen/html/classProducer_1_1RenderSur face.html There is also a tutorial on Producer on the link http://andesengineering.com/Producer/Tutorial/index.html I hope this answer your query.If not please let me know i will send you a code snippet on the same. best regards RJ On Mon, 2007-04-02 at 17:29 +, [EMAIL PROTECTED] wrote: Hi, When using Producer to set up a default window, is it possible to specify parameters that the window should have upon creation? I know that I can specify my own window to use, but unfortunately, the 3rd party tool that I am using in conjunction with OSG isn't set up to do that easily. Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] precise object occlusion
I need to PARTIALLY occlude some moving objects. Actually, using the osg::Occluder class and looking at the osgoccluder example, I realized that when my moving object move just a little bit behind the occluder plane, it is hidden completely...even though there is a large part of that object that should be still visible. How can I hide JUST the part of objects that are behind the occluder? By the way, beside the occluder plane, is possible to use any kind of geometry (I mean any 3D object...of any shape) as an occluder? Actually, my problem is to have 'transparent objects that should be able to occlude some other objects'...any suggestion? Thanks! PS: I'm new to osg ;) ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] CombatSimulatorProject
RJ wrote: Hi folks, I am trying to build CSP from source code. One of the dependencies openalpp (openAl++) i am not able to build from source code. Does any one has tried building this library from the source code ? best regards RJ i highly suggest to ask that question direct in the development discussion at http://csp.sourceforge.net/forum/ as there are all the csp developers at home. :) kind regards ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: Re: [osg-users] Render scene part to texture
Hi Robert, i've checked the osgdistortion example and the osgprerender/renderToTexture example again. Rendering a scene part following these examples will not work and i am not sure if i'am on the right way... For a better understanding what i want to realize, here a short example what i will do: Imagine a smal scene like a racing track. In this scene you have several cars moving on the track and some environment objects arround this track. During rendering this scene i want to create multiple views of this scene by using RTT's. In example for a tracking-cameras or rear/side-view projected as HUD elements on top of the scene/screen. My problem is, if i create the RTT based on the examples, the object which is the child of RTT-Camera is projected to the RTT. But only the object no one of the objects arraound this object (in example some of the other cars in the camera field of view or the sky in the backround ect.). If i add the scene to the camera or the group/translation where all Objects are attached, i only got an black screen or texture and sometimes a SIGFAULT. Now i am not sure, if this is it the right way to realize multiple views? I know i could do this by using more openProducer cameras, but this bar me from using shaders for some special effects which i will use on the rtts ... Best regards, Chris - Original Message - From: [EMAIL PROTECTED] Robert Osfield To: osg-users@openscenegraph.net Date: 12.02.2007 21:07:03 Subject: Re: [osg-users] Render scene part to texture HI Chris, I don't of its what you are looking for by the osgdistortion example uses inheritance of the view and projection matrices for its RTT camera to capture the scene as it would have been seen using normal OSG camera manipualation. The resulting texture is then distorted and applied to the screen. Robert. On 2/12/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hello, since some days i try to develop a tracking camera for my scene. The view of the camera should be rendered into a texture to show that texture as hud element combined with some shader effects. To realize that i've played arround with the render to texture example and the osgprerender example. So far so good. Rendering a single object to texture should not be the problem if i add the object as subgraph to the render to texture camera. But how should i realize this, if the object is a part of the rendering scene? In this case the object and all surrounding objects (normaly parallel and parent nodes in the graph) in the field of view of the camera should be rendered to this texture? Is this possible? I've checked the NPS tutorials, where the tracking camera is realized as separate view context. Is it possible to render the result of this second view to a texture? Best regards, Chris ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Question about osgdepthshadow example.
Hi, I'm trying to modify this example slightly so that only the depth buffer portion is rendered. The shadow is cool, but is there a way to just render the depth buffer? Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Status of Combat Simulator Project
Our SVN server is up again at https://www.zerobar.net/csp/ If You have any further questions do not hesitate to ask. Kind regards ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Status of Combat Simulator Project
Praveen Bhaniramka wrote: Folks, Would anyone on this list have information about the status of the CSP? in ongoing development. http://csp.sourceforge.net/wiki/Development I downloaded the Windows demos and definitely feel it would be fun to play with and possibly develop further as needed. But I am unable to get the SVN repository or register to the forums. it should be possible to register in the forum, we only have some stuff changed to prevent spam after bad experience. if you still can´t login catch me on icq 133668500 (please mention csp) seems that the svn is unreachable at the moment, sent onsight who run our development support server a message about it, should be fixed soon. there are several improvements of the svn version over the demo. happy welcome, great that you are interested in developing. :) kind regards, dust ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Performance drops when using 4 or more cameranodes (RTT)
Hi Robert, Thanks for that tip! It turns out that your suspicions were correct. I get the following output when I run my app: Setting up osg::CameraNode::FRAME_BUFFER_OBJECT RenderStage::runCameraSetUp(), FBO setup failed, FBO status= 0x8cd9 OpenGL extension 'ARB_render_texture' is not supported. Failed to aquire Graphics Context Setting up osg::CameraNode::FRAME_BUFFER What I don't understand is how an nVidia 7800 card does not support the ARB_render_texture extension. :? My guess is that something isn't being configured correctly, or that the extension has been replaced with something else, or perhaps it is a driver bug. Any thoughts? The card isn't exactly a low-end card. I've also checked to see if WGL_ARB_render_texture is supported, and it isn't either. Thanks, Brian On Thu, 25 Jan 2007 08:57:03 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian, How is the RTT being implementated? Are you sure its FBO? Running use OSG_NOTIFY_LEVEL set to DEBUG will print out what path the OSG uses for implementing the FBO. I have personally worked with FBO sizes way higher than window resolution, all the way up to 4k x 4k without problem. I must admit I have only done this with a full screen window, but I'd be very surprised if its any different. From the description of the performance you a getting it I does kinda sound like there issue with the OpenGL driver. Robert. On 24 Jan 2007 17:58:53 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi Robert, Actually, I have discovered what my problem was. On two of the machines that I was testing on, their window dimensions were 640x480. My original test machine was 512x512. When I changed my test machine to 640x480, I saw the same drop in FPS on it as well. To my knowledge, I am not modifying the size of the 512x512 textures, so I am curious as to why increasing my window dimensions from 512x512 to 640x480 would cause such a drastic frame rate drop. :? The only thing that I can think of would be that since the initial RTT texture is smaller than the actual window dimensions, this imposes a performance hit when the render to texture actually occurs. Of course, that's just a stab in the dark. Can you think of any suggestions that I could try to get my framerate back up? The final window dimensions will ultimately need to be full screen, but that kind of frame rate hit (~2-3 FPS) will just kill any usability. Even having just one floating point texture at 512x512 drops the frame rate to around 20FPS for a full screen app. BTW, the final texture for the RTT is being rendered at 512x512, not full screen. It is a secondary view. Thanks, Brian On Wed, 24 Jan 2007 09:13:51 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian Jeremy I know this doesn't help you guys, but for the record I used Linux almost everyday for the past 5 years, running with various NVidia hardware + drivers and haven't ever come across changes in performance w..r.t first boot up vs continued use. So its certainly possible to have OpenGL running near flawlessly year after year. But... there are dodgey drivers around and it sounds like you guys have got them installed, the quickest way to establish a pattern is to publish what drivers you are have a problem with and see if others are having the same problem. Try upgrading or stepping back driver versions. NVidia have a set of forums that you could look at. I can't recally the address but google will find it for you. Robert. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Performance drops when using 4 or more cameranodes (RTT)
Hi Robert, I finally found the bug (in my code, of course.) :) It is amazing how one missing line of code can really wreak havoc on performance. When I was setting up my cameraNodes, I set the first cameraNode that the scene renders to correctly. The other camera nodes that I use with multiple shaders were not set up correctly. I had forgotten to include the line: m_CameraNode-setViewport(0, 0, width, height); Adding this one line solved all of my problems. I cannot stress how appreciative I am of the debugging capabilities built in to OSG! I only wish that I had known about them earlier. Thanks, Brian Hi Robert, The osgprerender example using --hdr --fbo cow.osg works fine. The last lines in the debug output are: OpenGL extension 'GL_EXT_framebuffer_object' is supported. OpenGL extension 'GL_ARB_shader_objects' is supported. OpenGL extension 'GL_ARB_vertex_shader' is supported. OpenGL extension 'GL_ARB_fragment_shader' is supported. OpenGL extension 'GL_ARB_shading_language_100' is supported. glVersion=2, isGlslSupported=YES, glslLanguageVersion=1.1 Setting up osg::CameraNode::FRAME_BUFFER_OBJECT I'll take a closer look at that example. I'm obviously missing something. Whether it's in my initial setup for the openGL window, the texture2D object, or the FBO set up, I don't know. However, I should be able to track it down given that the example works. After looking at the example, it is quite confusing to me because the way the example sets up its FBO and texture seem to be exactly how my app does it. Thanks for the help, Brian HI Brian, I have 7800GTs and they work fine with FBO's. The pathway your code seems to be going through is for pbuffers though, so it looks like FBO isn't available, so its failling back to pbuffers, which is also not supported, so its falling back to frame buffer. Both FBO's and pbuffers are supported on your graphics card though, so I'm a bit perplexed as to what is amiss. Drivers? Do the OSG RTT examples like osgprerender work ok? By default it should use FBO, then fallback to pbuffers if they arn't supported. Robert. On 26 Jan 2007 16:40:35 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi Robert, Thanks for that tip! It turns out that your suspicions were correct. I get the following output when I run my app: Setting up osg::CameraNode::FRAME_BUFFER_OBJECT RenderStage::runCameraSetUp(), FBO setup failed, FBO status= 0x8cd9 OpenGL extension 'ARB_render_texture' is not supported. Failed to aquire Graphics Context Setting up osg::CameraNode::FRAME_BUFFER What I don't understand is how an nVidia 7800 card does not support the ARB_render_texture extension. :? My guess is that something isn't being configured correctly, or that the extension has been replaced with something else, or perhaps it is a driver bug. Any thoughts? The card isn't exactly a low-end card. I've also checked to see if WGL_ARB_render_texture is supported, and it isn't either. Thanks, Brian On Thu, 25 Jan 2007 08:57:03 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian, How is the RTT being implementated? Are you sure its FBO? Running use OSG_NOTIFY_LEVEL set to DEBUG will print out what path the OSG uses for implementing the FBO. I have personally worked with FBO sizes way higher than window resolution, all the way up to 4k x 4k without problem. I must admit I have only done this with a full screen window, but I'd be very surprised if its any different. From the description of the performance you a getting it I does kinda sound like there issue with the OpenGL driver. Robert. On 24 Jan 2007 17:58:53 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi Robert, Actually, I have discovered what my problem was. On two of the machines that I was testing on, their window dimensions were 640x480. My original test machine was 512x512. When I changed my test machine to 640x480, I saw the same drop in FPS on it as well. To my knowledge, I am not modifying the size of the 512x512 textures, so I am curious as to why increasing my window dimensions from 512x512 to 640x480 would cause such a drastic frame rate drop. :? The only thing that I can think of would be that since the initial RTT texture is smaller than the actual window dimensions, this imposes a performance hit when the render to texture actually occurs. Of course, that's just a stab in the dark. Can you think of any suggestions that I could try to get my framerate back up? The final window dimensions will ultimately need to be full screen, but that kind of frame rate hit (~2-3 FPS) will just kill any usability. Even having just one
