Re: [Flightgear-devel] Further 3D Clouds updates
Frederic Bouvier wrote: I cheched your patch in but I still have the checkerboard : http://frbouvi.free.fr/flightsim/fgfs_clouds_checker.jpg but only when the weather scenario is set to none. -Fred Thanks for checking in all my patches. I'm aware that there are still checkerboard issues. That's next on my list :) -Stuart - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Further 3D Clouds updates
Well, just tested, very nice, I withnessed an improvement in performance. My specs: Athlon XP2000, 768MB ram, NVidia FX 5900 @ 1680x1050 (system: linux with nvidia drivers), no AA Using: sg + fg cvs, OSG 2.7.5; everything enabled, custom settings, using metar, no ai/mp planes around Startup: 30fps 2D/ 20fps 3D (clouds at 6000ft, scattered, thickness 600ft) Flying 400ft/440knots: 26fps 2D/ 19fps 3D (clouds at 6000ft, scattered, thickness 600ft) Flying 6300ft/440knots: 35fps 2d/ 11-22fps (clouds at 6000ft, scattered, thickness 600ft) Note: Flying within the clouds resulted in 22fps, getting very close or through clouds 11fps Changing conditions: Low vis (7000m), clouds at 8000ft(broken) and 4000ft(broken), thickness 600ft Flying @ 8000ft/440knots: 14fps 2D/ 6fps 3D Low vis being also a frame rate killer? Changing conditions: Vis (25000m), clouds at 5500ft(broken), thickness 600ft Flying @ 8000ft/440knots: 25fps 2D/ 16fps 3D Flying @ 6000ft/440knots: 20fps 2D/ 12fps 3D Hope this helps a bit, Markus Stuart Buchanan wrote: Stuart Buchanan wrote: Vivian Meazza wrote: And I have yet to see any 3d clouds. Any clues on where I should be looking (yes the box is checked :-)) Something has changed in the environment manager which means that clouds generateion is now inconsistent. I'm still tracking it down, as my recent changes shouldn't have affected this. Well, the cause was a bug in my code, but it didn't expose itself until we moved to multiple cameras. The attached patch fixes the problem. I've also put in a new heuristic to improve the frame-rate. Clouds that are already sorted are likely to still be sorted in subsequent frames. Therefore I've put in a back-off mechanism for the bubble-sort pass. This should mean that if you stay completely stationary, once the clouds become sorted they will eventually only perform a bubble sort pass every 128 frames. It would be good to get a feel for how bad performance is with 3D clouds. At the moment I don't have a handle on whether performance is almost good enough, or completely unacceptable. I'd appreciate it if people would post their observations, providing details of their machine spec, graphics options and the frame-rate with and without 3D clouds - ideally with a description of the general load on the machine. Thanks, -Stuart - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Further 3D Clouds updates
I haven't applied the patch yet, so I can't give full specs. With CVS as of a few days ago, I did notice that OSG attributed almost double the 'update' time with 3D clouds on. I assume that Tim can comment on what The 'update' time is. I assume this means that (at least prior to your patch) the 3D clouds is current CPU limited. I can't comment on the 'draw' time difference, although it does appear that there are now extra 'cameras' in the last few weeks. On my multi-GPU system I used to have num-displays+1, now I seem to have num-displays*2. Is this expected? For multi-display, this would have an impac on framerate (particular with the serialized mode that is default. I'll discuss more fully in a different email - I just wanted to touch this here in case it is clouds related. Regards... Matthew On 11/26/08, Markus Zojer [EMAIL PROTECTED] wrote: Well, just tested, very nice, I withnessed an improvement in performance. My specs: Athlon XP2000, 768MB ram, NVidia FX 5900 @ 1680x1050 (system: linux with nvidia drivers), no AA Using: sg + fg cvs, OSG 2.7.5; everything enabled, custom settings, using metar, no ai/mp planes around Startup: 30fps 2D/ 20fps 3D (clouds at 6000ft, scattered, thickness 600ft) Flying 400ft/440knots: 26fps 2D/ 19fps 3D (clouds at 6000ft, scattered, thickness 600ft) Flying 6300ft/440knots: 35fps 2d/ 11-22fps (clouds at 6000ft, scattered, thickness 600ft) Note: Flying within the clouds resulted in 22fps, getting very close or through clouds 11fps Changing conditions: Low vis (7000m), clouds at 8000ft(broken) and 4000ft(broken), thickness 600ft Flying @ 8000ft/440knots: 14fps 2D/ 6fps 3D Low vis being also a frame rate killer? Changing conditions: Vis (25000m), clouds at 5500ft(broken), thickness 600ft Flying @ 8000ft/440knots: 25fps 2D/ 16fps 3D Flying @ 6000ft/440knots: 20fps 2D/ 12fps 3D Hope this helps a bit, Markus Stuart Buchanan wrote: Stuart Buchanan wrote: Vivian Meazza wrote: And I have yet to see any 3d clouds. Any clues on where