Re: [Flightgear-devel] Depth buffer issues with instruments and reindeer
Jim Wilson wrote:
http://www.spiderbark.com/fgfs/citation_instruments.png
It turns out these are in fact contained in a single 3D model for the entire
aircraft, so it has nothing to do with 2D. Apparently the problem is in the
models. ...
FWIW I'd like to suggest that it is a good idea for 3D modelers to test their
work at 16 bit depth buffer settings since a lot of folks are still running
modern laptop, consumer grade and Intel embedded chips at 16 bit for
performance reasons. Even though it involves moving layers further appart,
adjusting 3D instrument models to support 16 bit generally does no harm to the
appearance of the model at the normal viewing distance.
I disagree on changing the models or instruments. Looking at the code
the problem is obvious.
We ask for a depth buffer precision that is impossible to achieve. From
FG/Model/acmodel.cxx :
FGAircraftModel::FGAircraftModel ()
: _aircraft(0),
_selector(new ssgSelector),
_scene(new ssgRoot),
_nearplane(0.01f),
_farplane(1000.0f)
{
}
Jim, if you can compile FG, can you try with a near plane of 0.03 and/or
a far plane of 250.0 ?
Harald.
___
Flightgear-devel mailing list
Flightgear-devel@flightgear.org
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Depth buffer issues with instruments and reindeer
Harald JOHNSEN wrote:
Jim Wilson wrote:
http://www.spiderbark.com/fgfs/citation_instruments.png
It turns out these are in fact contained in a single 3D model for the
entire aircraft, so it has nothing to do with 2D. Apparently the
problem is in the models. ...
FWIW I'd like to suggest that it is a good idea for 3D modelers to
test their work at 16 bit depth buffer settings since a lot of folks
are still running modern laptop, consumer grade and Intel embedded
chips at 16 bit for performance reasons. Even though it involves
moving layers further appart, adjusting 3D instrument models to
support 16 bit generally does no harm to the appearance of the model
at the normal viewing distance.
I disagree on changing the models or instruments. Looking at the code
the problem is obvious.
We ask for a depth buffer precision that is impossible to achieve.
From FG/Model/acmodel.cxx :
FGAircraftModel::FGAircraftModel ()
: _aircraft(0),
_selector(new ssgSelector),
_scene(new ssgRoot),
_nearplane(0.01f),
_farplane(1000.0f)
{
}
Jim, if you can compile FG, can you try with a near plane of 0.03
and/or a far plane of 250.0 ?
For what it's worth, changing the far plane has little affect on the
depth buffer precision. The near buffer value is what dominates the
amount of depth buffer precision.
Curt.
--
Curtis Olsonhttp://www.flightgear.org/~curt
HumanFIRST Program http://www.humanfirst.umn.edu/
FlightGear Project http://www.flightgear.org
Unique text:2f585eeea02e2c79d7b1d8c4963bae2d
___
Flightgear-devel mailing list
Flightgear-devel@flightgear.org
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Depth buffer issues with instruments and reindeer
From: Harald JOHNSEN
snip
FGAircraftModel::FGAircraftModel ()
: _aircraft(0),
_selector(new ssgSelector),
_scene(new ssgRoot),
_nearplane(0.01f),
_farplane(1000.0f)
{
}
Jim, if you can compile FG, can you try with a near plane of 0.03 and/or
a far plane of 250.0 ?
Ah, good catch. I've got a feeling this was changed unintentionally. It came
in under a log entry that describes curts time of day patches from 2 years ago.
Interesting that others aren't seeing this though...
Best,
Jim
___
Flightgear-devel mailing list
Flightgear-devel@flightgear.org
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Depth buffer issues with instruments and reindeer
Doh! Let me try this again...I missed a big chunk of the 2nd sentence in the last message. Hopefully this will make more sense: Ok, there really is a problem with the model. The distance between the instrument face surface on the clock (and other instruments in the Citation II cockpit) and the panels backplane surface is only about 0.0001 units. --- --- -- - --- That nearplane value for the aircraft model should probably be 0.10f (about 3 inches) instead of the 0.01f that it is now. Best, Jim ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Depth buffer issues with instruments and reindeer
From: Jim Wilson
From: Harald JOHNSEN
snip
FGAircraftModel::FGAircraftModel ()
: _aircraft(0),
_selector(new ssgSelector),
_scene(new ssgRoot),
_nearplane(0.01f),
_farplane(1000.0f)
{
}
Jim, if you can compile FG, can you try with a near plane of 0.03 and/or
a far plane of 250.0 ?