Re: [osg-users] Performance drops when using 4 or more cameranodes (RTT)
Hi Robert, Actually, I have discovered what my problem was. On two of the machines that I was testing on, their window dimensions were 640x480. My original test machine was 512x512. When I changed my test machine to 640x480, I saw the same drop in FPS on it as well. To my knowledge, I am not modifying the size of the 512x512 textures, so I am curious as to why increasing my window dimensions from 512x512 to 640x480 would cause such a drastic frame rate drop. :? The only thing that I can think of would be that since the initial RTT texture is smaller than the actual window dimensions, this imposes a performance hit when the render to texture actually occurs. Of course, that's just a stab in the dark. Can you think of any suggestions that I could try to get my framerate back up? The final window dimensions will ultimately need to be full screen, but that kind of frame rate hit (~2-3 FPS) will just kill any usability. Even having just one floating point texture at 512x512 drops the frame rate to around 20FPS for a full screen app. BTW, the final texture for the RTT is being rendered at 512x512, not full screen. It is a secondary view. Thanks, Brian On Wed, 24 Jan 2007 09:13:51 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian Jeremy I know this doesn't help you guys, but for the record I used Linux almost everyday for the past 5 years, running with various NVidia hardware + drivers and haven't ever come across changes in performance w..r.t first boot up vs continued use. So its certainly possible to have OpenGL running near flawlessly year after year. But... there are dodgey drivers around and it sounds like you guys have got them installed, the quickest way to establish a pattern is to publish what drivers you are have a problem with and see if others are having the same problem. Try upgrading or stepping back driver versions. NVidia have a set of forums that you could look at. I can't recally the address but google will find it for you. Robert. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
RE: [osg-users] Performance drops when using 4 or more cameranodes (RTT)
Well, I'm perplexed for sure. I decided to uninstall my nvidia drivers, disabled my antivirus service, and reinstalled the drivers as per nvidia's suggestion. I rebooted, and the application ran great at 75FPS. I then decided to test other graphics settings because we have nearly identical systems that did not benefit from the drivers being reinstalled. The frame rate dropped down to 16FPS again, and I cannot get the original 75FPS to resurface. I am truly at a loss here. On Tue, 23 Jan 2007 12:21:30 -0500, Gordon Tomlinson [EMAIL PROTECTED] wrote : One yhing don't forget MIP maps on your textures you need a to times by 4 then divide by 3 to your total texture usage below to add in the mip map requirments of your textures But that does not expand you usage too much... Best Regards Gordon __ Gordon Tomlinson Email : gordon.tomlinson @ overwatch.com YIM/AIM: Gordon3dBrit MSN IM : Gordon3dBrit @ 3dSceneGraph.com __ Self defence is not a function of learning tricks but is a function of how quickly and intensely one can arouse one's instinct for survival - Master Tambo Tetsura -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of [EMAIL PROTECTED] Sent: Tuesday, January 23, 2007 12:20 PM To: osg users Subject: Re: [osg-users] Performance drops when using 4 or more cameranodes (RTT) Hi, First, to answer Anders' question, my general specs are: Windows XP Pro nVidia 7800 GTX (512MB) As for how much texture memory that I'm using, I simplified my application to only load the cow model. I have seven other textures loading: texN is width x height (internal format, source format, source type) tex0 is 512x512 (GL_RGBA16F_ARB, GL_RGBA, GL_UNSIGNED_BYTE) tex1 is 512x512 (GL_LUMINANCE, GL_LUMINANCE, GL_FLOAT) tex2-5 are all 512x512 (GL_RGBA16F_ARB, GL_RGBA, GL_FLOAT) tex6 is 512x512 (GL_RGBA, GL_RGBA, GL_FLOAT) Unless I'm misunderstanding how texture memory works, seven 512x512 textures in the worst case scenario would take up: (512*512) * 4bytes (RGBA) * 4bytes (32-bit) * 7 textures = 28MB The second 4bytes (32-bit) is assuming GL_FLOAT uses 32-bits per RGBA. I realize that this calculation is wrong in the sense that not all textures are using the same internal formats, source types, or source formats; but I would assume that this is the worst it could ever be for these textures. With that in mind, it would appear that I should have plenty of texture memory available. I'm wondering if there is some limit to the amount of texture memory that can be associated with FBOs. The same number of FBOs using texture formats of (GL_RGBA, GL_RGBA, GL_UNSIGNED_BYTE) works great, so I'm assuming that I haven't hit a limit as far as the number of FBOs is concerned. If there is a limit on the overall texture memory that can be associated with FBOs, is there a relatively painless way of determining what that maximum texture memory size is? Thanks, Brian On Tue, 23 Jan 2007 08:50:39 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian, I'd check to see if you graphics card has sufficient memory for what you are trying to do Robert. On 22 Jan 2007 23:27:35 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi, I'm using floating point textures as my targets for multipass RTT. In my tests I have noticed that if three or less camera nodes are set up using GL_RGBA16F_ARB internal format textures, the frame rate stays at a constant 75FPS when only the texture is rendered to the screen (i.e. the texture is rendered over the full window). If I add an additional camera node, the frame rate drops to 15 FPS. Five nodes yields 8FPS, and 5FPS are achieved with six nodes. Also, if I set the texture to render to 1/4 of the screen and position the main camera so that the actual scene being rendered to texture is visible, I'll get 25FPS for the three camera node implementation. I'm setting up my textures with internal formats of GL_RGBA16F_ARB, a source type of GL_FLOAT, and a source format of GL_RGBA. The FBO is set up like: fbo-setAttachment(GL_COLOR_ATTACHMENT0_EXT, osg::FrameBufferAttachment(texture)); fbo-setAttachment(GL_DEPTH_ATTACHMENT_EXT, osg::FrameBufferAttachment(new osg::RenderBuffer(512, 512, GL_DEPTH_COMPONENT24))); My min/max filters are set up as NEAREST. If I change my textures to internal formats of GL_RGBA and the source type as GL_UNSIGNED_BYTE, my performance is great. Any thoughts? Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org
RE: [osg-users] Performance drops when using 4 or more cameranodes (RTT)
Well, I managed to get the 75FPS back by doing a Restore on the nVidia control panel. The perplexing thing is that there are two other systems that I've tested on: one has a high frame rate and the other's is slow. Both are using nVidia 7800 GT cards with default settings. They are also both running WinXP Pro with nearly identical memory/hardware specs as well. I'd really like to get this issue resolved, however I do realize that it is not an OSG specific topic. Are there any good forums that someone could recommend? Thanks, Brian On Tue, 23 Jan 2007 17:24:54 -0500, Jeremy L. Moles [EMAIL PROTECTED] wrote : On Tue, 2007-01-23 at 22:23 +, [EMAIL PROTECTED] wrote: Well, I'm perplexed for sure. I decided to uninstall my nvidia drivers, disabled my antivirus service, and reinstalled the drivers as per nvidia's suggestion. I rebooted, and the application ran great at 75FPS. I then decided to test other graphics settings because we have nearly identical systems that did not benefit from the drivers being reinstalled. The frame rate dropped down to 16FPS again, and I cannot get the original 75FPS to resurface. I am truly at a loss here. I don't know if you're talking about Linux here (I don't think you are) but I see the exact same thing w/ my Quadro2500M (which is a beastly card) in Fedora 6. If I run an osg example app right when I boot I get N fps; if I close and run that same app again I get about N/3 fps. 100% repeatable and consistent. Doesn't happen with DirectX, so maybe it's an OpenGL thing. On Tue, 23 Jan 2007 12:21:30 -0500, Gordon Tomlinson [EMAIL PROTECTED] wrote : One yhing don't forget MIP maps on your textures you need a to times by 4 then divide by 3 to your total texture usage below to add in the mip map requirments of your textures But that does not expand you usage too much... Best Regards Gordon __ Gordon Tomlinson Email : gordon.tomlinson @ overwatch.com YIM/AIM: Gordon3dBrit MSN IM : Gordon3dBrit @ 3dSceneGraph.com __ Self defence is not a function of learning tricks but is a function of how quickly and intensely one can arouse one's instinct for survival - Master Tambo Tetsura -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of [EMAIL PROTECTED] Sent: Tuesday, January 23, 2007 12:20 PM To: osg users Subject: Re: [osg-users] Performance drops when using 4 or more cameranodes (RTT) Hi, First, to answer Anders' question, my general specs are: Windows XP Pro nVidia 7800 GTX (512MB) As for how much texture memory that I'm using, I simplified my application to only load the cow model. I have seven other textures loading: texN is width x height (internal format, source format, source type) tex0 is 512x512 (GL_RGBA16F_ARB, GL_RGBA, GL_UNSIGNED_BYTE) tex1 is 512x512 (GL_LUMINANCE, GL_LUMINANCE, GL_FLOAT) tex2-5 are all 512x512 (GL_RGBA16F_ARB, GL_RGBA, GL_FLOAT) tex6 is 512x512 (GL_RGBA, GL_RGBA, GL_FLOAT) Unless I'm misunderstanding how texture memory works, seven 512x512 textures in the worst case scenario would take up: (512*512) * 4bytes (RGBA) * 4bytes (32-bit) * 7 textures = 28MB The second 4bytes (32-bit) is assuming GL_FLOAT uses 32-bits per RGBA. I realize that this calculation is wrong in the sense that not all textures are using the same internal formats, source types, or source formats; but I would assume that this is the worst it could ever be for these textures. With that in mind, it would appear that I should have plenty of texture memory available. I'm wondering if there is some limit to the amount of texture memory that can be associated with FBOs. The same number of FBOs using texture formats of (GL_RGBA, GL_RGBA, GL_UNSIGNED_BYTE) works great, so I'm assuming that I haven't hit a limit as far as the number of FBOs is concerned. If there is a limit on the overall texture memory that can be associated with FBOs, is there a relatively painless way of determining what that maximum texture memory size is? Thanks, Brian On Tue, 23 Jan 2007 08:50:39 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian, I'd check to see if you graphics card has sufficient memory for what you are trying to do Robert. On 22 Jan 2007 23:27:35 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi, I'm using floating point textures as my targets for multipass RTT. In my tests I have noticed that if three or less camera nodes are set up using GL_RGBA16F_ARB internal format textures, the frame
RE: [osg-users] Performance drops when using 4 or more cameranodes (RTT)
Well, I managed to get the 75FPS back by doing a Restore on the nVidia control panel. The perplexing thing is that there are two other systems that I've tested on: one has a high frame rate and the other's is slow. Both are using nVidia 7800 GT cards with default settings. They are also both running WinXP Pro with nearly identical memory/hardware specs as well. I'd really like to get this issue resolved, however I do realize that it is not an OSG specific topic. Are there any good forums that someone could recommend? Thanks, Brian On Tue, 23 Jan 2007 17:24:54 -0500, Jeremy L. Moles [EMAIL PROTECTED] wrote : On Tue, 2007-01-23 at 22:23 +, [EMAIL PROTECTED] wrote: Well, I'm perplexed for sure. I decided to uninstall my nvidia drivers, disabled my antivirus service, and reinstalled the drivers as per nvidia's suggestion. I rebooted, and the application ran great at 75FPS. I then decided to test other graphics settings because we have nearly identical systems that did not benefit from the drivers being reinstalled. The frame rate dropped down to 16FPS again, and I cannot get the original 75FPS to resurface. I am truly at a loss here. I don't know if you're talking about Linux here (I don't think you are) but I see the exact same thing w/ my Quadro2500M (which is a beastly card) in Fedora 6. If I run an osg example app right when I boot I get N fps; if I close and run that same app again I get about N/3 fps. 100% repeatable and consistent. Doesn't happen with DirectX, so maybe it's an OpenGL thing. On Tue, 23 Jan 2007 12:21:30 -0500, Gordon Tomlinson [EMAIL PROTECTED] wrote : One yhing don't forget MIP maps on your textures you need a to times by 4 then divide by 3 to your total texture usage below to add in the mip map requirments of your textures But that does not expand you usage too much... Best Regards Gordon __ Gordon Tomlinson Email : gordon.tomlinson @ overwatch.com YIM/AIM: Gordon3dBrit MSN IM : Gordon3dBrit @ 3dSceneGraph.com __ Self defence is not a function of learning tricks but is a function of how quickly and intensely one can arouse one's instinct for survival - Master Tambo Tetsura -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of [EMAIL PROTECTED] Sent: Tuesday, January 23, 2007 12:20 PM To: osg users Subject: Re: [osg-users] Performance drops when using 4 or more cameranodes (RTT) Hi, First, to answer Anders' question, my general specs are: Windows XP Pro nVidia 7800 GTX (512MB) As for how much texture memory that I'm using, I simplified my application to only load the cow model. I have seven other textures loading: texN is width x height (internal format, source format, source type) tex0 is 512x512 (GL_RGBA16F_ARB, GL_RGBA, GL_UNSIGNED_BYTE) tex1 is 512x512 (GL_LUMINANCE, GL_LUMINANCE, GL_FLOAT) tex2-5 are all 512x512 (GL_RGBA16F_ARB, GL_RGBA, GL_FLOAT) tex6 is 512x512 (GL_RGBA, GL_RGBA, GL_FLOAT) Unless I'm misunderstanding how texture memory works, seven 512x512 textures in the worst case scenario would take up: (512*512) * 4bytes (RGBA) * 4bytes (32-bit) * 7 textures = 28MB The second 4bytes (32-bit) is assuming GL_FLOAT uses 32-bits per RGBA. I realize that this calculation is wrong in the sense that not all textures are using the same internal formats, source types, or source formats; but I would assume that this is the worst it could ever be for these textures. With that in mind, it would appear that I should have plenty of texture memory available. I'm wondering if there is some limit to the amount of texture memory that can be associated with FBOs. The same number of FBOs using texture formats of (GL_RGBA, GL_RGBA, GL_UNSIGNED_BYTE) works great, so I'm assuming that I haven't hit a limit as far as the number of FBOs is concerned. If there is a limit on the overall texture memory that can be associated with FBOs, is there a relatively painless way of determining what that maximum texture memory size is? Thanks, Brian On Tue, 23 Jan 2007 08:50:39 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian, I'd check to see if you graphics card has sufficient memory for what you are trying to do Robert. On 22 Jan 2007 23:27:35 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi, I'm using floating point textures as my targets for multipass RTT. In my tests I have noticed that if three or less camera nodes are set up using GL_RGBA16F_ARB internal format textures, the frame
