Bugs found in Windows v1.5.0 prerelease 1

[selden]

These are in the order that I happened to test them, using some old bug demos. Many old bugs found in previous versions of Celestia aren't present, of course!

1. Cel url compatibility is needed: most of the tests I have use urls to set up the viewpoint. None work any more :(


This is new:
2. Although there are no error messages in the console log, this star not drawn; Celesrtia displays bizarre radius & magnitude values.

Abs (app) mag: 1#J (1#J)
Radius: -1#JRsun

Code: Select all

Star "CylStar"
{
        OrbitBarycenter "Sol"
        SpectralType "M5"
        AppMag -4.5


        EllipticalOrbit {
                Period            1.0000
                SemiMajorAxis     1.0000
        }
}

The folowing 4 problems are demonstrated by two Addons contained in the Zip file
http://www.lepp.cornell.edu/~seb/celest ... issues.zip
(1.2MB)

3. From some viewpoints, smooth edges are rendered as jagged intrusions. See the folder "model_issues"

For items 4-6, see the folder "positioning_errors"

4. The numeric precision of longitude and latitude is inadequate in LongLat: objects are placed at different long,lat positions when only their altitude values differ.

5. the same problem, buut worse, is seen for Locations: their locations vary in realtime: they jump around so much they're unreadable.

6. Getting too close to a location label causes it to disappear.

[chris]

Selden,

Are you saying that these are all old bugs?

--Chris

[selden]

1 is new to 1.5, but has been complained about for a while, of course.

2 is new to me.

3 is new to 1.5. I'd been intending to mention it for a while, but kept hoping it'd disappear while you were making frequent changes to the depth sorting code.

(There's another depth sorting issue that i'll report shortly, which I've stumbled over in one of my as-yet-unmentioned projects.)

4 & 5 have been around since 1.4 at least. I was pretty sure I'd reported those positioning resolution limitations to the developers' list around the time of 1.4.1.

6 is newly being reported: I think I noticed it a while ago but was having problems characterizing it. I managed to create a url for it last night while testing 4 & 5. I'm wondering if it might be somehow related to the Shuttle rendering issue reported in another thread yesterday.

[chris]

This is new:
2. Although there are no error messages in the console log, this star not drawn; Celesrtia displays bizarre radius & magnitude values.

Abs (app) mag: 1#J (1#J)
Radius: -1#JRsun

Code: Select all

Star "CylStar"
{
        OrbitBarycenter "Sol"
        SpectralType "M5"
        AppMag -4.5


        EllipticalOrbit {
                Period            1.0000
                SemiMajorAxis     1.0000
        }
}

I'm surprised that it didn't happen in old versions of Celestia. The problem with the definition is that Celestia needs to calculate the intrinsic brightness of the star from the apparent magnitude, but it can't do so when it's right at the origin. You should specify the absolute magnitude instead. I changed the stc parser to print an error message and reject stars near the origin that have their brightness specified as an apparent rather than absolute magnitude.

--Chris

[selden]

7.

Here's the example of another depth sorting bug



a) An emissive surface is beneath an opaque one but is drawn partially above it: note the stripes on the diagonal bars.

b) A similar but not identical problem is described in #3 above: intersections of models sometimes are drawn with jagged borders, although the edges are straight. Note the edges of the rectangular cutouts in the cylinder in the lower portion of the image.

An Addon to demonstrate these two problems is available at http://www.lepp.cornell.edu/~seb/celest ... orting.zip (1.3MB)

[Joe]

Allow me to add on seldon's list

8. there is no comet tails (for Halley). The buit from latest CVS has the tail displayed, but it looks there is a polygon back-face culling problem:

[t00fri]

h?¶ h?¶...

that looks more like a door to another "baby universe" rather than a comet tail

+++++++++++++++++++++++++
Note the WRONG star brightness! It appears that with
some background lighting the stars are shining much
brighter than without. The opposite is true of course.
+++++++++++++++++++++++++

Bye Fridger

[Johaen]

I'm not sure if this is a glitch, or more just a misunderstanding on my part on what it should look like.



This screenshot is taken from the back side of Titan, with Ambient Light set to off. Note the Sun is over the limb. Shouldn't the back side of Titan be completely black? Or is there enough light scatter where it would actually be bright like that?

[t00fri]

I'm not sure if this is a glitch, or more just a misunderstanding on my part on what it should look like.



This screenshot is taken from the back side of Titan, with Ambient Light set to off. Note the Sun is over the limb. Shouldn't the back side of Titan be completely black? Or is there enough light scatter where it would actually be bright like that?

Sure! That's the whole great news about modelling the strong amount of backscattering in Titan's atmosphere.

Here is a TRUE photo reconstructed from the R,G,B filtered images by Cassini. It was done by our old friend Matt McIrvin:



Bye Fridger

[Johaen]

Here is a TRUE photo reconstructed from the R,G,B filtered images by Cassini. It was done by our old friend Matt McIrvin:

Except to me, it looks like the back side is completely back. Or really close. Certainly nowhere near as bright as it is in my screenshot.