I should be looking (yes the box is checked :-)) Something has changed in the environment manager which means that clouds generateion is now inconsistent. I'm still tracking it down, as my recent changes shouldn't have affected this. Well, the cause was a bug in my code, but it didn't expose itself until we moved to multiple cameras. The attached patch fixes the problem. I've also put in a new heuristic to improve the frame-rate. Clouds that are already sorted are likely to still be sorted in subsequent frames. Therefore I've put in a back-off mechanism for the bubble-sort pass. This should mean that if you stay completely stationary, once the clouds become sorted they will eventually only perform a bubble sort pass every 128 frames. It would be good to get a feel for how bad performance is with 3D clouds. At the moment I don't have a handle on whether performance is almost good enough, or completely unacceptable. I'd appreciate it if people would post their observations, providing details of their machine spec, graphics options and the frame-rate with and without 3D clouds - ideally with a description of the general load on the machine. Thanks, -Stuart - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- Sent from my mobile device - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list
Re: [Flightgear-devel] Further 3D Clouds updates
Hi, Well, just tested, very nice, I withnessed an improvement in performance. My specs: Athlon XP2000, 768MB ram, NVidia FX 5900 @ 1680x1050 (system: linux with nvidia drivers), no AA Using: sg + fg cvs, OSG 2.7.5; everything enabled, custom settings, using metar, no ai/mp planes around Startup: 30fps 2D/ 20fps 3D (clouds at 6000ft, scattered, thickness 600ft) Flying 400ft/440knots: 26fps 2D/ 19fps 3D (clouds at 6000ft, scattered, thickness 600ft) Flying 6300ft/440knots: 35fps 2d/ 11-22fps (clouds at 6000ft, scattered, thickness 600ft) Note: Flying within the clouds resulted in 22fps, getting very close or through clouds 11fps Changing conditions: Low vis (7000m), clouds at 8000ft(broken) and 4000ft(broken), thickness 600ft Flying @ 8000ft/440knots: 14fps 2D/ 6fps 3D Low vis being also a frame rate killer? Changing conditions: Vis (25000m), clouds at 5500ft(broken), thickness 600ft Flying @ 8000ft/440knots: 25fps 2D/ 16fps 3D Flying @ 6000ft/440knots: 20fps 2D/ 12fps 3D Hope this helps a bit, Markus Here comes my test: Intel DualCore E6750 2.66 Ghz 3,6 GB RAM Nvidia 8800GT @1280x768 full AA FGFS-OSG win32 built by Fred from 11/26/2008 everything enabled Real wether fetch, framerate-throttle at 60fps Startup on TNCM: 60fps 2D/ 49fps 3D (fair weather, visibility: 25000m ) Flying 400ft/440knots: 60fps 2D/ 53fps 3D (fair weather, visibility: 25000m) Flying 6300ft/440knots: 60ps 2d/ 60fps (fair weather, visibility: 25000m) Note: Flying within the clouds no reducing of fps visible No with METAR enabled: LSZH startup cloud base at 1304ft thickness 600ft 38fps 2D/ 21fps 3d (visibility: 8000m) LSZH heading 160 8500ft altitude cloud base 1304ft thickness 600ft 48fps 2d/ 20fps 3d (visibility: 8000m) The same for beeing in the air about 2400-6500 ft altitude scattered cloudlayer seems to be a framerate-killer So I noticed the same like Markus: increasing visibility increases the fps! Regarding the ugly-borders: together with increasing visibility (decreasing fog) and better texture (third version in developement here on my pc) the ugly-borders nearly disappear and the clouds looks much better (fewer ugly edges in the clouds- sorting looks much better; much softer shading!) Is there a way to exclude the 3d-clouds by overdrawn from the fog? (like it is possible with the terrain?) http://www.flightgear.org/forums/download/file.php?id=382 Regards HHS - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Further 3D Clouds updates
Stuart Buchanan wrote: Vivian Meazza wrote: And I have yet to see any 3d clouds. Any clues on where I should be looking (yes the box is checked :-)) Something has changed in the environment manager which means that clouds generateion is now inconsistent. I'm still tracking it down, as my recent changes shouldn't have affected this. Well, the cause was a bug in my code, but it didn't expose itself until we moved to multiple cameras. The attached patch fixes the problem. I've also put in a new heuristic to improve the frame-rate. Clouds that are already sorted are likely to still be sorted in subsequent frames. Therefore I've put in a back-off mechanism for the bubble-sort pass. This should mean that if you stay completely stationary, once the clouds become sorted they will eventually only perform a bubble sort pass every 128 frames. It would be good to get a feel for how bad performance is with 3D clouds. At the moment I don't have a handle on whether performance is almost good enough, or completely unacceptable. I'd appreciate it if people would post their observations, providing details of their machine spec, graphics options and the frame-rate with and without 3D clouds - ideally with a description of the general load on the machine. Thanks, -Stuart clouds.diff.gz Description: GNU Zip compressed data - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Further 3D Clouds updates