Re: [osg-users] osgVideo
Here comes another question; Rendering in some systems due to the cpu rate or something else may not be as fast as possible, that is they may not show the graphics in real-time; are these screen capturing software programs able to handle this, may be by writing the color buffer first rather than taking the actual screen image? Hope still some people following this thread :)) Any ideas are appreciated.. Thanks in advance :) On 11/10/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Using the route described above, you can grab an image from a frame in a post-draw callback and write it to a file regardless of its size and aspect ratio I think, since it is not used as a texture. Any video editing tool (not just AVID) should be able to grab a sequence of frames and make an AVI out of it. I am sure even Window's Movie Maker, the free one that comes with XP will do that. I also ran across a little snippet of code somewhere that you could use to write out an AVI, and even asks for what compression you want using the CODECs installed. I might be able to dig it up if you are interested. The part about running with a constant framerate is critical too, because it takes quite a bit of processor/disk to do it all and the animation will not be smooth. -- Rick On 11/9/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: When we're putting together promo videos, we just dump out each frame as a tga then use Avid to edit the sequence and turn it into a useable video. Admittedly we have a high-end Avid suite for this, but the dumping of the frames goes from the PS3, down the network and onto our PCs. We run at about 1 fps, but you can capture quite a lot of video by letting it run overnight. In our case, we have several machines capturing different parts of the game, then let the video guys sift through it all. There's no reason why you couldn't set up a 512x512 render target at the top of your scenegraph, change the aspect ratio to that of 720x480 (3:2) so it appears stretched (or if your card supports non-power-of-two textures, then you can do it at native resolution), then dump out each frame to an image file. I don't have any code to hand, but OSG can cope with that easily. When you come to encode it, you can stretch it back to the desired resolution. Oh, and don't forget to make your animation dependent on framerate (i.e. assume you're running at a constant framerate), so the animation runs smoothly at full speed; otherwise you'll pick up any variation in rendering speed. To put it all together, Avid did a home edition of their video editing software that was free to download and is fairly powerful, but I can't find it on their page, maybe they stopped doing it. If you're on Linux, there's a free bit of software for video editing, called Cinelerra (http://heroinewarrior.com/cinelerra.php3). If you want to automate the process, and you use Linux, Transcode (http://www.transcoding.org/cgi-bin/transcode) will import individual frames and spit out a video in a bunch of different formats. It might take some time to set up to your liking, but it'll happily cope with most things you can throw at it; it'll even mux any audio you might want to use. John Donovan Sony Computer Entertainment Europe http://www.scee.com [EMAIL PROTECTED] wrote on 09/11/2006 00:45:14: Hello, PLEASE! Does anybody have a way to automatically create an mpg or avi from an OSG fly through? So, that it plays back in real-time? Basically, I want to generate videos for demo purposes. ie. loop update frame render scene capture screen down convert screen capture to acceptable format (720x480) end loop concatenate screen captures into video file. Recommendations are also welcome. Thanks. Zach ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ** This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify [EMAIL PROTECTED] This footnote also confirms that this email message has been checked for all known viruses. ** Sony Computer Entertainment Europe ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ -- CodeSchbeke Software Solutions Senior Project Team Computer Engineering Department Middle East
[osg-users] Performance drops when using 4 or more camera nodes (RTT)
Hi, I'm using floating point textures as my targets for multipass RTT. In my tests I have noticed that if three or less camera nodes are set up using GL_RGBA16F_ARB internal format textures, the frame rate stays at a constant 75FPS when only the texture is rendered to the screen (i.e. the texture is rendered over the full window). If I add an additional camera node, the frame rate drops to 15 FPS. Five nodes yields 8FPS, and 5FPS are achieved with six nodes. Also, if I set the texture to render to 1/4 of the screen and position the main camera so that the actual scene being rendered to texture is visible, I'll get 25FPS for the three camera node implementation. I'm setting up my textures with internal formats of GL_RGBA16F_ARB, a source type of GL_FLOAT, and a source format of GL_RGBA. The FBO is set up like: fbo-setAttachment(GL_COLOR_ATTACHMENT0_EXT, osg::FrameBufferAttachment(texture)); fbo-setAttachment(GL_DEPTH_ATTACHMENT_EXT, osg::FrameBufferAttachment(new osg::RenderBuffer(512, 512, GL_DEPTH_COMPONENT24))); My min/max filters are set up as NEAREST. If I change my textures to internal formats of GL_RGBA and the source type as GL_UNSIGNED_BYTE, my performance is great. Any thoughts? Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] GL_RGBA16F_ARB Texture source formats
Hi, I am also trying to get a grasp on this subject. Basically, I have n shader programs. Shaders 1 to n-1 need to use floating point textures. As I understand it, the first shader to be rendered needs to have an internal format of GL_RGBA16F_ARB, a source format of GL_RGBA, and a source type of GL_UNSIGNED_BYTE. The floating point shaders need to have internal formats of GL_RGBA16F_ARB and source formats and types of GL_FLOAT The last shader would need to have an internal format of GL_RGBA, a source format and type of GL_FLOAT. Does this seem correct? When I try this approach, I get an error: RenderStage::drawInner(,) OpenGL errorNo= 0x502 Obviously, I'm doing something wrong, but I am uncertain as to what might be the problem. Thanks, Brian On Wed, 17 Jan 2007 12:19:50 +, David Spilling [EMAIL PROTECTED] wrote : Robert, Thanks for that. I can see that this makes sense when (for example) you have an image in disk in RGBA8, and you load it into a texture with a different internal format (e.g. RGBA16F), then a conversion takes place on the way. But if you have a RTT, does the source format actually matter at all? By setting a texture internal format to RGBA16F and attaching that to a cameraNode, you are saying that the render target is RGBA16F, and shaders and the like operating on the graph underneath the camera can output float values; theres no RGBA8 to 16F conversion going on, is there? David ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Overriding shader values using osgNV
Hi, I noticed in the archives that osgNV is no longer being maintained, but I was wondering if anyone knows if it supports overriding the shader uniform values like osg::Uniform does. For example, is there an osgNV equivalent to: osg::Uniform* uniform = new osg::Uniform( osg::Uniform::SAMPLER_2D, shaderEnabled); uniform-set(1); rootStateSet-addUniform(uniform, osg::StateAttribute::OVERRIDE); This lets me modify GLSL shader uniforms with no problem. The problem is that we have some shaders written with Cg that need to be customized for a main camera and a CameraNode. Test GLSL shaders work fine when modifying uniform values as shown above, but I've yet to figure out how to do this with osgNV. Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Drawables, RTT, and shaders
Hi, It's been about a month since I looked at this, but I'm finally able to get back to it. Can you please explain how to override the inherited setting for the RTT CameraNode's StateSet? I know how to add a Uniform value to a StateSet, but I am not familiar with how to override the value for the osg::Uniform of a parent node. I know how to override or inherit values using the setMode and/or setAttributesAndModes functions of the StateSet, but I have never used them in conjunction with a shader's Uniform value. Thanks, Brian Hi Brian, You should be able to define a StateSet on the RTT Camera subgraph that contains a osg::Uniform that overrides the inherited setting. For maximum performance you might want to just use a different shader though, one that doesn't have any lighting path. Robert. Hi, I have an application that I'm developing that uses RTT to render's an image to a quad in my scene that is based on the overall scene. Best analogy that I can give is that the rendered quad is similar to a security camera. The main display consists of a drawable that has a shader program attached to it to produce lighting effects. The shader has a parameter that allows the lighting effect to be toggled. The security camera is rendered using a series of camera nodes. Each camera node has a quad, a texture, and a shader attached to the geometry node for the quad. All this works nicely. The problem that I have run into is that ideally, I would like to toggle the lighting effects shader for the RTT. Currently, the main display's shader has lighting enabled, but the lighting effects are carried into the RTT for the security camera. Ideally, I'd like to enable the main display's lighting effects, but disable them for the RTT so that I can produce more realistic effects like night vision if that is possible. One thing to keep in mind is that all lighting effects are done through the shader, but I do have a flag in the shader program that can disable the effect. I'm guessing that either node visitors or setting a cull mask might be the way to go, but I haven't had any luck with either of those methods. It's very possible that I'm going about it the wrong way. Any ideas? Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Drawables, RTT, and shaders
Hi, It's been about a month since I looked at this, but I'm finally able to get back to it. Can you please explain how to override the inherited setting for the RTT CameraNode's StateSet? I know how to add a Uniform value to a StateSet, but I am not familiar with how to override the value for the osg::Uniform of a parent node. I know how to override or inherit values using the setMode and/or setAttributesAndModes functions of the StateSet, but I have never used them in conjunction with a shader's Uniform value. Thanks, Brian Hi Brian, You should be able to define a StateSet on the RTT Camera subgraph that contains a osg::Uniform that overrides the inherited setting. For maximum performance you might want to just use a different shader though, one that doesn't have any lighting path. Robert. Hi, I have an application that I'm developing that uses RTT to render's an image to a quad in my scene that is based on the overall scene. Best analogy that I can give is that the rendered quad is similar to a security camera. The main display consists of a drawable that has a shader program attached to it to produce lighting effects. The shader has a parameter that allows the lighting effect to be toggled. The security camera is rendered using a series of camera nodes. Each camera node has a quad, a texture, and a shader attached to the geometry node for the quad. All this works nicely. The problem that I have run into is that ideally, I would like to toggle the lighting effects shader for the RTT. Currently, the main display's shader has lighting enabled, but the lighting effects are carried into the RTT for the security camera. Ideally, I'd like to enable the main display's lighting effects, but disable them for the RTT so that I can produce more realistic effects like night vision if that is possible. One thing to keep in mind is that all lighting effects are done through the shader, but I do have a flag in the shader program that can disable the effect. I'm guessing that either node visitors or setting a cull mask might be the way to go, but I haven't had any luck with either of those methods. It's very possible that I'm going about it the wrong way. Any ideas? Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Drawables, RTT, and shaders
