Re: [Flightgear-devel] Now Rembrandt here...
Gene Buckle On Thu, 12 Apr 2012, Frederic Bouvier wrote: Here is a video with a steady view to see shadow stability. http://youtu.be/JtEXIn2yL94 I also added 3 different sequences with different levels of filtering. Filtering is not yet configurable but is selectable in the sunlight shader with simple code modification. Fred, is there a reason the shadows on the instrument faces are so jumpy? Probably :) one reason is the moving sun that few other games have (not speaking of flight simulators, of course) -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Sat, 14 Apr 2012, Frederic Bouvier wrote: Fred, is there a reason the shadows on the instrument faces are so jumpy? Probably :) one reason is the moving sun that few other games have (not speaking of flight simulators, of course) Ok. :) I noticed the Anteworld demo of Outerra has the same issue when I was fiddling with it yesterday. The heavy filtering does help quite a bit, but there should be some way to eliminate that jitter. It's a shame we can't just slow the planet down. :D g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. http://www.diy-cockpits.org/coll - Go Collimated or Go Home. Some people collect things for a hobby. Geeks collect hobbies. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://www.scarletdme.org - Get it _today_! Buying desktop hardware and installing a server OS doesn't make a server-class system any more than sitting in a puddle makes you a duck. [Cipher in a.s.r] -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
I have the same bug when the skydome scattering shader is enable. Are you sure you have disable skydome scattering shader ? Look, I literally wrote the part of the effect file that controls what shaders run when the skydome shader is enabled and what shaders don't, I did a fair share of the skydome shader code, and I know very well how it looks like when it's on. So may we assume for the purpose of the discussion that when I say the skydome shader is off, it is actually off? Otherwise I feel we'll not get anywhere. Another possibility is that an extension check fails for you : extension-supportedGL_ARB_texture_rg/extension-supported Try to locate that line in Effects/model-default.eff and Effects/terrain-default.eff and remove it or comment it out Will try. In any case, the information that some fallback code is probably be running is helpful already, I should be able to check this easily by setting gl_FragColor to blue in the shader that ought to be running and investigate from there. Here is a video with a steady view to see shadow stability. Okay, thanks, that's helpful. My flickering is a bit worse than in the first case in this video, but not much. The filtering seems to improve things a lot - what's the downside of it? Cheers, * Thorsten -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
Will try. In any case, the information that some fallback code is probably be running is helpful already, I should be able to check this easily by setting gl_FragColor to blue in the shader that ought to be running and investigate from there. If you can post a screenshot with the buffers displayed, that might help to locate the problem. Regards, -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Fri, 13 Apr 2012 06:23:23 +, Renk wrote in message e495a106ff5f31448739e79d34138c190ec04...@mbs3.ad.jyu.fi: Here is a video with a steady view to see shadow stability. Okay, thanks, that's helpful. My flickering is a bit worse than in the first case in this video, but not much. ..what happens if you run your video card at a slower frame rate? -- ..med vennlig hilsen = with Kind Regards from Arnt Karlsen ...with a number of polar bear hunters in his ancestry... Scenarios always come in sets of three: best case, worst case, and just in case. -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Thu, 12 Apr 2012, Frederic Bouvier wrote: Here is a video with a steady view to see shadow stability. http://youtu.be/JtEXIn2yL94 I also added 3 different sequences with different levels of filtering. Filtering is not yet configurable but is selectable in the sunlight shader with simple code modification. Fred, is there a reason the shadows on the instrument faces are so jumpy? g. -- Proud owner of F-15C 80-0007 http://www.f15sim.com - The only one of its kind. http://www.diy-cockpits.org/coll - Go Collimated or Go Home. Some people collect things for a hobby. Geeks collect hobbies. ScarletDME - The red hot Data Management Environment A Multi-Value database for the masses, not the classes. http://www.scarletdme.org - Get it _today_! Buying desktop hardware and installing a server OS doesn't make a server-class system any more than sitting in a puddle makes you a duck. [Cipher in a.s.r] -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
I think that is what we have for now. You can do better by increasing your shadow map size to 8192 or 16384, but at the 16384 resolution my performance goes into the tank, and at 8192, there are still many shadow artifacts due to lack of resolution. (clearly blocky/xelated/aliasing edges, something that looks like z-buffer fighting for objects further way, etc.) Hopefully this will improve with future tuning. Okay, just to be sure, I've now also compared what I see to the IAR-80 on Vinson video by Fred and to the Cub flying around Kufstein video by HHS. I'm seeing a number of issues: 1) The shadows around the aircraft have a ragged egde. That I understand is a function of the shadow map size. I can't go beyond 4096, I get an error on the console trying to go higher - but 4096 works fine with acceptable framerates, the edge is just a limit of my GPU and that is okay. 2) I see shadows flickering (I tried the Cub cockpit for comparison) when I'm in level flight where they shouldn't move - the effect is a bit like a shadow cast by a candle flickering in the wind. In Heiko's video that is sometimes visible (he doesn't fly straight much, and when the shadows actually move the effect is masked). I've never seen it in Fred's videos - is it not there, or just not in the video? Personally, I find that flickering maddening - I've ended my test flight after 5 minutes because I was starting to get a headache. 