Also, another thing I noticed, while I was looking at solarsys.ssc to see if I can figure out how that backscattering works:

Code: Select all

"Titan" "Sol/Saturn"
{
   Texture  "titan.*"
        Color      [ 0.8 0.2 0.45 ]
       # HazeColor [ 0.781 0.629 0.975 ]
   HazeDensity          0.6
   Radius            2575

    Atmosphere {
        Height         500.0
         Lower      [ 0.477 0.367 0.211 ]
        Upper      [ 0.96 0.805 0.461 ]
        Sky        [ 0.3 0 0 ]
     
         Mie              0.0001
         MieAsymmetry    -0.55
         Rayleigh   [ 0.0 0.0 0.00017 ]
         Absorption [ 0.000075 0.00030 0.00025 ]
         MieScaleHeight 220.0
         CloudHeight    220.0
         CloudSpeed      65.0
         CloudMap "titan-clouds.*"
    }
   
   CustomOrbit "titan"
   EllipticalOrbit
   {
   Period            15.945421
   SemiMajorAxis 1221850
   Eccentricity       0.0292
   Inclination        0.33
   MeanAnomaly      120
   }

   Obliquity        0.6
   EquatorAscendingNode   186.6
   RotationOffset      318.9

   Albedo                    0.21
}

Is there a reason the haze color is commented out for Titan? It's not comented out any other place in the .ssc.

[t00fri]

...
Except to me, it looks like the back side is completely back. Or really close. Certainly nowhere near as bright as it is in my screenshot.

How about your ambient illumination setting?? Hit the '{' key.
Also all this is for the OpenGL 2.0 render path, where HazeColor does not work

Bye Fridger

[danielj]

I tested with Beteugelse and nothing happened.It??s a problem related with the Sun.
Look:
http://img501.imageshack.us/my.php?imag ... arszw5.jpg
This image is from the star type Fuzzy point and Open GL 2.0.Only with normal and multitexture,the sun didn??t appear that large.So it??s not only a problem with scaled disks.

Talking about that,look at these:
http://img524.imageshack.us/my.php?imag ... cesqp8.jpg
It looks like the Sun??s corona...

With scaled disks,this is the Sun with multitexture render path,as seen from Venus:
http://img524.imageshack.us/my.php?imag ... osezo3.jpg

And this,with scaled disks and Open GL 2.0 path:
http://img524.imageshack.us/my.php?imag ... ovath7.jpg

Very weird,indeed.

[Johaen]

How about your ambient illumination setting?? Hit the '{' key.

This screenshot is taken from the back side of Titan, with Ambient Light set to off.

Like I said, maybe it's just me, and the way I'm looking at it. My monitor at home is a pretty decent CRT. But here at work I have a nice LCD, and my screenshot doesn't look anywhere near as bright as it does at home, although it's still brighter than that picture you posted. Maybe it's just me... *shrug*

[Adirondack]

With my latest CVS version (Dec 7th) the Earth turns red when cloudshadows and nightmaps
are enabled in OpenGL2.0 render path.
When I switch off cloudshadows or nightmaps everything is fine with Earth.

XP SP2, GT6600

Adirondack

[rra]

I have the very same problem ,
also turning off the atmosphere and leaving the other
options enabled helps in restoring the normal Earth color.


Ren?©

(also XP SP2 GT6600 !!)

[selden]

Ren?© and Adirondack,

My home system also uses a GF 6600GT, with ForceWare v91.47.

Using CVS code built at 6am on December 8th, I see the shader error that you report when

1. I have another star orbiting the Sun, so that there are two or more light sources illuminating the Earth

[edit]
OR

2. if a solar eclipse is in progress
which makes this a rather more urgent problem.

Are either of these what you are doing?
[/edit]
If not, it might help if you included a copy of shaders.log and a description of any Addons that you're using.

(My understanding is that eliminating the limit in the number of shaders that can be applied to a surface requires a major rewrite of how they are applied. I suspect this won't be fixed in 1.5.0)

p.s. Ren?©, I deleted your duplicate post.

[chris]

For items 4-6, see the folder "positioning_errors"

4. The numeric precision of longitude and latitude is inadequate in LongLat: objects are placed at different long,lat positions when only their altitude values differ.

5. the same problem, buut worse, is seen for Locations: their locations vary in realtime: they jump around so much they're unreadable.

6. Getting too close to a location label causes it to disappear.

These bugs should all be fixed by my recent checkins.

--Chris

[ElChristou]

Fresh compile from today, orbits disapears playing with fov and rotation in an unexpective way... (specially between 12h to 20h on the celestial grid)

[t00fri]

Chris,

it would be VERY valuable to know WHAT bugs you fixed precisely in your large checkin tonight. Are these just the 3 you were mentioning above?

How about your star code? Anything improved there?

Bye Fridger

[t00fri]

Here is glorious

"Titan and it's atmosphere" in 1.5.0pre1 CVS (5 minutes ago)



My feeling is that things tend to get worse somehow...

Bye Fridger