Hi, Sometimes I tend to forget to look at the actual function declaration in the header files. I just now noticed the second parameter that allows the override status to be set. Guess I'm just tired. Thanks, Brian Hi, It's been about a month since I looked at this, but I'm finally able to get back to it. Can you please explain how to override the inherited setting for the RTT CameraNode's StateSet? I know how to add a Uniform value to a StateSet, but I am not familiar with how to override the value for the osg::Uniform of a parent node. I know how to override or inherit values using the setMode and/or setAttributesAndModes functions of the StateSet, but I have never used them in conjunction with a shader's Uniform value. Thanks, Brian Hi Brian, You should be able to define a StateSet on the RTT Camera subgraph that contains a osg::Uniform that overrides the inherited setting. For maximum performance you might want to just use a different shader though, one that doesn't have any lighting path. Robert. Hi, I have an application that I'm developing that uses RTT to render's an image to a quad in my scene that is based on the overall scene. Best analogy that I can give is that the rendered quad is similar to a security camera. The main display consists of a drawable that has a shader program attached to it to produce lighting effects. The shader has a parameter that allows the lighting effect to be toggled. The security camera is rendered using a series of camera nodes. Each camera node has a quad, a texture, and a shader attached to the geometry node for the quad. All this works nicely. The problem that I have run into is that ideally, I would like to toggle the lighting effects shader for the RTT. Currently, the main display's shader has lighting enabled, but the lighting effects are carried into the RTT for the security camera. Ideally, I'd like to enable the main display's lighting effects, but disable them for the RTT so that I can produce more realistic effects like night vision if that is possible. One thing to keep in mind is that all lighting effects are done through the shader, but I do have a flag in the shader program that can disable the effect. I'm guessing that either node visitors or setting a cull mask might be the way to go, but I haven't had any luck with either of those methods. It's very possible that I'm going about it the wrong way. Any ideas? Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Is it possible to share textures between RenderSurface objects?
I decided to go the extra mile and create another app that basically duplicates the Producer camera method with CameraNodes. The original Producer camera application was more of a test than anything. Long story short... the two running side-by-side were nearly identical. A little finagling of the Producer version's POV have the two running side by side, and I cannot tell the difference between them. Thanks for the help. I especially am thankful for your confirming that the two methods should produce identical results. Not to mention, the CameraNode version works at a higher framerate than the Producer method. Thanks, Brian On 28 Dec 2006 03:14:39 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote : Thanks for the suggestion, but it did not work. Instead, the RTT is no longer rendered and the texture, providing any exists, appears gray. I also get the following error when trying this: RenderStage::drawInner(,) OpenGL errorNo= 0x506 which translates to INVALID_FRAMEBUFFER_OPERATION_EXT Any other thoughts? Thanks, Brian On Wed, 27 Dec 2006 15:43:17 -0800, Don Burns [EMAIL PROTECTED] wrote : Try calling Producer::RenderSurface::shareAllGLContexts(true) before you create any RenderSurfaces (or call Camera::frame() ); -don On 27 Dec 2006 23:17:37 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi, I have developed an application that uses one camera that uses camera nodes to render a scene to a texture. It works fine and has no problems (that I know of.) I have recently been asked to create another window with a camera in it that simply displays the texture rendered in my original application window. I've tried setting up the new window's camera with either a shared RenderSurface object from the original application's window or its own private RenderSurface object. In both cases, I cannot get the texture, an osg::Texture2D object created for the original camera, to display. What I wind up seeing is my scene along with a *partial* rendered scene in the original window, and in the new window, I see my quad rendered as a gray image. The gray image persists even when the texture is not added to the geometry's state set, and the partial rendered scene that had appeared in the original window also disappears. This leads me to the conclusion that the texture itself is not being seen by the second camera. Is there a way to get the second camera to recognize and correctly display a texture rendered by the first camera? The RTT does its job fine, and a quad using the rendered texture can be displayed in the original window. It's only when I add the second camera that I have problems. Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Is it possible to share textures between RenderSurface objects?
Hi Robert, It's a puzzling problem for sure. In each implementation, I have verified that all of the shaders are working. However, in the Producer version, the images produced by the shader's histogram are different than those produced by the CameraNode implementation. It appears that imagery is being lost in the CameraNode version. By lost, I mean that the textures generated in each pass don't seem to be accumulating as I would expect. They do accumulate, or use image information from a previous pass, but the outcome is different. To further complicate things, my coworker/boss is convinced that we have to go the Producer route, and he may be right for all I know. His basic reasoning is that everyone else is doing it that way. If I may ask, which method (Producer or CameraNode) is the most efficient in terms of speed, or are they about the same? If it's the CameraNode version, then I'd love to convince my coworker otherwise; and he would certainly be open to the idea. However, until I can duplicate the images using CameraNodes, that is not going to be easy. Right now, I'm just trying to appease him by finding a solution to the Producer method. If it helps you understand, we're in the process of trying to do various IR sensor effects using shaders. So, the first rendering pass might perform a noise filter, the next might be a blur filter, and then a sharpening filter. Thanks, Brian On Wed, 3 Jan 2007 19:16:45 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian, In what way is the rendering not what you expect when you use osg::CameraNode? Its all OpenGL under the hood so most likely causes are some difference in OpenGL state or the depth of the colour buffer. Robert. On 3 Jan 2007 16:54:12 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi Robert, Yes, I agree that using the CameraNode would be easier and more efficient than using the Producer. I have also implemented the code below using CameraNodes, and as far as I can tell the same shaders are being used in the same order. However, the output is clearly different. It's as though there is another step being performed by Producer that is causing the imagery to be rendered correctly. I'd like to say that my CameraNode implementation is the correct one, but I've been convinced otherwise. What I plan to do is to simplify the scenario a bit and work from there. As you've pointed out, the current implementation is a bit complex. Thanks, Brian On Wed, 3 Jan 2007 10:37:00 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian, Your set up is too complex for me to get my head around remotely. Personally I wouldn't use Producer for FBO work, the core OSG's osg::CameraNode (now named osg::Camera in CVS) has support for FBO while Producer doesn't. Please see the OSG RTT examples such as osgprerender, osgprerendercubemap, osgshadowtexture etc. Robert. On 2 Jan 2007 23:45:37 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi Robert, Thanks again for the tips, but I'm still struggling with this issue. Here's what I'm attempting at a slightly deeper level. I'm using two cameras. One is my main scene, and the other is used to do multipass rendering. The multipass render camera uses a drawable object that does nothing but clear the color and depth buffers. Attached to this multipass camera are n additional cameras and a rendering surface. Each camera has a newly allocated drawable object as described above. Associated with each additional camera (Producer::Camera) is a new FBO, Geode, texture, scene handler, and render buffer. The FBOs are set up so that the output of the previous render will be used as the input to the next rendering cycle. Each texture has its own shader that is associated with it. In the multipass camera's main rendering function, all the additional cameras are rendered to produce a final image that is output to the display. This all works fine, and hopefully what I wrote makes sense. The problem that I'm having is that because the multipass camera has its own rendering surface, it also has its own window. Since this is a secondary camera, I'd like to take the final imagery that is generated for the multipass camera and use it in the main application window. I had hoped that it would be a simple matter of attaching a quad to a geode and then attaching the secondary window's final texture to the geode's stateset. Unfortunately, the only thing rendered to the main application's geode is a gray quad, whereas the secondary window is rendering correctly. I have verified that the pipeline is rendering the multipass camera before the main window's camera. I have also tried creating a camera node in the main camera using the final texture
Re: [osg-users] Is it possible to share textures between RenderSurface objects?
Hi Robert, Yes, I agree that using the CameraNode would be easier and more efficient than using the Producer. I have also implemented the code below using CameraNodes, and as far as I can tell the same shaders are being used in the same order. However, the output is clearly different. It's as though there is another step being performed by Producer that is causing the imagery to be rendered correctly. I'd like to say that my CameraNode implementation is the correct one, but I've been convinced otherwise. What I plan to do is to simplify the scenario a bit and work from there. As you've pointed out, the current implementation is a bit complex. Thanks, Brian On Wed, 3 Jan 2007 10:37:00 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian, Your set up is too complex for me to get my head around remotely. Personally I wouldn't use Producer for FBO work, the core OSG's osg::CameraNode (now named osg::Camera in CVS) has support for FBO while Producer doesn't. Please see the OSG RTT examples such as osgprerender, osgprerendercubemap, osgshadowtexture etc. Robert. On 2 Jan 2007 23:45:37 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi Robert, Thanks again for the tips, but I'm still struggling with this issue. Here's what I'm attempting at a slightly deeper level. I'm using two cameras. One is my main scene, and the other is used to do multipass rendering. The multipass render camera uses a drawable object that does nothing but clear the color and depth buffers. Attached to this multipass camera are n additional cameras and a rendering surface. Each camera has a newly allocated drawable object as described above. Associated with each additional camera (Producer::Camera) is a new FBO, Geode, texture, scene handler, and render buffer. The FBOs are set up so that the output of the previous render will be used as the input to the next rendering cycle. Each texture has its own shader that is associated with it. In the multipass camera's main rendering function, all the additional cameras are rendered to produce a final image that is output to the display. This all works fine, and hopefully what I wrote makes sense. The problem that I'm having is that because the multipass camera has its own rendering surface, it also has its own window. Since this is a secondary camera, I'd like to take the final imagery that is generated for the multipass camera and use it in the main application window. I had hoped that it would be a simple matter of attaching a quad to a geode and then attaching the secondary window's final texture to the geode's stateset. Unfortunately, the only thing rendered to the main application's geode is a gray quad, whereas the secondary window is rendering correctly. I have verified that the pipeline is rendering the multipass camera before the main window's camera. I have also tried creating a camera node in the main camera using the final texture, FBO, and rendering surface from the multipass camera. This, too, failed. I hope that my explanation of what I'm trying to do makes sense. If not, I'd be more than happy to try and clarify. Thanks, Brian Hi Brian, I'm just about to head away for a trip, so some quick pointers. Yes you can share texture objects between separate graphics contexts in certain situations, but its isn't universally supported by graphics drivers/hardware. You can get the OSG/Producer to do it use but I can't recall the details off the top of my head, so you'll need to dig into the API of Producer. The quickest thing is probably to just have a HUD on the same window that has your RTT texture. See the osgshadowtexture example for just this type of usage. Robert. On 27 Dec 2006 23:17:37 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi, I have developed an application that uses one camera that uses camera nodes to render a scene to a texture. It works fine and has no problems (that I know of.) I have recently been asked to create another window with a camera in it that simply displays the texture rendered in my original application window. I've tried setting up the new window's camera with either a shared RenderSurface object from the original application's window or its own private RenderSurface object. In both cases, I cannot get the texture, an osg::Texture2D object created for the original camera, to display. What I wind up seeing is my scene along with a *partial* rendered scene in the original window, and in the new window, I see my quad rendered as a gray image. The gray image persists even when the texture is not added to the geometry's state set, and the partial rendered scene that had appeared in the original window also disappears. This leads me to the conclusion
Re: [osg-users] Is it possible to share textures between RenderSurface objects?