3) On my box, all three panels in the screen edges show the same image - not so on Fred's videos - is this the intended behaviour? 4) Fred's IAR-80 video has the camera circling around an IAR-80 parked on the Vinson. I tried to do this as well, but for me the effect comes out very differently. The amount of shadow I see depends on the angle of the view axis with the sunlight direction - under some angles I see dark shadows, under some angles I see no shadows at all whereas in the video the shadows behave as expected, i.e. they stay the same independent of view angle. Under some angles (sun in my back) I also get a whiteout of bright textures. I haven't seen this issue in any of the videos, but it likewise looks very unrealistic - I watch an object, and as I overfly it, the shadow goes away... (Needless to say, I've switched any lightfield rendering scheme which would interfere off... Also, most tests I've done with FGData master rather than my customized branch) As I said earlier, I respect the amount of work which has gone into that very much and I'm also aware that this is brand new technology with some quirks to be sorted out. But at least I can't really work with it or fly the state which is on my system. So - since it's not working here - can I help in any way to pin down the issues and so help you fixing them? Cheers, * Thorsten -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On 12 Apr 2012, at 08:24, Renk Thorsten wrote: 3) On my box, all three panels in the screen edges show the same image - not so on Fred's videos - is this the intended behaviour? This is the important one - it means the multiple render targets isn't working, so you have no chance of seeing anything sensible with Rembrandt. (As I understand it) James -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
Thorsten wrote: I think that is what we have for now. You can do better by increasing your shadow map size to 8192 or 16384, but at the 16384 resolution my performance goes into the tank, and at 8192, there are still many shadow artifacts due to lack of resolution. (clearly blocky/xelated/aliasing edges, something that looks like z-buffer fighting for objects further way, etc.) Hopefully this will improve with future tuning. Okay, just to be sure, I've now also compared what I see to the IAR-80 on Vinson video by Fred and to the Cub flying around Kufstein video by HHS. I'm seeing a number of issues: 1) The shadows around the aircraft have a ragged egde. That I understand is a function of the shadow map size. I can't go beyond 4096, I get an error on the console trying to go higher - but 4096 works fine with acceptable framerates, the edge is just a limit of my GPU and that is okay. 2) I see shadows flickering (I tried the Cub cockpit for comparison) when I'm in level flight where they shouldn't move - the effect is a bit like a shadow cast by a candle flickering in the wind. In Heiko's video that is sometimes visible (he doesn't fly straight much, and when the shadows actually move the effect is masked). I've never seen it in Fred's videos - is it not there, or just not in the video? Personally, I find that flickering maddening - I've ended my test flight after 5 minutes because I was starting to get a headache. 3) On my box, all three panels in the screen edges show the same image - not so on Fred's videos - is this the intended behaviour? 4) Fred's IAR-80 video has the camera circling around an IAR-80 parked on the Vinson. I tried to do this as well, but for me the effect comes out very differently. The amount of shadow I see depends on the angle of the view axis with the sunlight direction - under some angles I see dark shadows, under some angles I see no shadows at all whereas in the video the shadows behave as expected, i.e. they stay the same independent of view angle. Under some angles (sun in my back) I also get a whiteout of bright textures. I haven't seen this issue in any of the videos, but it likewise looks very unrealistic - I watch an object, and as I overfly it, the shadow goes away... The disappearing shadow was caused by the sun camera and range animation and was fixed a while back - can you confirm that you are using the very latest fg/sg/fgdata? And it's probably worth checking that you have the very latest nVidia drivers. It looks to me as if you are pushing the very limits of your GPU; I think you might have to accept that you are not going to be able to use the added facilities provided by Rembrandt. Vivian Vivian -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Thu, 2012-04-12 at 09:01 +0100, James Turner wrote: On 12 Apr 2012, at 08:24, Renk Thorsten wrote: 3) On my box, all three panels in the screen edges show the same image - not so on Fred's videos - is this the intended behaviour? This is the important one - it means the multiple render targets isn't working, so you have no chance of seeing anything sensible with Rembrandt. (As I understand it) Also I see that --prop:/sim/rendering/no-16bit-buffer=true falls back to 8-but normal buffers. That might be just a bit to low for normal buffers. Erik -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Thu, 2012-04-12 at 10:29 +0200, Erik Hofman wrote: On Thu, 2012-04-12 at 09:01 +0100, James Turner wrote: On 12 Apr 2012, at 08:24, Renk Thorsten wrote: 3) On my box, all three panels in the screen edges show the same image - not so on Fred's videos - is this the intended behaviour? This is the important one - it means the multiple render targets isn't working, so you have no chance of seeing anything sensible with Rembrandt. (As I understand it) Also I see that --prop:/sim/rendering/no-16bit-buffer=true falls back to 8-but normal buffers. That might be just a bit to low for normal buffers. And it looks that way, Replacing GL_RGBA8 with GL_RGBA16 removes the flickering and still got me a working system. Erik -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
