Celestia Forums

I just installed pre11 on my system, and still found some old bugs to be there. Here are some:

Orbit cutting, seen with OrbitPathSamplePoints and RingSystemSections set to higher values:

100

150

200

300

and so on ...

Emissive Models with planetary bodies in sight:

Cassini with Earth - not always:

Cassini with Earth - but a little turn:

Cassini with Earth just outside of view:

Cassini with Moon:


maxim

Funny, I reported both bugs (I assume they are the same) a couple of days ago on the developers mailing list...

maxim wrote:Emissive Models with planetary bodies in sight:
[...]

Can you confirm that this happens only when using the OpenGL vertex program (possibly in combination with NVidia combiners) renderpath?

I have problems getting this bug with Cassini, ISS works much "better":
Cassini (only Huygens is rendered wrong):
cel://follow/Sol%3ACassini/2004-08-08T19%3A55%3A35.23074?x=vO2O0t8UEdqQDA&y=FrDx5uKHE1T//////////w&z=bi+be+rX2kx4/////////w&ow=-0.788470&ox=0.069537&oy=-0.601787&oz=-0.106449&select=Sol:Cassini&fov=7.458925&ts=1.000000<d=0&rf=56247&lm=0
ISS:
cel://follow/Sol%3AEarth%3AISS/2004-08-08T19%3A56%3A23.35931?x=k3uh/HNrukvJDA&y=CuGGKBc59f///////////w&z=Pv8qm61CBgYL&ow=0.994105&ox=-0.026193&oy=0.079938&oz=-0.068411&select=Sol:Earth:ISS&fov=32.545578&ts=1.000000<d=0&rf=56247&lm=0

(I wish it would be possible to insert these URLs without breaking the layout - sorry)

Harald

Harald,

Are both of your URLs supposed to exhibit this problem?

Because I don't see it with either of them. Everything seems to be properly shaded and not glowing on my system.

System:
256MB 500MHz P3, WinXP Pro SP1
128MB GF4 Ti4200, drivers v61.76

Celestia built from cvs today
Celestia v1.3.2pre11
Celestia v1.3.2pre10

Added slightly later:
I don't see the extra orbit lines, either, with any of these versions of Celestia.

Which versions of drivers are you guys using?
Maybe they're both due to driver bugs?

Harry wrote:Funny, I reported both bugs (I assume they are the same) a couple of days ago on the developers mailing list...

maxim wrote:Emissive Models with planetary bodies in sight:
[...]

Can you confirm that this happens only when using the OpenGL vertex program (possibly in combination with NVidia combiners) renderpath?

I have problems getting this bug with Cassini, ISS works much "better":
Cassini (only Huygens is rendered wrong):
cel://follow/Sol%3ACassini/2004-08-08T19%3A55%3A35.23074?x=vO2O0t8UEdqQDA&y=FrDx5uKHE1T//////////w&z=bi+be+rX2kx4/////////w&ow=-0.788470&ox=0.069537&oy=-0.601787&oz=-0.106449&select=Sol:Cassini&fov=7.458925&ts=1.000000<d=0&rf=56247&lm=0
ISS:
cel://follow/Sol%3AEarth%3AISS/2004-08-08T19%3A56%3A23.35931?x=k3uh/HNrukvJDA&y=CuGGKBc59f///////////w&z=Pv8qm61CBgYL&ow=0.994105&ox=-0.026193&oy=0.079938&oz=-0.068411&select=Sol:Earth:ISS&fov=32.545578&ts=1.000000<d=0&rf=56247&lm=0

(I wish it would be possible to insert these URLs without breaking the layout - sorry)

Harald

Harald,

sorry, but on my system everything looks right with your cel://urls (as I wrote you already this afternoon). I also toggled through all render paths...NOTHING.

My card: FX 5900Ultra/256MB, latest NV driver.

OrbitPathSamplePoints 1000
RingSystemSections 500
ShadowTextureSize 512
EclipseTextureSize 256

Bye Fridger

Harry wrote:Can you confirm that this happens only when using the OpenGL vertex program (possibly in combination with NVidia combiners) renderpath?

Yes, I can confirm that.

selden wrote:Which versions of drivers are you guys using?

Not the newest one - still 53.03 - Wasn't there a problem with 61.xx drivers and MX cards??
I'll gonna verify those with the newest drivers.

maxim

PS: You'll always find my system setup in my profile.

selden wrote:Are both of your URLs supposed to exhibit this problem?