Hi Robert, Thanks again for the tips, but I'm still struggling with this issue. Here's what I'm attempting at a slightly deeper level. I'm using two cameras. One is my main scene, and the other is used to do multipass rendering. The multipass render camera uses a drawable object that does nothing but clear the color and depth buffers. Attached to this multipass camera are n additional cameras and a rendering surface. Each camera has a newly allocated drawable object as described above. Associated with each additional camera (Producer::Camera) is a new FBO, Geode, texture, scene handler, and render buffer. The FBOs are set up so that the output of the previous render will be used as the input to the next rendering cycle. Each texture has its own shader that is associated with it. In the multipass camera's main rendering function, all the additional cameras are rendered to produce a final image that is output to the display. This all works fine, and hopefully what I wrote makes sense. The problem that I'm having is that because the multipass camera has its own rendering surface, it also has its own window. Since this is a secondary camera, I'd like to take the final imagery that is generated for the multipass camera and use it in the main application window. I had hoped that it would be a simple matter of attaching a quad to a geode and then attaching the secondary window's final texture to the geode's stateset. Unfortunately, the only thing rendered to the main application's geode is a gray quad, whereas the secondary window is rendering correctly. I have verified that the pipeline is rendering the multipass camera before the main window's camera. I have also tried creating a camera node in the main camera using the final texture, FBO, and rendering surface from the multipass camera. This, too, failed. I hope that my explanation of what I'm trying to do makes sense. If not, I'd be more than happy to try and clarify. Thanks, Brian Hi Brian, I'm just about to head away for a trip, so some quick pointers. Yes you can share texture objects between separate graphics contexts in certain situations, but its isn't universally supported by graphics drivers/hardware. You can get the OSG/Producer to do it use but I can't recall the details off the top of my head, so you'll need to dig into the API of Producer. The quickest thing is probably to just have a HUD on the same window that has your RTT texture. See the osgshadowtexture example for just this type of usage. Robert. On 27 Dec 2006 23:17:37 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi, I have developed an application that uses one camera that uses camera nodes to render a scene to a texture. It works fine and has no problems (that I know of.) I have recently been asked to create another window with a camera in it that simply displays the texture rendered in my original application window. I've tried setting up the new window's camera with either a shared RenderSurface object from the original application's window or its own private RenderSurface object. In both cases, I cannot get the texture, an osg::Texture2D object created for the original camera, to display. What I wind up seeing is my scene along with a *partial* rendered scene in the original window, and in the new window, I see my quad rendered as a gray image. The gray image persists even when the texture is not added to the geometry's state set, and the partial rendered scene that had appeared in the original window also disappears. This leads me to the conclusion that the texture itself is not being seen by the second camera. Is there a way to get the second camera to recognize and correctly display a texture rendered by the first camera? The RTT does its job fine, and a quad using the rendered texture can be displayed in the original window. It's only when I add the second camera that I have problems. Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Use CEGUI with OSG?
I use CEGUI as well and have seen problems of this sort. My application uses multitexturing, so this may not apply to what you are doing. I discovered that the problem was usually solved by doing the following: In the call to OpenGLRenderer::initPerFrameStates(void), I found that I needed to call the following functions before rendering. The glActiveTextureARB call is an extension function, and the texid is the id of the texture being used. // put the call to glActiveTextureARB after the calls to the following // functions glPushClientAttrib(GL_CLIENT_ALL_ATTRIB_BITS); glPushAttrib(GL_ALL_ATTRIB_BITS); glActiveTextureARB(GL_TEXTURE0); This will help to ensure that the active texture is reset to GL_TEXTURE0, which CEGUI uses by default, in the event that some multi-texturing aspect of your program has it set to some other value by the time OpenGLRenderer::doRender is called. I have seen other situations that cause crazy texturing problems like you spoke of even with these changes, however they were generally solved the same way. For instance, if you have drawable objects, you might need to push and pop the attributes and reset the active texture. This may not be the most effecient method, but it works for my application. Incidentally, check out CEGUI's web site/forum for answers to a lot of questions. Also, the IRC channel, #CEGUI on irc.freenode.net, is a great source for help. This is really more of a CEGUI-specific question, but hopefully my suggestions will help you in the meantime. -Brian On Fri, 29 Dec 2006 21:20:14 +0800 (CST), lin [EMAIL PROTECTED] wrote : How can I use CEGUI with OSG? We integrate CEGUI for HUDs, In the application, sometimes the HUD is right, sometimes the HUD's texture is wrong, sometimes the HUD disappears. can somebody give me a demo? I have saw the message:adaptation of CEGUI and OSG at: ftp://yag2002.homeftp.org/VRC/VRC-sources-July-31-2005.zip,but I can't download it. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Is it possible to share textures between RenderSurface objects?
Hi, I have developed an application that uses one camera that uses camera nodes to render a scene to a texture. It works fine and has no problems (that I know of.) I have recently been asked to create another window with a camera in it that simply displays the texture rendered in my original application window. I've tried setting up the new window's camera with either a shared RenderSurface object from the original application's window or its own private RenderSurface object. In both cases, I cannot get the texture, an osg::Texture2D object created for the original camera, to display. What I wind up seeing is my scene along with a *partial* rendered scene in the original window, and in the new window, I see my quad rendered as a gray image. The gray image persists even when the texture is not added to the geometry's state set, and the partial rendered scene that had appeared in the original window also disappears. This leads me to the conclusion that the texture itself is not being seen by the second camera. Is there a way to get the second camera to recognize and correctly display a texture rendered by the first camera? The RTT does its job fine, and a quad using the rendered texture can be displayed in the original window. It's only when I add the second camera that I have problems. Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Drawables, RTT, and shaders
Hi, I have an application that I'm developing that uses RTT to render's an image to a quad in my scene that is based on the overall scene. Best analogy that I can give is that the rendered quad is similar to a security camera. The main display consists of a drawable that has a shader program attached to it to produce lighting effects. The shader has a parameter that allows the lighting effect to be toggled. The security camera is rendered using a series of camera nodes. Each camera node has a quad, a texture, and a shader attached to the geometry node for the quad. All this works nicely. The problem that I have run into is that ideally, I would like to toggle the lighting effects shader for the RTT. Currently, the main display's shader has lighting enabled, but the lighting effects are carried into the RTT for the security camera. Ideally, I'd like to enable the main display's lighting effects, but disable them for the RTT so that I can produce more realistic effects like night vision if that is possible. One thing to keep in mind is that all lighting effects are done through the shader, but I do have a flag in the shader program that can disable the effect. I'm guessing that either node visitors or setting a cull mask might be the way to go, but I haven't had any luck with either of those methods. It's very possible that I'm going about it the wrong way. Any ideas? Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] link error: RenderSurface_X11.obj : error LNK2001: unresolved external symbol
I am using the stable zipped download from main osg website. I have checked my options settings but everything looks ok so Im at a loss atm. I will keep looking for something I've missed . cheers neighborlee --- Hi, What version of Producer are you using? I haven't seen any reports of similiar problems before so it sounds like either a new problem in CVS or that your install/compile has screwed up for some reason. ReaderSurface_X11.cpp if compiled at all should just compile to an empty file in anything other than Unix. Robert. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] link error: RenderSurface_X11.obj : error LNK2001: unresolved external symbol
I am using vc++express 2005 atm and getting this error during compile ( OT compiled just fine in debug) of producer in debug mode: RenderSurface_X11.obj : error LNK2001: unresolved external symbol I think I may have delted something from vc++ options and while I reinstalled it did not help. thx anyone nl ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Creating an Image using GL_LUMINANCE and GL_FLOAT
Hi,
I'm trying to generate an image using GL_LUMINANCE and GL_FLOAT, however no
matter what values (ranging from 0 to 1.0) I use in my image buffer, I
always get an image that has RGB values of 205,205,205. The number 205
is typically used when the debugger initializes variables, so this leads me
to believe that my Image object is incorrectly set up.
If I write the image to a file or simply attach it to a Texture2D object, I
only see a pure gray image.
osg::Image* NoiseImage = new osg::Image();
// generate random values between 0 and 1
for (int i=0; iNoiseBufferSize; i++)
{
g_NoiseImageF[i] = (float)(rand() % 4096) / 4096.0f;
}
NoiseImage-setImage(width,
height,
1,
GL_LUMINANCE,
GL_LUMINANCE,
GL_FLOAT,
(unsigned char*) g_NoiseImageF,
osg::Image::NO_DELETE);
osgDB::writeImageFile(*NoiseImage, c:/noise.bmp);
I've looked at various examples in the OSG examples. One example set the
Image's internal texture format to GL_LUMINANCE_FLOAT32_ATI. I tried this,
but it simply hung my application. I tried using GL_LUMINANCE for the
image's texture format, but that had no effect either. Any help would be
most appreciated.