The disappearing shadow was caused by the sun camera and range animation and was fixed a while back - can you confirm that you are using the very latest fg/sg/fgdata? And it's probably worth checking that you have the very latest nVidia drivers. FG, SG and FGData are pulled and compiled as of yesterday noon, should be recent enough. The driver is somewhat older - as with anything which works fine on my computer, I'm reluctant to fiddle with it because on past occasions I have found myself struggling for a few days just to restore the previous state of the system when an update went wrong, and I simply don't have the time. It looks to me as if you are pushing the very limits of your GPU; I think you might have to accept that you are not going to be able to use the added facilities provided by Rembrandt. I have no problem with that (modulo the concerns I expressed yesterday) - I'm just trying to determine how my time is best spent. If there's a consensus that my GPU isn't up to the job, then I'll extend my atmosphere rendering framework for non-Rembrandt rendering and if anyone wants it in Rembrandt someone else has to port it because I can't develop something when I can't see the results. If Rembrandt can be made to work on my machine, then I can write stuff within the Rembrandt framework at some point. That's all there is to it. Cheers, * Thorsten -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Thu, 2012-04-12 at 10:35 +0200, Erik Hofman wrote: On Thu, 2012-04-12 at 10:29 +0200, Erik Hofman wrote: On Thu, 2012-04-12 at 09:01 +0100, James Turner wrote: On 12 Apr 2012, at 08:24, Renk Thorsten wrote: 3) On my box, all three panels in the screen edges show the same image - not so on Fred's videos - is this the intended behaviour? This is the important one - it means the multiple render targets isn't working, so you have no chance of seeing anything sensible with Rembrandt. (As I understand it) Also I see that --prop:/sim/rendering/no-16bit-buffer=true falls back to 8-but normal buffers. That might be just a bit to low for normal buffers. And it looks that way, Replacing GL_RGBA8 with GL_RGBA16 removes the flickering and still got me a working system. even GL_RGB16 works. Erik -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
De: James Turner On 12 Apr 2012, at 08:24, Renk Thorsten wrote: 3) On my box, all three panels in the screen edges show the same image - not so on Fred's videos - is this the intended behaviour? This is the important one - it means the multiple render targets isn't working, so you have no chance of seeing anything sensible with Rembrandt. (As I understand it) Glad to see someone else if able to make the correct diagnosis :) If the internal buffer are not correct, lighting equations done in the ambient, sunlight and fog pass will work on bogus inputs. I'll add that you get that either if an effect not converted to Rembrandt is used, or if the default shader has a build error and OSG fallback to fixed functions. So if the lightfield or the skydome is not enabled, check the console for a shader build error. Regards, -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
De: Erik Hofman e...@ehofman.com On Thu, 2012-04-12 at 10:35 +0200, Erik Hofman wrote: On Thu, 2012-04-12 at 10:29 +0200, Erik Hofman wrote: On Thu, 2012-04-12 at 09:01 +0100, James Turner wrote: On 12 Apr 2012, at 08:24, Renk Thorsten wrote: 3) On my box, all three panels in the screen edges show the same image - not so on Fred's videos - is this the intended behaviour? This is the important one - it means the multiple render targets isn't working, so you have no chance of seeing anything sensible with Rembrandt. (As I understand it) Also I see that --prop:/sim/rendering/no-16bit-buffer=true falls back to 8-but normal buffers. That might be just a bit to low for normal buffers. And it looks that way, Replacing GL_RGBA8 with GL_RGBA16 removes the flickering and still got me a working system. even GL_RGB16 works. For all the buffers or only the normal buffer ? You're not supposed to have multiple render target of different element size. RGBA8 = 32bits, as well as RG16. What is your card brand and model ? I came across a blog post that compares multiple ways to compress the normals in an 8bit texture and I will try that shortly. Regards, -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Thu, 2012-04-12 at 10:43 +0200, Erik Hofman wrote: even GL_RGB16 works. And therefore I propose this patch. Erik diff --git a/src/Main/renderer.cxx b/src/Main/renderer.cxx index 1c6e77b..1c5ea0a 100644 --- a/src/Main/renderer.cxx +++ b/src/Main/renderer.cxx @@ -65,6 +65,7 @@ #include osgViewer/Renderer #include simgear/math/SGMath.hxx +#include simgear/screen/extensions.hxx #include simgear/scene/material/matlib.hxx #include simgear/scene/material/EffectCullVisitor.hxx #include simgear/scene/material/Effect.hxx @@ -735,7 +736,7 @@ void buildDeferredBuffers( flightgear::CameraInfo* info, int shadowMapSize, bool if (normal16) info-addBuffer(flightgear::RenderBufferInfo::NORMAL_BUFFER, buildDeferredBuffer( 0x822C /*GL_RG16*/, 0x8227 /*GL_RG*/, GL_UNSIGNED_SHORT, osg::Texture::CLAMP_TO_BORDER) ); else -info-addBuffer(flightgear::RenderBufferInfo::NORMAL_BUFFER, buildDeferredBuffer( GL_RGBA8, GL_RGBA, GL_UNSIGNED_SHORT, osg::Texture::CLAMP_TO_BORDER) ); +info-addBuffer(flightgear::RenderBufferInfo::NORMAL_BUFFER, buildDeferredBuffer( GL_RGB16, GL_RGB, GL_UNSIGNED_SHORT, osg::Texture::CLAMP_TO_BORDER) ); info-addBuffer(flightgear::RenderBufferInfo::DIFFUSE_BUFFER, buildDeferredBuffer( GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, osg::Texture::CLAMP_TO_BORDER) ); info-addBuffer(flightgear::RenderBufferInfo::SPEC_EMIS_BUFFER, buildDeferredBuffer( GL_RGBA8, GL_RGBA, GL_UNSIGNED_BYTE, osg::Texture::CLAMP_TO_BORDER) ); @@ -1356,7 +1357,10 @@ FGRenderer::buildDeferredPipeline(flightgear::CameraGroup* cgroup, unsigned flag osg::GraphicsContext* gc) { CameraInfo* info = new CameraInfo(flags); - buildDeferredBuffers( info, _shadowMapSize, !fgGetBool(/sim/rendering/no-16bit-buffer, false ) ); + +bool gl_ext_texture_rg = SGIsOpenGLExtensionSupported(GL_ARB_texture_rg) + !fgGetBool(/sim/rendering/no-16bit-buffer, false ); + buildDeferredBuffers( info, _shadowMapSize, gl_ext_texture_rg ); osg::Camera* geometryCamera = buildDeferredGeometryCamera( info, gc ); cgroup-getViewer()-addSlave(geometryCamera, false); -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Thu, 2012-04-12 at 11:02 +0200, Frederic Bouvier wrote: De: Erik Hofman e...@ehofman.com even GL_RGB16 works. For all the buffers or only the normal buffer ? I only tested the normal buffer. You're not supposed to have multiple render target of different element size. RGBA8 = 32bits, as well as RG16. What is your card brand and model ? It's a NVidia GeForce 9600GT I came across a blog post that compares multiple ways to compress the normals in an 8bit texture and I will try that shortly. That might be a good idea, at least to save texture memory. Erik -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