Ah, good catch. I've got a feeling this was changed unintentionally. It
came in under a log entry that describes curts time of
day patches from 2 years ago.
Interesting that others aren't seeing this though...
Ok, their really is a problem with the model. The distance between the
instrument face surface on the clock (and other instruments in the Citation II
cockpit) is about 0.0001 units.
That nearplane value for the aircraft model should probably be 0.10f (about 3
inches) instead of the 0.01f that it is now.
Best,
Jim
___
Flightgear-devel mailing list
Flightgear-devel@flightgear.org
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d
[Flightgear-devel] Depth buffer issues with instruments and reindeer
Just thought I'd mention a couple things. 1) Some of the aircraft panel instruments display z-fighting even at the higher depth buffer setting... # grep Depth XFree86.0.log (**) NVIDIA(0): Depth 24, (--) framebuffer bpp 32 (--) Depth 24 pixmap format is 32 bpp ...and I'm wondering if there's been some change in the 2D panel code or something else that might be producing this. The Citation II seems to be especially bad. Are other folks seeing this? 2) Raindeer in the santa model should be spelled reindeer. It might be a common misspelling but it is still a misspelling ;-) Best, Jim ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Depth buffer issues with instruments and reindeer
Jim Wilson wrote: 2) Raindeer in the santa model should be spelled reindeer. It might be a common misspelling but it is still a misspelling ;-) Ehr Alright, but only because it's you who mentioned it. I could have called it rendier instead as it's called in Dutch, but I chose to call it after Rainman I guess ... ;-) Erik ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Depth buffer issues with instruments and reindeer
Jim Wilson wrote: Just thought I'd mention a couple things. 1) Some of the aircraft panel instruments display z-fighting even at the higher depth buffer setting... # grep Depth XFree86.0.log (**) NVIDIA(0): Depth 24, (--) framebuffer bpp 32 (--) Depth 24 pixmap format is 32 bpp ...and I'm wondering if there's been some change in the 2D panel code or something else that might be producing this. The Citation II seems to be especially bad. Are other folks seeing this? Depth tests are disabled when drawing 2D or 2.5D panels. I've checked the Citation-II, I can not see any z-fighting. Harald. ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
Re: [Flightgear-devel] Depth buffer issues with instruments and reindeer
From: Harald JOHNSEN Jim Wilson wrote: Just thought I'd mention a couple things. 1) Some of the aircraft panel instruments display z-fighting even at the higher depth buffer setting... # grep Depth XFree86.0.log (**) NVIDIA(0): Depth 24, (--) framebuffer bpp 32 (--) Depth 24 pixmap format is 32 bpp ...and I'm wondering if there's been some change in the 2D panel code or something else that might be producing this. The Citation II seems to be especially bad. Are other folks seeing this? Depth tests are disabled when drawing 2D or 2.5D panels. I've checked the Citation-II, I can not see any z-fighting. http://www.spiderbark.com/fgfs/citation_instruments.png It turns out these are in fact contained in a single 3D model for the entire aircraft, so it has nothing to do with 2D. Apparently the problem is in the models. I'll have to go back and make a list of the others. Of course now this raises a question about how my nvidia card is operating a 16bpp buffer when the logs all show it is configured properly. FWIW I'd like to suggest that it is a good idea for 3D modelers to test their work at 16 bit depth buffer settings since a lot of folks are still running modern laptop, consumer grade and Intel embedded chips at 16 bit for performance reasons. Even though it involves moving layers further appart, adjusting 3D instrument models to support 16 bit generally does no harm to the appearance of the model at the normal viewing distance. Best, Jim ___ Flightgear-devel mailing list Flightgear-devel@flightgear.org http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