Thanks,
Brian
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
Re: [osg-users] Multiple shaders and RTT
Hi Robert, I understand about reviewing the code. I only included it in case it might prove helpful. Anyway, while I haven't gotten this to work just yet, I have managed to confirm a few things. Separately, both CameraNodes can be added and render their respective textures fine, so I know that they are set up correctly in that regard. I've set up a quad osg::geometry and have added it to the camera node and have associated a texture with it for test purposes. CameraNode1 - quad with test texture | |-- CameraNode2 - cow node attached When I render my final quad using CameraNode1's RTT texture, it correctly displays the rendered textured quad. However, I am assuming that in order to render the quad's texture with CameraNode2's rendered scene, I would need to use the texture associated with CameraNode2's texture that is attached to its FBO for RTT. CameraNode1 - quad with CameraNode2's rendered texture | |-- CameraNode2 - cow node attached When I try this approach, I see a black square. Both camera nodes are set to PRE_RENDER. Thanks, Brian On Wed, 22 Nov 2006 09:04:58 +, Robert Osfield [EMAIL PROTECTED] wrote : Hi Brian, Doing code reviews of user code is bit beyond what I have time for. I'd recommend getting the nested RTT camera working as you want first, without the enclosing/parent RTT camera. Then once you are happy with this then add the enclosing RTT camera. Robert. On 22 Nov 2006 00:46:56 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi Robert, One quick question. Applogies I didn't quite differentiate the different options properly. A NESTED_RENDER woulnd't generally be used for RTT, as all the drawing goes into the current RenderStage, rather than have then drop into a RenderStage drawn before the enclosing RenderStage. CameraNode - PRE_RENDER - one | |-- CameraNode - PRE_RENDER - two | |-- Cow node. I'm thinking that I should set up CameraNode 1 as an ortho 2D camera. However, I'm having a dumb moment and cannot seem to get it set up correctly. I set up my CameraNodes like: osg::ref_ptrosg::FrameBufferObject fbo1 = new osg::FrameBufferObject(); fbo1-setAttachment(GL_COLOR_ATTACHMENT0_EXT, osg::FrameBufferAttachment(CameraTexture1)); osg::RenderBuffer *renderBuffer1 = NULL; renderBuffer1 = new osg::RenderBuffer(512, 512, GL_DEPTH_COMPONENT24))); fbo1-setAttachment(GL_DEPTH_ATTACHMENT_EXT, osg::FrameBufferAttachment(renderBuffer1)); CameraNode1 = new osg::CameraNode; CameraNode1-setClearColor(osg::Vec4(0.f, 0.5f, 0.f, 1.f)); CameraNode1-setClearMask(GL_COLOR_BUFFER_BIT); CameraNode1-setViewport(0, 0, width, height); // set up projection. CameraNode1-setProjectionMatrixAsOrtho2D(0.0, width, 0.0, height); // set view CameraNode1-setReferenceFrame(osg::Transform::ABSOLUTE_RF); // set the camera to render before the main camera. CameraNode1-setRenderOrder(osg::CameraNode::PRE_RENDER); osg::CameraNode::RenderTargetImplementation renderImplementation; renderImplementation = osg::CameraNode::FRAME_BUFFER_OBJECT; CameraNode1-setRenderTargetImplementation(renderImplementation); CameraNode1-attach(osg::CameraNode::COLOR_BUFFER, CameraTexture1); and CameraNode2 like osg::ref_ptrosg::FrameBufferObject fbo2 = new osg::FrameBufferObject(); fbo2-setAttachment(GL_COLOR_ATTACHMENT0_EXT, osg::FrameBufferAttachment(CameraTexture2)); osg::RenderBuffer *renderBuffer2 = new osg::RenderBuffer(512, 512, GL_DEPTH_COMPONENT24)) fbo2-setAttachment(GL_DEPTH_ATTACHMENT_EXT, osg::FrameBufferAttachment(renderBuffer2)); CameraNode2 = new osg::CameraNode; CameraNode2-setClearColor(osg::Vec4(1.f, 0.f, 0.f, 1.f)); CameraNode2-setClearMask(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT); CameraNode2-setViewport(0, 0, width, height); CameraNode2-setComputeNearFarMode( osg::CullSettings::DO_NOT_COMPUTE_NEAR_FAR); CameraNode2-setProjectionMatrixAsPerspective( 30.0, 1.0, 0.1, 5.0); // set view CameraNode2-setReferenceFrame(osg::Transform::ABSOLUTE_RF); // set the camera to render before the main camera. CameraNode2-setRenderOrder(osg::CameraNode::PRE_RENDER); //osg::CameraNode::RenderTargetImplementation renderImplementation; renderImplementation = osg::CameraNode::FRAME_BUFFER_OBJECT; CameraNode2-setRenderTargetImplementation(renderImplementation); CameraNode2-attach(osg::CameraNode::COLOR_BUFFER, CameraTexture2); I then add CameraNode2 to CameraNode1 like CameraNode1-addChild(CameraNode2); I then add the main scene (a cow for now) node to CameraNode2 like CameraNode2-addChild(CowNode); If I
Re: [osg-users] Preframe callbacks with CameraNode
Hi, I have been trying to disable lighting in my RTT CameraNode. Well, I have discovered that it is not a true light that I am concerned with here. The lighting is being implemented through a shader that a coworker wrote, so the stateset management has been working all along. However, since the shader has no knowledge of this, it continues to produce lighting effects. While I am using Delta3D as my rendering package, the core OSG is exposed to me. I have set up a preframe callback for my main Producer camera, and I'm using a node visitor with an update callback for the CameraNode. Closer inspection of my code (it's been several months since I set it up originally) reveals that my rendering is set up like CameraNode Producer camera - scene - CameraNode - drawables No RTT Producer camera - scene - drawables I'm using osgNV to modify the shader uniform for disabling the light shader, and while both the preframe callback and node visitor update callback are hit, only the changes made by the preframe callback are taken into account. Setting the light uniform in the preframe callback to render renders the light for both cameras even though I'm saying otherwise in the node visitor update callback. I'm obviously going about this the wrong way, but I don't know what other way to go. I've also tried setting up a frame buffer object update callback like: osg::FrameBufferObject *fbo = new osg::FrameBufferObject(); // actually, fbo is a ref pointer, but to help it fit on one line // I removed that code. fbo-setUpdateCallback(new FboCallback(true)); This had absolutely no effect. The callback's () operator was never hit. Any help would be greatly appreciated. Thanks, Brian ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Preframe callbacks with CameraNode
Hi Robert, I realize that enabling or disabling of GL_LIGHTING will have no effect in regard to the shaders. I have implemeneted a way to change a uniform so that lighting can be toggled. My initial attempt at toggling the light uniform was with callbacks. This, as you pointed out, will not work. I know enough OSG to be dangerous, but that is about it. How can I use the OSG standard state inheritance mechanism to help me toggle the light uniform? Thanks, Brian HI Brian, If shaders are being used for lighting then all you can do is change uniforms or change the shaders themelves, no enabling or disabling of the GL_LIGHTING etc will make a difference. Putting in callbacks won't help either - you have to substitute the shaders being used, and the OSG's standard state inheritance mechanism or just traversing through the scene will help. Robert. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Preframe callbacks with CameraNode
Hi Robert, You had mentioned substituting the shaders via scene traversal or state inheritance. I know how to enable/disable/substitute (perhaps cripple) the shader. My problem is setting the scene up properly (either through scene traversal or state inheritance) such that the RTT's CameraNode is crippled while the main Producer camera's shader is fully intact. Basically, I just don't know how to implement the state inheritance or scene traversal aspects such that my crippling and enabling functions (that I have written) can be called prior to the rendering of their respective cameras. Thanks, Brian Hi Brian, Gees you really need to talk to the guys who set up the shaders about this, I can't comment on stuff I have nothing to do with... Robert. On 3 Nov 2006 17:27:14 -, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi Robert, I realize that enabling or disabling of GL_LIGHTING will have no effect in regard to the shaders. I have implemeneted a way to change a uniform so that lighting can be toggled. My initial attempt at toggling the light uniform was with callbacks. This, as you pointed out, will not work. I know enough OSG to be dangerous, but that is about it. How can I use the OSG standard state inheritance mechanism to help me toggle the light uniform? Thanks, Brian HI Brian, If shaders are being used for lighting then all you can do is change uniforms or change the shaders themelves, no enabling or disabling of the GL_LIGHTING etc will make a difference. Putting in callbacks won't help either - you have to substitute the shaders being used, and the OSG's standard state inheritance mechanism or just traversing through the scene will help. Robert. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Preframe callbacks with CameraNode
Hi Robert,
I have tried that thinking that it should work, but for whatever reason, it
did not.
I used code like:
/
osg::StateSet *stateset = cameraNode-getOrCreateStateSet();
if (stateset)
{
stateset-setMode(GL_LIGHTING,
osg::StateAttribute::OVERRIDE|osg::StateAttribute::OFF);
stateset-setMode(GL_LIGHT0, osg::StateAttribute::OFF);
cameraNode-setStateSet(stateset);
}
//
Is this what you meant? Perhaps I am calling it in the wrong place. When
is the best time to call these functions?
Thanks,
Brian
Hi Brian,
To disable the light for the subgraph under the CameraNode it should
be sufficient to attach a StateSet to the CameraNode and disable the
lighting mode there, overriding the mode for the whole subgraph below.
There shouldn't be any need to using a callback for this.
Robert.
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
Re: [osg-users] Preframe callbacks with CameraNode
Hi Robert, I appreciate your helping me with this problem. Perhaps a different view will help here. In the Producer::Camera I have camera-scene-light_drawable-geometry_drawables In my RTT camera node I have cameraNode-scene-light_drawable-geometry_drawables Attached to the cameraNode is a FBO. If I disable the lighting in the cameraNode, and the light_drawable re- enables it, could that be my problem? I'm using a 3rd party tool that is based on OSG (Delta3D) so I have no option but to use the light drawable. My thought was to intercept the rendering by using preframe callbacks: one for the Producer camera and one for the cameraNode. In these callbacks, I could modify the state of the light_drawable. My initial thought was to remove the drawable in the cameraNode preframe callback and add it back in the Producer::Camera preframe callback, but I fear that by doing that, my scene would change to: camera-scene-geometry_drawables-light_drawable and I don't know what effect that would have on the rendering of my scene. Of course, I can't figure out how to get a preframe callback for the cameraNode, so I cannot test this theory. I tried using a node visitor and removing the light drawable in it, but that caused bad things to happen (i.e. program crashed.) I think that removing the drawable in the node visitor's update callback caused the node traversal to break. Thanks again. Brian Hi Brian, From the osgtext example you'll find : camera-getOrCreateStateSet()-setMode (GL_LIGHTING,osg::StateAttribute::OFF); I have just doubled checked that its functioning by commenting this line out and indeed the lighting goes on and problems occur with rendering. Robert. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Preframe callbacks with CameraNode
Hi Robert, That's ok. I have verified using the osgviewer application that disabling the light (pressing 'L') works with the terrain data that I am using. That alone makes me feel sane. This is obviously a problem either with the way I am using Delta3D or with Delta3D itself. Thanks for your help. Brian Hi Brian, I'm afriad I can't support custom stuff that Delta3D is doing. If Delta3D manages OpenGL state they way the OSG is designed to handle state then the trick that I and Don suggested, and what you have tried, should work, the fact it doesn't suggest a problem outwith the OSG. Robert. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] KDevelop and OSG
I was wondering if anyone has got a workingOSG/KDevelop project up and running. I seem to have trouble - mainly I think the extentionless header problem. Isthis solvedby simplyadding extentions onto the headers ? Kind Regards Paul ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Compilation problem with ungif/giflib
Andreas Goebel wrote: [EMAIL PROTECTED] schrieb: Hello, I tried to compile OpenSceneGraph, but it failed with error: g++ -O2 -L/usr/X11R6/lib64 -L/usr/freeware/lib64 -L../../../../lib/Linux64 -L/home/tuuli/bin/OSG_OP_OT-1.2/OpenThreads/lib/Linux64 -L/home/tuuli/bin/OSG_OP_OT-1.2/Producer/lib/Linux64 -shared ReaderWriterGIF.o -lstdc++ -losgGA -losgDB -losgUtil -losg -lungif -lOpenThreads -o osgdb_gif.so /usr/lib/gcc/x86_64-pc-linux-gnu/4.1.1/../../../../x86_64-pc-linux-gnu/bin/ld: cannot find -lungif collect2: ld returned 1 exit status gmake[4]: *** [osgdb_gif.so] Error 1 I don't have ungif installed, but giflib 4.1.4. How can I resolve this without installing ungif? I'm running Gentoo with gcc 4.1.1 and glibc 2.4-r3. Thanks. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ Hi, I think you could go to the make folder and edit makedefs, there is a section giflib. Be sure to change it for your platform. I hope it´s right, I don´t have it here at the moment so I can´t look. Regards, Andreas That worked, thanks! ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Terrain Databases ?