De: Erik Hofman e...@ehofman.com On Thu, 2012-04-12 at 11:02 +0200, Frederic Bouvier wrote: What is your card brand and model ? It's a NVidia GeForce 9600GT I think Emilian has the same card and I don't think he had these problems. Maybe it's a good idea to collect user experience with Rembrandt on a Wiki page ? Regards, -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
The driver is somewhat older - as with anything which works fine on my computer, I'm reluctant to fiddle with it because on past occasions I have found myself struggling for a few days just to restore the previous state of the system when an update went wrong, and I simply don't have the time. I don't think it will change anything as NVidia is not making driver updates for this model of GPU for a long time. Regards, -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Thu, 2012-04-12 at 11:24 +0200, Frederic Bouvier wrote: De: Erik Hofman e...@ehofman.com On Thu, 2012-04-12 at 11:02 +0200, Frederic Bouvier wrote: What is your card brand and model ? It's a NVidia GeForce 9600GT I think Emilian has the same card and I don't think he had these problems. Maybe it's a good idea to collect user experience with Rembrandt on a Wiki page ? I'm using Linux, maybe that's the difference? BTW today there was a driver update but that didn't change anything regarding this 'problem'. Erik -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
De: Erik Hofman On Thu, 2012-04-12 at 11:24 +0200, Frederic Bouvier wrote: De: Erik Hofman e...@ehofman.com On Thu, 2012-04-12 at 11:02 +0200, Frederic Bouvier wrote: What is your card brand and model ? It's a NVidia GeForce 9600GT I think Emilian has the same card and I don't think he had these problems. Maybe it's a good idea to collect user experience with Rembrandt on a Wiki page ? I'm using Linux, maybe that's the difference? BTW today there was a driver update but that didn't change anything regarding this 'problem'. Updates are now for newer cards -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Thu, 2012-04-12 at 11:35 +0200, Frederic Bouvier wrote: De: Erik Hofman On Thu, 2012-04-12 at 11:24 +0200, Frederic Bouvier wrote: De: Erik Hofman e...@ehofman.com On Thu, 2012-04-12 at 11:02 +0200, Frederic Bouvier wrote: What is your card brand and model ? It's a NVidia GeForce 9600GT Updates are now for newer cards Hm, according to this site the 195.36.24 driver does support GL_ARB_texture_rg for Linux: http://feedback.wildfiregames.com/report/opengl/device/GeForce%209600% 20GT But I still have version 173.14.22 Erik -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Thu, 2012-04-12 at 11:31 +0200, Erik Hofman wrote: On Thu, 2012-04-12 at 11:24 +0200, Frederic Bouvier wrote: De: Erik Hofman e...@ehofman.com On Thu, 2012-04-12 at 11:02 +0200, Frederic Bouvier wrote: What is your card brand and model ? It's a NVidia GeForce 9600GT I think Emilian has the same card and I don't think he had these problems. Maybe it's a good idea to collect user experience with Rembrandt on a Wiki page ? I'm using Linux, maybe that's the difference? Ah the 32-bit drivers don't support this extension while the 64-bit drivers do. Erik -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
I came across a blog post that compares multiple ways to compress the normals in an 8bit texture and I will try that shortly. That might be a good idea, at least to save texture memory. This blog entry is now cited as reference in the Rembrandt wiki page. -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
On Thursday 12 April 2012 11:24:23 Frederic Bouvier wrote: De: Erik Hofman e...@ehofman.com On Thu, 2012-04-12 at 11:02 +0200, Frederic Bouvier wrote: What is your card brand and model ? It's a NVidia GeForce 9600GT I think Emilian has the same card and I don't think he had these problems. Maybe it's a good idea to collect user experience with Rembrandt on a Wiki page ? Regards, -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel I've got an 8600GT non mobile, and it works just fine(minus the perf hit with shadows on, due to the huge vertex number in the suncamera scene). I'm using the latest driver 253.40 as of yesterday. Might be a good idea to update the drivers. On Linux they support any cards from the 6xxx series up with this branch of drivers, while fx and older are supposed to use the older driver series. HTH, Emilian -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
[Flightgear-devel] Now Rembrandt here...
The amount of shadow I see depends on the angle of the view axis with the sunlight direction - under some angles I see dark shadows, under some angles I see no shadows at all Hi Thorsten, I have the same bug when the skydome scattering shader is enable. Are you sure you have disable skydome scattering shader ? Cheers, Clément -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
Thorsten, I'll add that you get that either if an effect not converted to Rembrandt is used, or if the default shader has a build error and OSG fallback to fixed functions. So if the lightfield or the skydome is not enabled, check the console for a shader build error. Another possibility is that an extension check fails for you : extension-supportedGL_ARB_texture_rg/extension-supported Try to locate that line in Effects/model-default.eff and Effects/terrain-default.eff and remove it or comment it out Regards, -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Now Rembrandt here...