Yes, see these screenshots:
http://www.h-schmidt.net/celestia/tmp/cassini.jpg
http://www.h-schmidt.net/celestia/tmp/iss.jpg

System:
256MB 500MHz P3, WinXP Pro SP1
128MB GF4 Ti4200, drivers v61.76

I see these problems on Linux, 32MB GF2MX, driver version 1.0-5336 - I think they occured on windows too, but I can't test ATM (and don't know which driver version I am using).

I don't see the extra orbit lines, either, with any of these versions of Celestia.
Which versions of drivers are you guys using?
Maybe they're both due to driver bugs?

I am sure I've seen them on both Linux and Windows, but it's possible that it's a cross-platform driver bug

Harald

t00fri wrote:sorry, but on my system everything looks right with your cel://urls

I probably should upgrade the NV driver on Windows and see if that solves the problem. So much for "never change a working system"

(as I wrote you already this afternoon).

I received a mail regarding the splash-screen, but none regarding the bugs.

Harald

Harald,

my NV driver is the /latest/ : 1.0-6106 that also sports this neat GUI for setting opengl parameters like anti-alias on the fly...

Bye Fridger

PS: I never saw any orbit crossings also on my XP-laptop (ATI mobile card) that knows however only 'basic' and 'multitexture' render paths...

Well, it seems both Harald and Maxim are seeing two bugs with Nvidia MX cards and v53 drivers.

Neither Fridger nor I see them with v61 drivers and Nvidia FX and Ti cards respectively.

It'd be nice if someone else could report whether or not they see them and what cards and driver versions they have.

selden wrote:Well, it seems both Harald and Maxim are seeing two bugs with Nvidia MX cards and v53 drivers.

Neither Fridger nor I see them with v61 drivers and Nvidia FX and Ti cards respectively.

Just upgraded to the latest&greatest: v61.77.

Bugs are still there.

Any ideas?

Harald

Since I installed 1.3.2 pre11 a week or so ago, I've got to say that I don't appear to be having any of the problems highlighted above, for either the models or the orbit paths.

I am currently using the "Open GL vertex program / Nvidea Combiners" render path.

I have altered some of the default .cfg settings to;


I also have my driver settings for antialiasing set to application controlled and have antialiasing switched on within Celestia.

Here's the details of the system I'm using ;

AMD Athlon 1800 256mb DDR:
256mb DDR Nvidea GeForce.FX5200 running on Nvidea 61.77 Drivers
WinXPHome:
FullScreen 1280x1024x32

regards,
TERRIER

Same as Harry:

Installed V61.77 of nvidia drivers, but the bugs are still there.
Seems to be a special MX cards issue.

maxim

Does it make a difference when you change rendering modes va CTRL + V?

I will see if these show up on my system tonight...

I can also report some rendering bugs tonight when I get the chance...One of which has to do with a central star and models overlaying the same position coded as such in ssc files not dsc files...The central star will dissapear at certain angles when the observer is located within the bounding box of the model.

Rassilon wrote:Does it make a difference when you change rendering modes va CTRL + V?

The rendering problems only occur when using the OpenGL vertex program renderpath, either alone or in combination with NVidia combiners.

The orbit-lines happen with all renderpaths.

Harald

Just tried the cel URLs on my ATI 9700 WinXP system and they work okay -- on the three render paths for ATI cards.

I also have the model shading problem and the orbit line problem.
I have Win XP, Pentium4 1.8gHz, 512MB ram, GeForce4 MX440 with latest 6177 driver. The orbit line problem is new to me, but I think I had the shading problem on earlier Celestia releases.
The shading problem only occurs with the OpenGL Vertex renderpaths, but the orbit line problem is always there.

The interesting thing is that the the shading problem goes away if I turn off the clouds ???

Dave.

Dave-Montreal wrote:The interesting thing is that the the shading problem goes away if I turn off the clouds ???

Yes, that often happens, but not in all cases. I couldn't track down a definite behaviour there - it seems also that there are distance slots, where the problem vanishes and arises again.

maxim

I'm looking into both of these problems, but I can't get the extra orbit lines to show up. To anyone who is able to reproduce the orbit bug: does it happen for asteroid orbits, or just the major planets of our solar system?

--Chris

For all eliptic orbits as it seems: planets, moons, spacecrafts, asteroids and comets
Not for xyz trajectories.

Do you need example pics?

maxim

Sorry, that was me!

maxim