Does anyone know where you can download terrain databases for particular regions of Europe (France/Germany) ? I'm sorry but I have no experience with paged database terrains, I've ben trying to create terrain using Demeter ( because it offersgreater control over texturing and texture splatting) but Demeterwont compile againstOSG version 1.0. I think osgDem/osgTerrain is probably the way to go - but does it offer the kind of multitexturing/texture splatting like Demeter ? or is that texture information given to you in the database ? like an Overall_texture for the given terrain. Thanks for your time. Paul ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] RenderBin Error
Thanks for the responses. Firstly, I have disabled all of my own GPU shader code. This didn't cause any change in the error. The scene consists simply of a single object and a single terrain database. Removing the terrain database does get rid of the problem, but that also doesn't seem to help much seeing as it is a complex scene node.The program has been set up based on the OSG examples for running FBO's, shadowing, and terrain paging. I tried to follow these examples as closely as possible when creating the scene. I guess I am most interested in the OSG changes that could have caused this error to be found between v 1.1 and v1.2. Identical code under v1.1 doesn't show this error (99% of the time, anyway). One peculiar thing did start to show up which might lend a clue. I am rendering the entire scene on the FBO and then using another basic GLUT node to draw a graph about performance of the app. The app saves the contents of the FBO out to bitmaps. From time to time, the screen buffer is getting saved to the bitmaps. Nothing consistent, but it does happen. I am tending to think this is related to the error but I am unsure how to narrow this down any further.I appreciate how hard it is to debug such a vague error through email. If nothing else, how can I disable this error?Robert Osfield [EMAIL PROTECTED] wrote: Hi,The OpenGL error could be caused by many different errors in the scene graph setup so I can't really guess at what might be wrong. I'd recommend simplifying your scene and till the error goes away, then you should be better able to pin point the problem. Also check the RTT examples like osgprerender and osgshadowtexture etc, do they produce errors too? This shouldn't produce errors, and if they do it suggest so problem with OpenGL driver support.Robert. On 9/14/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: I upgraded to OSG 1.2 today. I am now consistently getting the following error message whenever I use FBO's: "Warning: detected OpenGL error 'invalid operation' after RenderBin::draw(,)" I would get this occasionally (1% of the time) with v1.1 but now I am getting it EVERY time with v1.2. I assume there is something that I am doing wrong, but as I "see" no problems with the rendered scene, I was wondering what types of operations could be causing this or what I could do to attempt to track it down. Particularly, what operations would cause an error on the FBO that wouldn't cause an error for a normally rendered scene? Were there any changes to the RenderBin which would cause v1.2 to catch an error that v1.1 would not consistently catch? Any help would be appreciated! Thanks!Do you Yahoo!? Get on board. You're invited to try the new Yahoo! Mail. ___osg-users mailing listosg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___osg-users mailing listosg-users@openscenegraph.nethttp://openscenegraph.net/mailman/listinfo/osg-usershttp://www.openscenegraph.org/ Talk is cheap. Use Yahoo! Messenger to make PC-to-Phone calls. Great rates starting at 1¢/min.___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] RenderBin Error
I upgraded to OSG 1.2 today. I am now consistently getting the following error message whenever I use FBO's: "Warning: detected OpenGL error 'invalid operation' after RenderBin::draw(,)" I would get this occasionally (1% of the time) with v1.1 but now I am getting it EVERY time with v1.2. I assume there is something that I am doing wrong, but as I "see" no problems with the rendered scene, I was wondering what types of operations could be causing this or what I could do to attempt to track it down. Particularly, what operations would cause an error on the FBO that wouldn't cause an error for a normally rendered scene? Were there any changes to the RenderBin which would cause v1.2 to catch an error that v1.1 would not consistently catch? Any help would be appreciated! Thanks! Do you Yahoo!? Get on board. You're invited to try the new Yahoo! Mail.___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Does anyone know how to use osgShadow nodekit ?
hi,all Where I can get a sample about osgShadow nodekit? thanks. lilinx [EMAIL PROTECTED] 2006-09-03 ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] New version of Blender exporter
Rubén López wrote: Hi, I've just uploaded a new version of the blender exporter with lots of enhancements by Ben Discoe, Alessandro Pirovano and Vincent Vivanloc. In fact I've released two consecutive versions: 2.42 and 2.42b, the second one has also a progressbar :) Please, check it here: https://projects.blender.org/projects/osgexport/ ... and send me any bug report, feature request or patch that you want. I want to say thanks to the contributors. This version wouldn't be possible without them. we have a problem http://csp.sourceforge.net/forum/viewtopic.php?p=3435#3435 guess it has more to do with the fact that transform matrices are not cleaned up before the export. I ran into that kind of issue with the animation module of blender. kind regards ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Can't find removeDrawables in osg.dll dynamic lib error message?
Hi folks, I had the same problem because I also use Delta3D. At the moment I removed all Delta-Files but is there another possibility to use both at the same time or do I have to wait until the Delta3D-Team changed to OSG 1.1? Kerstin. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Animation Path
Hello. Here is one question- How to switch to some time label in animation path. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] Queston about animation path
Hello.
Here is one question-
How to switch to some time label in animation path.
And another one-
There is a FAQ about exporter. One of the questions is:
Q: How do I extract the animations, cameras, switches etc. from the exported
file.
A: The following code should do it:
std::vector(osg::PositionAttitudeTransform*) _cameras;
std::vector(osg::Switch*) _switches;
std::vector(osg::AnimationPathCallback*) _animations;
/**
* Find all special nodes in the scenegraph.
*/
void Previewer::findSpecialNodes(osg::Node* node){
// Find animations
if(dynamic_cast(osg::AnimationPathCallback*) (node-getUpdateCallback())){
_animations.push_back(dynamic_cast(osg::AnimationPathCallback*)
(node-getUpdateCallback())-getAnimationPath());
}
// Find cameras
if(dynamic_cast(osg::PositionAttitudeTransform*) (node)){
_cameras.push_back(static_cast(osg::PositionAttitudeTransform*) (node));
}
// Find switches
else if(dynamic_cast(osg::Switch*) (node)){
_switches.push_back(static_cast(osg::Switch*) (node));
}
// Traverse any group node
else if(dynamic_cast(osg::Group*) (node)){
osg::Group* group = static_cast(osg::Group*) (node);
for(unsigned int i=0; igetNumChildren(); i++)
findSpecialNodes(group-getChild(i));
}
}
-
Why doesn't it work? What is incorrect?
Thanx.
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
[osg-users] AnimationPath help
I▓ve got the following question:
How do I extract the animations from the exported (OSG)file.
The code (FAQ OsgExp for 3DMAX) not work:
What exactly should be changed?
std::vector(osg::PositionAttitudeTransform*) _cameras;
std::vector(osg::Switch*) _switches;
std::vector(osg::AnimationPathCallback*) _animations;
/**
* Find all special nodes in the scenegraph.
*/
void Previewer::findSpecialNodes(osg::Node* node){
// Find animations
if(dynamic_cast(osg::AnimationPathCallback*) (node-getUpdateCallback())){
_animations.push_back(dynamic_cast(osg::AnimationPathCallback*)
(node-getUpdateCallback())-getAnimationPath());
}
// Find cameras
if(dynamic_cast(osg::PositionAttitudeTransform*) (node)){
_cameras.push_back(static_cast(osg::PositionAttitudeTransform*) (node));
}
// Find switches
else if(dynamic_cast(osg::Switch*) (node)){
_switches.push_back(static_cast(osg::Switch*) (node));
}
// Traverse any group node
else if(dynamic_cast(osg::Group*) (node)){
osg::Group* group = static_cast(osg::Group*) (node);
for(unsigned int i=0; igetNumChildren(); i++)
findSpecialNodes(group-getChild(i));
}
}
___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
[osg-users] TXPNode::updateEye()
I am running OSG v1.1 with two sceneviews. Each of the sceneviews is looking at the same scene, and the scene contains a txp node. When I switched from OSG v1.0 to v1.1, I started getting this error on the second sceneView. TXPNode::updateEye() no pageManager created The error seems to be printing out during the cull on that sceneView. The sceneview printing out this error is drawing the entire scene except the txp node, while the first sceneview is drawing the txp node just fine. (To summarize: I am running two SceneViews simultaneously. They are each looking at a scene containing a txp node and other objects. Only the first sceneview is drawing the txp node. The second is not drawing the txp node and is producing the error.) I tried to modify the osgSimplePager to reproduce the problem, but it seemed to run two sceneViews without a problem. I have come to the conclusion that it is surely a problem in my code, but I can't track it down. I seem to be making all the same calls that osgSimplePager uses and I seem to be making them in the same order. Can anyone give me some ideas as to what might cause this error? I am running a version 2.x Terrex database and the same code/database worked fine under OSG v1.0 with an updated TXP loader. Thanks! Talk is cheap. Use Yahoo! Messenger to make PC-to-Phone calls. Great rates starting at 1¢/min.___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] TXPNode::updateEye()
Thanks for the quick response. Unfortunately, as I am already assigning unique contextID's for each sceneView, that is not the cause of my trouble here. It is something relating to the latest version of the txp loader, I believe. The CVS version of the txp loader I was using prior to the 1.1 release was working fine and not producing this error. However, I could see how setting up the scene graph wrong might cause a problem as well. I just don't understand the txp loader well enough to track down this problem.Any additional thoughts?Robert Osfield [EMAIL PROTECTED] wrote: Hi,The most common problem with multiple SceneView is not incrementingthe contextID for each seperate graphics contexts, see the FAQ.I can't comment much on the internals of the TXP loader though, it hasbeen updated for the 1.1 release.Robert.On 7/22/06, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: I am running OSG v1.1 with two sceneviews. Each of the sceneviews is looking at the same scene, and the scene contains a txp node. When I switched from OSG v1.0 to v1.1, I started getting this error on the second sceneView. TXPNode::updateEye() no pageManager created The error seems to be printing out during the cull on that sceneView. The sceneview printing out this error is drawing the entire scene except the txp node, while the first sceneview is drawing the txp node just fine. (To summarize: I am running two SceneView's simultaneously. They are each looking at a scene containing a txp node and other objects. Only the first sceneview is drawing the txp node. The second is not drawing the txp node and is producing the error.) I tried to modify the osgSimplePager to reproduce the problem, but it seemed to run two sceneViews without a problem. I have come to the conclusion that it is surely a problem in my code, but I can't track it down. I seem to be making all the same calls that osgSimplePager uses and I seem to be making them in the same order. Can anyone give me some ideas as to what might cause this error? I am running a version 2.x Terrex database and the same code/database worked fine under OSG v1.0 with an updated TXP loader. Thanks! Talk is cheap. Use Yahoo! Messenger to make PC-to-Phone calls. Great rates starting at 1�/min. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/___osg-users mailing listosg-users@openscenegraph.nethttp://openscenegraph.net/mailman/listinfo/osg-usershttp://www.openscenegraph.org/ Yahoo! Music Unlimited - Access over 1 million songs. Try it free. ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