1) The shadows around the aircraft have a ragged egde. That I understand is a function of the shadow map size. I can't go beyond 4096, I get an error on the console trying to go higher - but 4096 works fine with acceptable framerates, the edge is just a limit of my GPU and that is okay. 2) I see shadows flickering (I tried the Cub cockpit for comparison) when I'm in level flight where they shouldn't move - the effect is a bit like a shadow cast by a candle flickering in the wind. In Heiko's video that is sometimes visible (he doesn't fly straight much, and when the shadows actually move the effect is masked). I've never seen it in Fred's videos - is it not there, or just not in the video? Personally, I find that flickering maddening - I've ended my test flight after 5 minutes because I was starting to get a headache. Here is a video with a steady view to see shadow stability. http://youtu.be/JtEXIn2yL94 I also added 3 different sequences with different levels of filtering. Filtering is not yet configurable but is selectable in the sunlight shader with simple code modification. Regards, -Fred -- For Developers, A Lot Can Happen In A Second. Boundary is the first to Know...and Tell You. Monitor Your Applications in Ultra-Fine Resolution. Try it FREE! http://p.sf.net/sfu/Boundary-d2dvs2 ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
The first line says it all. Okay... so what does it mean? I'd first try to reduce the size of the shadow map : --prop:/sim/rendering/shadows/map-size=1024 No success, problem persists. or reduce the window size : --geometry=800x600 to reduce the memory footprint. Same as above, problem persists. Also when I use small shadow map in addition to 800x600 window no success (leaving aside the fact that Flightgear running in a window of less than a quarter of my screen isn't really thrilling...) . You have a laptop right ? Maybe you have shared memory between the CPU and the GPU (NVidia calls that TurboCache) and you didn't reserve enough memory for the GPU. This is a BIOS setting. Allocate the more you can to the GPU and retry. Can't do - this is a SONY VAIO laptop, which means you don't get to set anything relevant in the BIOS unless you start hacking it. :-( Somehow I have the feeling that this isn't a memory problem though. Experimenting with Earthview, I can load a sphere textured with 5 4096x4096 and 11 2048x2048 texture sheets allright without bothering with any smart texture management in addition to a normal Flightgear scenery and it works just fine - going through the numbers, that's quite a lot of raw data to be stored somewhere. My GPU isn't exactly new, but given what cloud and weather I can run and what others get as framerates out of the same scene, it is still rather competitive and probably better than what most people in the forum run (i.e. those who don't invest in a dedicated high-end GPU). Plus, so far with one exception (landmass effect at high quality) pretty much everything so far ran just fine out of the box with very acceptable framerates. So, I have a feeling that Rembrandt might be going to leave a lot of folks with blank screens at this point. I don't want to be negative here, maybe it is a trivial problem, but this isn't really a shader which I don't really need to see, this gives me an unusuable Flightgear. Cheers, * Thorsten -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
On Wed, 2012-04-11 at 06:25 +, Renk Thorsten wrote: The first line says it all. Okay... so what does it mean? I'd first try to reduce the size of the shadow map : --prop:/sim/rendering/shadows/map-size=1024 No success, problem persists. or reduce the window size : --geometry=800x600 to reduce the memory footprint. Same as above, problem persists. Also when I use small shadow map in addition to 800x600 window no success (leaving aside the fact that Flightgear running in a window of less than a quarter of my screen isn't really thrilling...) . In fact I have the same problem with a AMD-X2/3Ghz with 2Gb memory and a GeForce 9600GT: fgfs --enable-rembrandt --prop:/sim/rendering/shadows/map-size=1024 --geometry=800x60 RenderStage::runCameraSetUp(), FBO setup failed, FBO status= 0x8cd6 Warning: RenderStage::runCameraSetUp(State) Pbuffer does not support multiple color outputs. Erik -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
On Wed, 2012-04-11 at 06:25 +, Renk Thorsten wrote: The first line says it all. Okay... so what does it mean? I'd first try to reduce the size of the shadow map : --prop:/sim/rendering/shadows/map-size=1024 No success, problem persists. or reduce the window size : --geometry=800x600 to reduce the memory footprint. Same as above, problem persists. Also when I use small shadow map in addition to 800x600 window no success (leaving aside the fact that Flightgear running in a window of less than a quarter of my screen isn't really thrilling...) . Maybe this may help developers, it's about the same message: http://forum.openscenegraph.org/viewtopic.php?t=8905 Erik -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
De: Erik Hofman On Wed, 2012-04-11 at 06:25 +, Renk Thorsten wrote: The first line says it all. Okay... so what does it mean? It means that a required Framebuffer Object failed to setup and the fallback isn't OK for Rembrandt. I'd first try to reduce the size of the shadow map : --prop:/sim/rendering/shadows/map-size=1024 No success, problem persists. or reduce the window size : --geometry=800x600 to reduce the memory footprint. Same as above, problem persists. Also when I use small shadow map in addition to 800x600 window no success (leaving aside the fact that Flightgear running in a window of less than a quarter of my screen isn't really thrilling...) . Maybe this may help developers, it's about the same message: http://forum.openscenegraph.org/viewtopic.php?t=8905 This issue is related to iOS and OGL ES, that is a bit different. There is the concept of implicit attachment in OSG, so INCOMPLETE_ATTACHMENT shouldn't occur if an attachment was missing in the first place. I am thinking of an explicit attachment that failed silently. There is a way to increase the OSG log level, but I don't remember it for the moment. I have to make guess as I don't have a card that exhibit that issue. You can try to edit fg/src/Main/renderer.cxx and change GL_DEPTH_COMPONENT32 to GL_DEPTH_COMPONENT24. You may also try to add the --bpp=32 option to the fgfs command line. Last, make sure that you didn't enable multithreading in preferences.xml (AutomaticSelection or something else) Plus, so far with one exception (landmass effect at high quality) pretty much everything so far ran just fine out of the box with very acceptable framerates. So, I have a feeling that Rembrandt might be going to leave a lot of folks with blank screens at this point. I don't want to be negative here, maybe it is a trivial problem, but this isn't really a shader which I don't really need to see, this gives me an unusuable Flightgear. Be sure I value your feedback, but we are exploring new lands here. There is not so much OSG deferred rendering example or real application around, so please be forgiving. And I don't think Flightgear is unusable for anybody. The Rembrandt renderer is optional and the classical/2.6 renderer should work for everybody. Regards, -Fred -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