Frederic Bouvier wrote Attached is a small patch for 3D clouds. It provide the following: 1) Proper spherical distribution of sprites (previously they were distributed cylindrically - whoops) 2) Better shading, so the bottom of the cloud is darker than the top. 3) Fixed a couple of texture sizing bugs. I don't think any of these are controversial, so if someone would commit it, I'd be grateful. I applied your patch. I noticed this message being repeated endlessly when clouds are activated : Warning: detected OpenGL error 'valeur non valide' after RenderBin::draw(,) I also noticed that even in the Thunderstorm scenario, the cloud coverage is very sparse. Maybe the message tells us it doesn't draw every cloud instance. In fair weather, I have no clouds ( my cloud density is set at 100% in the rendering options ) I've just applied he latest Sg patch on Win XP/msvc9; builds without error. I still get: Warning: detected OpenGL error 'invalid value' after RenderBin::draw(,) RenderStage::drawInner(,) FBO status= 0x8cd5 And I have yet to see any 3d clouds. Any clues on where I should be looking (yes the box is checked :-)) Vivian - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Further 3D Clouds updates
Hi, just to say that the detected OpenGL error is not new for me and is here from much more than one year (always built from cvs sources), I also remember that about one year later someone said to use grep -v to avoid the ugly output on the terminal, which I currently do with a launch bash script from this time (I think it was on IRC channel). best regards seb. Vivian Meazza wrote : Frederic Bouvier wrote I applied your patch. I noticed this message being repeated endlessly when clouds are activated : Warning: detected OpenGL error 'valeur non valide' after RenderBin::draw(,) I also noticed that even in the Thunderstorm scenario, the cloud coverage is very sparse. Maybe the message tells us it doesn't draw every cloud instance. In fair weather, I have no clouds ( my cloud density is set at 100% in the rendering options ) I've just applied he latest Sg patch on Win XP/msvc9; builds without error. I still get: Warning: detected OpenGL error 'invalid value' after RenderBin::draw(,) RenderStage::drawInner(,) FBO status= 0x8cd5 And I have yet to see any 3d clouds. Any clues on where I should be looking (yes the box is checked :-)) Vivian - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Further 3D Clouds updates
Hi Stuart, Stuart Buchanan a écrit : Hi All, Attached is a small patch for 3D clouds. It provide the following: 1) Proper spherical distribution of sprites (previously they were distributed cylindrically - whoops) 2) Better shading, so the bottom of the cloud is darker than the top. 3) Fixed a couple of texture sizing bugs. I don't think any of these are controversial, so if someone would commit it, I'd be grateful. I applied your patch. I noticed this message being repeated endlessly when clouds are activated : Warning: detected OpenGL error 'valeur non valide' after RenderBin::draw(,) I also noticed that even in the Thunderstorm scenario, the cloud coverage is very sparse. Maybe the message tells us it doesn't draw every cloud instance. In fair weather, I have no clouds ( my cloud density is set at 100% in the rendering options ) -Fred -- Frédéric Bouvier http://my.fotolia.com/frfoto/ Photo gallery http://fgsd.sourceforge.net/FlightGear Scenery Designer - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Further 3D Clouds updates
Frederic Bouvier wrote: I applied your patch. I noticed this message being repeated endlessly when clouds are activated : Warning: detected OpenGL error 'valeur non valide' after RenderBin::draw(,) I'm not sure what is going on here. I'm seeing the same warning, but I haven't yet managed to track it down. I also noticed that even in the Thunderstorm scenario, the cloud coverage is very sparse. Maybe the message tells us it doesn't draw every cloud instance. Possibly, though there is an existing bug that not all cloud tiles are being drawn. In fair weather, I have no clouds ( my cloud density is set at 100% in the rendering options ) This is a new bug, but I don't think it has been introduced by the patch. Still working on it :) -Stuart - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] Further 3D Clouds updates
Frederic Bouvier wrote:
Stuart Buchanan a écrit :
Hi All,
Attached is a small patch for 3D clouds.