[osg-users] FBO and Culling
I am having some problems with the FBO working as I would have expected. I have included some code snippets to help. It seems that culling an FBO node results in some kind of corruption of the node. Let me explain what I am doing:I create a Texture2D that is of size (4096, 4096). This is obviously larger than my real frame buffer. (Hence the need for the FBO.) texture-setTextureSize(4096, 4096); texture-setInternalFormat(GL_RGB); texture-setFilter(osg::Texture2D::MIN_FILTER,osg::Texture2D::LINEAR); texture-setFilter(osg::Texture2D::MAG_FILTER,osg::Texture2D::LINEAR);I next create an osgCameraNode. I set the viewport to the same size as the texture. camera-setClearColor(osg::Vec4(0.1f,0.1f,0.1f,1.0f));
camera-setClearMask(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT); camera-setReferenceFrame(osg::Transform::ABSOLUTE_RF); camera-setViewport(0, 0, 4096, 4096); camera-setRenderOrder(osg::CameraNode::PRE_RENDER); camera-setRenderTargetImplementation(osg::CameraNode::RenderTargetImplementation::FRAME_BUFFER_OBJECT);Then, I finish setting up my FBO, establishing who is who's child, etc. image-allocateImage(4096, 4096, 1, GL_RGB, GL_UNSIGNED_BYTE); camera-attach(osg::CameraNode::COLOR_BUFFER, this-_pBufferImage); texture-setImage(0, this-_pBufferImage); camera-addChild(sceneManager-getScene().get()); camera-setProjectionMatrixAsPerspective(this-getFieldOfView(),
1.4, 10.0, 4.0);Now, I have other camera's running while this FBO exists. Having a FBO running of this size really kills performance, so I wanted a way to "turn off" the use of this FBO. I created a culling callback and added it. Then I added the whole thing to my master scene node. fboParent-setCullCallback(this-_fboCullCallback); fboParent-addChild(this-camera); masterSceneNode-addChild(this-fboParent);I implemented the culling callback as:CullCallback::CullCallback() : osg::NodeCallback(){ this-_skipCallback = false;} void CullCallback::operator()(osg::Node* node, osg::NodeVisitor* nv){ if(this-_skipCallback == false) { traverse(node,
nv); }}void CullCallback::setSkipCull(bool x){ this-_skipCallback = x;} Ok, now for the problem: If I leave everything alone, the FBO works fine. If I "setSkipCull(true)", then the FBO doesn''t render, as expected. The problem comes when I "setSkipCull(false)" after setting it to TRUE. The FBO does render, but only the size of the current viewport, not the original 4096x4096. The rest of the buffer is black. I guess I don't understand why skipping the cull on the node would result in this behavior. Is there a better way to turn the FBO rendering on and off? Is this a problem in the osg code or my code? I have tried to re-initialize the values for the FBO, its texture, etc, but it doesn't have any effect. Any input would be greatly
appreciated! __Do You Yahoo!?Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___
osg-users mailing list
osg-users@openscenegraph.net
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
[osg-users] 1 process, 2 viewers
Hello, Is it possible to run in the same process (same application) but on two separate threads, two osgProducer::Viewer instances (1 scene graph/viewer) ? I did some trials and it works more or less but: - it happens that text objects (osgText::Text) have no texture - for some reason large LODs like bluemarble are not working. Thanks, Antoine - ALICE SECURITE ENFANTS - Protégez vos enfants des dangers d'Internet en installant Sécurité Enfants, le contrôle parental d'Alice. http://www.aliceadsl.fr/securitepc/default_copa.asp ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] 1 process, 2 viewers
Hello Robert, Thanks for the answer. Actually, getting two osg::State is what I am looking for. One of the viewer is a true viewer displaying a scene graph the other is running a parrallel terrain generation. The latter should not break the first. I have seen that osgTerrain::DataSet has a setState(osg::State*) method but was not able to use it properly. Is there a way to create an independent graphic context without going for a full osgProducer::Viewer. I also had a look at the osgcamera example but it currently does not work on my machine (Error while redimensionning texture because the OpenGL context is not valid) and it is written in the source code that it is not stable yet. It looks like another level of abstraction for the OSG runtime and camera without using Producer. Is that right? Antoine -- Initial Header --- From : [EMAIL PROTECTED] To : osg users osg-users@openscenegraph.net Cc : Date : Fri, 23 Jun 2006 14:42:32 +0100 Subject : Re: [osg-users] 1 process, 2 viewers Hi Antoine, Is there a reason why you can't use two Camera's in a singe viewer, or just use an OsgCameraGroup and manage things yourself. Also do the viewers share the same graphics context? If not then you'll need to ensure they use the seperate osg::State objects for each context otherwise the OSG will only have one buffer of display lists/texture objects for two seperate graphis context and will certainly prodcue artifacts like you describe. It should be possible to modify things, but having two seperate viewers running sequnetially is a bit of hack. However it'd be better to have just one viewer with multiple windows and let osgProducer automatically manage the seperate osg::State for you. Robert. On 6/23/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hello, Is it possible to run in the same process (same application) but on two separate threads, two osgProducer::Viewer instances (1 scene graph/viewer) ? I did some trials and it works more or less but: - it happens that text objects (osgText::Text) have no texture - for some reason large LODs like bluemarble are not working. Thanks, Antoine - ALICE SECURITE ENFANTS - Protégez vos enfants des dangers d'Internet en installant Sécurité Enfants, le contrôle parental d'Alice. http://www.aliceadsl.fr/securitepc/default_copa.asp ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ - ALICE SECURITE ENFANTS - Protégez vos enfants des dangers d'Internet en installant Sécurité Enfants, le contrôle parental d'Alice. http://www.aliceadsl.fr/securitepc/default_copa.asp ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Terrapage 2.2 support in OSG
Is there a way to tell that this pager is done? That is: There is no more work for the pager to do until the camera moves again? I am interested in querying the "complete" state of the pager. Any ideas?Andrew Sampson [EMAIL PROTECTED] wrote: Terrex made a lot of changes (especially differences in whitespace), so the diff between the forked version and OSG 1.0's txp plugin is yucky. I did my best, but keep in mind that this is the result of a 4-way merge (kinda... terrex-0.9.8, stock-0.9.8, stock-0.9.9, stock-1.0). I guess I should clarify: performing a diff betweenOSG 1.0's txp plugin and the forked version is yucky(lots of whitespace changes, large chunks of new code,etc). The code itself is quite solid. It's beentested with just about every database I haveavailable, and has been stable so far. It actuallyseems a little quicker than the stock 1.0 code, but Icould be imagining that.-Andrew__Do You Yahoo!?Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___osg-users mailing listosg-users@openscenegraph.nethttp://openscenegraph.net/mailman/listinfo/osg-usershttp://www.openscenegraph.org/ __Do You Yahoo!?Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] Producer Texture Access
Will the glReadPixels work from the pBufferCamera? I guess it wasn't clear to me that I could read from the pbuffer with that call. How would glReadPixels know to read data from the Producer PBuffer Camera vs. the rendered Frame Buffer? From the documentation for glReadPixels: "glReadPixels returns pixel data from the frame buffer, starting with the pixel whose lower left corner is at location (x, y), into client memory starting at location pixels." The reason I had used the producer camera is because I needed to render an image larger than the actual frame buffer. I had originally tried this with the regular osg::CameraNode. While I was able to successfully access the pixel data out of the pBuffer, i was not able to get OSG to fill an area larger than the frame buffer. I went through each of the available render target implementations (FRAME_BUFFER, FRAME_BUFFER_OBJECT, PIXEL_BUFFER, etc.) with no luck. Is there a trick there? Looking through the mailing list archives, it seemed as though the Producer camera was the only one that could render to a texture larger than the frame buffer. I would have preferred to use the osg::CameraNode because that was working fine until I needed the larger image size. Robert Osfield [EMAIL PROTECTED] wrote: On 5/9/06, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: I have build a small application modeled after the one in Producer/Test/PBuffer. The major difference here is that where that example assigns the PBuffer's texure to another camera as in: camera-getRenderSurface()-setReadDrawable(pBufferCamera-getRenderSurface()); I want direct access to the texture in pBufferCamera's render surface. Ideally, I would want a pointer to some standard GL_RGB formatted OpenGL image. I have tried a couple things but I haven't found exactly how to get the texture data from the pBufferCamera. Any Ideas?You'll need to do a glReadPixels. Note that osg::Image has thisalready available as a method.Also consider just using the osg::CameraNode for the job, it can helpyou seemlessly work with FBO's, pixel buffers and standard framebuffer. See the osgprerender which includes a segment which has aread to an osg::Image.Robert.___osg-users mailing listosg-users@openscenegraph.nethttp://openscenegraph.net/mailman/listinfo/osg-usershttp://www.openscenegraph.org/ New Yahoo! Messenger with Voice. Call regular phones from your PC and save big.___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
Re: [osg-users] text
Ok, I found that example, thanksOn 5/2/06, Philip Hahn [EMAIL PROTECTED] wrote: Bahar-It sounds like want is a heads-up display. Here is a tutorial explaining how to accomplish it: http://www.nps.navy.mil/cs/sullivan/osgtutorials/osgText.htm Also the osgtext example included with OSG shows how to accomplish this (look at the function createHUDText() )Hope this helps,Philip Hahn On 5/2/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hi all,I want some string to be located in my scene as text. I use osgText for this. But the text is seen as a 3d object in the scene, I mean as the camera moves the string stays in its place like any object in the scene. I want the text to be located in its place statically even if the camera moves. Is there any way to do this or should I make the text follow the camera and take its rotation and positions too?The code is as this://osgText::Font* font = osgText::readFontFile(fonts/arial.ttf); osg::Geode* geode = new osg::Geode; osg::Vec4 layoutColor(1.0f,1.0f,0.0f,1.0f); float windowHeight = 400.0f; float windowWidth = 300.0f; float margin = 50.0f; float layoutCharacterSize = 20.0f; osgText::Text* text = new osgText::Text; text-setFont(font); text-setColor(layoutColor); text-setCharacterSize(layoutCharacterSize); text-setPosition(osg::Vec3(0.0f,0.0f,10.0f)); // the default layout is left to right, typically used in languages // originating from europe such as English, French, German, Spanish etc.. text-setLayout(osgText::Text::LEFT_TO_RIGHT); text-setText(I am the passenger); geode-addDrawable(text); root-addChild(geode); //Thanks...--bahar-- CodeSchbeke Software SolutionsSenior Project TeamComputer Engineering Department Middle East Technical University Ankara/TURKEYhttp://senior.ceng.metu.edu.tr/2006/codeschbeke ___osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/ ___osg-users mailing listosg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/-- CodeSchbeke Software SolutionsSenior Project TeamComputer Engineering DepartmentMiddle East Technical UniversityAnkara/TURKEY http://senior.ceng.metu.edu.tr/2006/codeschbeke ___ osg-users mailing list osg-users@openscenegraph.net http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