Be sure I value your feedback, but we are exploring new lands here. There is not so much OSG deferred rendering example or real application around, so please be forgiving. And I don't think Flightgear is unusable for anybody. The Rembrandt renderer is optional and the classical/2.6 renderer should work for everybody. Sorry if that came across the wrong way. I am well aware of this - I am probably more worried about aircraft and/or scenery being converted and committed at this stage than about project Rembrandt as such, since the default renderer is working just fine. What also bothers me is the attitude I've come across with other people (not you!) which goes like 'Don't bother writing something for the 2.6 renderer because Rembrandt will be there.' - which sounds more like replacement than optional. I'm also observing statements being made in the forum by various people - some are rather cautious, others raise expectations for a release which may backfire badly if there turn out to be issues with many cards. So I'm not writing this out of the blue. Personally, I can live without shadows if my GPU turns out not to support this at all in the end - but I can't really if all aircraft at night use Rembrandt in a non-optional way and all I get to see with default rendering is darkness. And so on. I'll explore your suggestions and let you know what happens. Cheers, * Thorsten -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
On Wed, 2012-04-11 at 10:18 +0200, Frederic Bouvier wrote: I have to make guess as I don't have a card that exhibit that issue. You can try to edit fg/src/Main/renderer.cxx and change GL_DEPTH_COMPONENT32 to GL_DEPTH_COMPONENT24. You may also try to add the --bpp=32 option to the fgfs command line. Last, make sure that you didn't enable multithreading in preferences.xml (AutomaticSelection or something else) None of these seems to help but when I apply the attached patch (as suggested by http://markmail.org/message/yfuz7je43bdzt6h2) at least the warnings are gone and I see the scenery (but not yet perfect). Erik diff --git a/src/Main/renderer.cxx b/src/Main/renderer.cxx index a4848d3..9f4ed10 100644 --- a/src/Main/renderer.cxx +++ b/src/Main/renderer.cxx @@ -761,9 +761,9 @@ osg::Camera* FGRenderer::buildDeferredGeometryCamera( flightgear::CameraInfo* in camera-setRenderTargetImplementation( osg::Camera::FRAME_BUFFER_OBJECT ); camera-setViewport( new osg::Viewport ); attachBufferToCamera( info, camera, osg::Camera::DEPTH_BUFFER, flightgear::GEOMETRY_CAMERA, flightgear::RenderBufferInfo::DEPTH_BUFFER ); -attachBufferToCamera( info, camera, osg::Camera::COLOR_BUFFER0, flightgear::GEOMETRY_CAMERA, flightgear::RenderBufferInfo::NORMAL_BUFFER ); -attachBufferToCamera( info, camera, osg::Camera::COLOR_BUFFER1, flightgear::GEOMETRY_CAMERA, flightgear::RenderBufferInfo::DIFFUSE_BUFFER ); -attachBufferToCamera( info, camera, osg::Camera::COLOR_BUFFER2, flightgear::GEOMETRY_CAMERA, flightgear::RenderBufferInfo::SPEC_EMIS_BUFFER ); +attachBufferToCamera( info, camera, osg::Camera::COLOR_BUFFER, flightgear::GEOMETRY_CAMERA, flightgear::RenderBufferInfo::NORMAL_BUFFER ); +attachBufferToCamera( info, camera, osg::Camera::COLOR_BUFFER, flightgear::GEOMETRY_CAMERA, flightgear::RenderBufferInfo::DIFFUSE_BUFFER ); +attachBufferToCamera( info, camera, osg::Camera::COLOR_BUFFER, flightgear::GEOMETRY_CAMERA, flightgear::RenderBufferInfo::SPEC_EMIS_BUFFER ); camera-setDrawBuffer(GL_FRONT); camera-setReadBuffer(GL_FRONT); -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
Thorsten wrote: -Original Message- From: Renk [mailto:thorsten.i.r...@jyu.fi] Sent: 11 April 2012 09:33 To: FlightGear developers discussions Subject: Re: [Flightgear-devel] No Rembrandt here... Be sure I value your feedback, but we are exploring new lands here. There is not so much OSG deferred rendering example or real application around, so please be forgiving. And I don't think Flightgear is unusable for anybody. The Rembrandt renderer is optional and the classical/2.6 renderer should work for everybody. Sorry if that came across the wrong way. I am well aware of this - I am probably more worried about aircraft and/or scenery being converted and committed at this stage than about project Rembrandt as such, since the default renderer is working just fine. What also bothers me is the attitude I've come across with other people (not you!) which goes like 'Don't bother writing something for the 2.6 renderer because Rembrandt will be there.' - which sounds more like replacement than optional. I'm also observing statements being made in the forum by various people - some are rather cautious, others raise expectations for a release which may backfire badly if there turn out to be issues with many cards. So I'm not writing this out of the blue. Personally, I can live without shadows if my GPU turns out not to support this at all in the end - but I can't really if all aircraft at night use Rembrandt in a non-optional way and all I get to see with default rendering is darkness. And so on. I'll explore your suggestions and let you know what happens. Ah the dangers of forums! Whether or not an aircraft is converted to Rembrandt depends on the aircraft developer. Personally, I have taken the route of adding a Rembrandt version to the inventory and leaving a version still fully compatible with 2.6.0. This is by way of being a test-bed for Rembrandt. So far it has taken several weeks, and most of it is just eye-candy. This is partly my learning curve, partly bugs, and not least the sheer size of the task. I think that you are quite right to express your concerns. I suppose in the future there might be aircraft or scenery developed purely for Rembrandt (and that would be a poor decision by the developer), but at the moment I cannot see why your fears would be realised while Rembrandt remains optional. Vivian -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
Vivian, or anyone else with optional-Rembrandt experience, feel free to add some instructions on how to make an aircraft support both renderers to http://wiki.flightgear.org/Project_Rembrandt#Porting_aircraft So we can forward aircraft devs to that (once Rembrandt is stable/complete). Cheers, Gijs -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