It provide the following:
1) Proper spherical distribution of sprites (previously they were
distributed
cylindrically - whoops)
2) Better shading, so the bottom of the cloud is darker than the top.
3) Fixed a couple of texture sizing bugs.
I don't think any of these are controversial, so if someone would commit
it,
I'd be grateful.
I applied your patch. I noticed this message being repeated endlessly
when clouds are activated :
Warning: detected OpenGL error 'valeur non valide' after RenderBin::draw(,)
Fixed in the patch below. For some reason the shader didn't like index 16 being
used...
The patch also fixes the chequer-board effect that was causing very sparse
cloud cover.
I've still to work out why the Fair Weather scenario isn't displaying clouds
consistently.
-Stuart
Index: CloudShaderGeometry.hxx
===
RCS file:
/var/cvs/SimGear-0.3/source/simgear/scene/sky/CloudShaderGeometry.hxx,v
retrieving revision 1.3
diff -u -p -r1.3 CloudShaderGeometry.hxx
--- CloudShaderGeometry.hxx23 Nov 2008 12:14:57 -1.3
+++ CloudShaderGeometry.hxx24 Nov 2008 01:19:56 -
@@ -42,12 +42,12 @@ class CloudShaderGeometry : public osg::
{
public:
+const static unsigned int CLOUD_HEIGHT = 10;
const static unsigned int TEXTURE_INDEX_X = 11;
const static unsigned int TEXTURE_INDEX_Y = 12;
const static unsigned int WIDTH = 13;
const static unsigned int HEIGHT = 14;
const static unsigned int SHADE = 15;
-const static unsigned int CLOUD_HEIGHT = 16;
CloudShaderGeometry()
{
Index: cloudfield.cxx
===
RCS file: /var/cvs/SimGear-0.3/source/simgear/scene/sky/cloudfield.cxx,v
retrieving revision 1.22
diff -u -p -r1.22 cloudfield.cxx
--- cloudfield.cxx23 Nov 2008 12:14:56 -1.22
+++ cloudfield.cxx24 Nov 2008 01:19:56 -
@@ -157,14 +157,16 @@ SGCloudField::SGCloudField() :
}
}
+int leafs = QUADTREE_SIZE / BRANCH_SIZE;
+
for (int x = 0; x QUADTREE_SIZE; x++) {
for (int y = 0; y QUADTREE_SIZE; y++) {
field_group[x][y]= new osg::Switch;
field_group[x][y]-setName(3D cloud group);
// Work out where to put this node in the quad tree
-int i = (int) (BRANCH_SIZE * ((float) x) / ((float)
QUADTREE_SIZE));
-int j = (int) (BRANCH_SIZE * ((float) y) / ((float)
QUADTREE_SIZE));
+int i = x / leafs;
+int j = y / leafs;
quad[i][j]-addChild(field_group[x][y].get(), 0.0f, 2.0f);
}
}
-
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100url=/
___
Flightgear-devel mailing list
Flightgear-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/flightgear-devel
[Flightgear-devel] Further 3D Clouds updates
Hi All, Attached is a small patch for 3D clouds. It provide the following: 1) Proper spherical distribution of sprites (previously they were distributed cylindrically - whoops) 2) Better shading, so the bottom of the cloud is darker than the top. 3) Fixed a couple of texture sizing bugs. I don't think any of these are controversial, so if someone would commit it, I'd be grateful. Thanks, -Stuart simgear.patch.gz Description: GNU Zip compressed data - This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100url=/___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