Hi Erik, De: Erik Hofman On Wed, 2012-04-11 at 10:18 +0200, Frederic Bouvier wrote: I have to make guess as I don't have a card that exhibit that issue. You can try to edit fg/src/Main/renderer.cxx and change GL_DEPTH_COMPONENT32 to GL_DEPTH_COMPONENT24. You may also try to add the --bpp=32 option to the fgfs command line. Last, make sure that you didn't enable multithreading in preferences.xml (AutomaticSelection or something else) None of these seems to help but when I apply the attached patch (as suggested by http://markmail.org/message/yfuz7je43bdzt6h2) at least the warnings are gone and I see the scenery (but not yet perfect). With this patch you are trading a bug for a bug. Assigning the same attachment to three buffers as the same effect than assigning three different values to the same variable. Regards, -Fred -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
On Wed, 2012-04-11 at 13:35 +0200, Frederic Bouvier wrote: Hi Erik, With this patch you are trading a bug for a bug. Assigning the same attachment to three buffers as the same effect than assigning three different values to the same variable. I was already afraid something like that was happening. Erik -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
On Wed, 2012-04-11 at 13:35 +0200, Frederic Bouvier wrote: Hi Erik, With this patch you are trading a bug for a bug. Assigning the same attachment to three buffers as the same effect than assigning three different values to the same variable. I was already afraid something like that was happening. Try --prop:/sim/rendering/no-16bit-buffer=true although the symptoms were a bit different Regards, -Fred -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
On Wed, 2012-04-11 at 14:15 +0200, Frederic Bouvier wrote: On Wed, 2012-04-11 at 13:35 +0200, Frederic Bouvier wrote: Hi Erik, With this patch you are trading a bug for a bug. Assigning the same attachment to three buffers as the same effect than assigning three different values to the same variable. I was already afraid something like that was happening. Try --prop:/sim/rendering/no-16bit-buffer=true although the symptoms were a bit different Yes! that works. Erik -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
Hi Thorsten, I think that is what we have for now. You can do better by increasing your shadow map size to 8192 or 16384, but at the 16384 resolution my performance goes into the tank, and at 8192, there are still many shadow artifacts due to lack of resolution. (clearly blocky/xelated/aliasing edges, something that looks like z-buffer fighting for objects further way, etc.) Hopefully this will improve with future tuning. I'd love to be able to overfly the SFO terminal at 1000' for instance and have solid shadows on the AI aircraft, solid shadows on the light poles, solid shadows from the terminal building, etc. I'm hoping that will ultimately be possible. Otherwise, just getting one good clear shadow of our own aircraft onto the ground and onto ourselves would be a good 90% solution. Curt. On Wed, Apr 11, 2012 at 7:52 AM, Renk Thorsten thorsten.i.r...@jyu.fiwrote: Try --prop:/sim/rendering/no-16bit-buffer=true although the symptoms were a bit different Same here, this sort of works up to the 4096 map size. I do see shadows and lights at night at KSFO, but the shadows are very... restless - they seem to flicker quite a bit, and all in all the light I get to see depends very much on the view angle - is that normal? The grainy shadow edge seems to be a function of the map size, it gets a lot worse when I go to 1024... Thanks for the fix in any case! Cheers, * Thorsten -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- Curtis Olson: http://www.atiak.com - http://aem.umn.edu/~uav/ http://www.flightgear.org - http://gallinazo.flightgear.org -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
[Flightgear-devel] No Rembrandt here...
Just freshly pulled and compiled GIT, fresh FGData master, trying to start with --enable-rembrandt results in garbage on the screen and an impressive list of errors - (I've omitted messages with meaning known to me in the following as well as repetitions). Running without Rembrandt looks fine to me on first glenace. Graphics card is an NVIDIA GeForce 8600M running under Linux with the nvidia driver. Cheers, * Thorsten RenderStage::runCameraSetUp(), FBO setup failed, FBO status= 0x8cd6 Warning: RenderStage::runCameraSetUp(State) Pbuffer does not support multiple olor outputs. Warning: RenderStage::runCameraSetUp(State) Pbuffer does not support multiple olor outputs. Warning: RenderStage::runCameraSetUp(State) Pbuffer does not support multiple olor outputs. Failed to create beacon for unknown runway 'KONT 26L OM'. Failed to create beacon for unknown runway 'KONT 26R OM'. Failed to create beacon for unknown runway 'KSAV 01 OM'. Failed to create beacon for unknown runway 'KSAV 10 OM'. Failed to create beacon for unknown runway 'PAYA 11 OM'. Failed to create beacon for unknown runway 'WARQ 26 OM'. Failed to create beacon for unknown runway 'WARR 10 OM'. Failed to create beacon for unknown runway 'ZULS 27R OM'. Failed to create beacon for unknown runway 'EDDB 07 MM'. Failed to create beacon for unknown runway 'EDDB 25 MM'. Failed to create beacon for unknown runway 'KIAD 01C MM'. Failed to create beacon for unknown runway 'KONT 08L MM'. Failed to create beacon for unknown runway 'KONT 26L MM'. Failed to create beacon for unknown runway 'KONT 26R MM'. Failed to create beacon for unknown runway 'KSAV 01 MM'. Failed to create beacon for unknown runway 'KSAV 10 MM'. Failed to create beacon for unknown runway 'PAED 06 MM'. Failed to create beacon for unknown runway 'PAYA 11 MM'. Failed to create beacon for unknown runway 'WADD 27 MM'. Failed to create beacon for unknown runway 'WALL 25 MM'. Failed to create beacon for unknown runway 'WAOO 10 MM'. Failed to create beacon for unknown runway 'WAOP 34 MM'. Failed to create beacon for unknown runway 'WARJ 09 MM'. Failed to create beacon for unknown runway 'WARQ 26 MM'. Failed to create beacon for unknown runway 'WARR 10 MM'. Failed to create beacon for unknown runway 'WIDD 04 MM'. Failed to create beacon for unknown runway 'WIHH 24 MM'. Failed to create beacon for unknown runway 'KIAD 19C IM'. Failed to create beacon for unknown runway 'KONT 26L IM'. RenderStage::runCameraSetUp(), FBO setup failed, FBO status= 0x8cd6 Warning: detected OpenGL error 'invalid enumerant' at end of SceneView::draw() Loading local weather routines... Animated jetways ... initialized loading scenario 'nimitz_demo' Image /home/fgfs/FGData/fgdata/Textures/Water/waves-ver10-nm.dds uses compressed textures which cannot be supported on some systems. Please decompress this texture for improved portability. (...) RenderStage::runCameraSetUp(), FBO setup failed, FBO status= 0x8cd6 Warning: detected OpenGL error 'invalid enumerant' at start of State::apply() (...) -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] No Rembrandt here...
Hi Thorsten, The first line says it all. I'd first try to reduce the size of the shadow map : --prop:/sim/rendering/shadows/map-size=1024 or reduce the window size : --geometry=800x600 to reduce the memory footprint. You have a laptop right ? Maybe you have shared memory between the CPU and the GPU (NVidia calls that TurboCache) and you didn't reserve enough memory for the GPU. This is a BIOS setting. Allocate the more you can to the GPU and retry. Regards, -Fred - Mail original - De: Renk Thorsten Just freshly pulled and compiled GIT, fresh FGData master, trying to start with --enable-rembrandt results in garbage on the screen and an impressive list of errors - (I've omitted messages with meaning known to me in the following as well as repetitions). Running without Rembrandt looks fine to me on first glenace. Graphics card is an NVIDIA GeForce 8600M running under Linux with the nvidia driver. Cheers, * Thorsten RenderStage::runCameraSetUp(), FBO setup failed, FBO status= 0x8cd6 Warning: RenderStage::runCameraSetUp(State) Pbuffer does not support multiple olor outputs. Warning: RenderStage::runCameraSetUp(State) Pbuffer does not support multiple olor outputs. Warning: RenderStage::runCameraSetUp(State) Pbuffer does not support multiple olor outputs. Failed to create beacon for unknown runway 'KONT 26L OM'. Failed to create beacon for unknown runway 'KONT 26R OM'. Failed to create beacon for unknown runway 'KSAV 01 OM'. Failed to create beacon for unknown runway 'KSAV 10 OM'. Failed to create beacon for unknown runway 'PAYA 11 OM'. Failed to create beacon for unknown runway 'WARQ 26 OM'. Failed to create beacon for unknown runway 'WARR 10 OM'. Failed to create beacon for unknown runway 'ZULS 27R OM'. Failed to create beacon for unknown runway 'EDDB 07 MM'. Failed to create beacon for unknown runway 'EDDB 25 MM'. Failed to create beacon for unknown runway 'KIAD 01C MM'. Failed to create beacon for unknown runway 'KONT 08L MM'. Failed to create beacon for unknown runway 'KONT 26L MM'. Failed to create beacon for unknown runway 'KONT 26R MM'. Failed to create beacon for unknown runway 'KSAV 01 MM'. Failed to create beacon for unknown runway 'KSAV 10 MM'. Failed to create beacon for unknown runway 'PAED 06 MM'. Failed to create beacon for unknown runway 'PAYA 11 MM'. Failed to create beacon for unknown runway 'WADD 27 MM'. Failed to create beacon for unknown runway 'WALL 25 MM'. Failed to create beacon for unknown runway 'WAOO 10 MM'. Failed to create beacon for unknown runway 'WAOP 34 MM'. Failed to create beacon for unknown runway 'WARJ 09 MM'. Failed to create beacon for unknown runway 'WARQ 26 MM'. Failed to create beacon for unknown runway 'WARR 10 MM'. Failed to create beacon for unknown runway 'WIDD 04 MM'. Failed to create beacon for unknown runway 'WIHH 24 MM'. Failed to create beacon for unknown runway 'KIAD 19C IM'. Failed to create beacon for unknown runway 'KONT 26L IM'. RenderStage::runCameraSetUp(), FBO setup failed, FBO status= 0x8cd6 Warning: detected OpenGL error 'invalid enumerant' at end of SceneView::draw() Loading local weather routines... Animated jetways ... initialized loading scenario 'nimitz_demo' Image /home/fgfs/FGData/fgdata/Textures/Water/waves-ver10-nm.dds uses compressed textures which cannot be supported on some systems. Please decompress this texture for improved portability. (...) RenderStage::runCameraSetUp(), FBO setup failed, FBO status= 0x8cd6 Warning: detected OpenGL error 'invalid enumerant' at start of State::apply() (...) -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- Better than sec? Nothing is better than sec when it comes to monitoring Big Data applications. Try Boundary one-second resolution app monitoring today. Free. http://p.sf.net/sfu/Boundary-dev2dev ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel