Moon eclipse

[Mad Boris]

There's been a patch posted on the add-on repository which apparently correctly renders the appearance of moons during lunar eclipses, but you have to modify the source code and compile it to get it to run. I can't do this : is there any way of integrating the patch directly into the program ?

[MB]

I am the guy who makes this try to render eclipse. It is straightforward to compile under Linux, but since the "Windows version" is ahead of the "Linux version ", I got frustrated and despite several hours to try to compile with visual C++ (6.0) for windows, every .exe crash!!!. If some guy have time to do such compile, Then the .exe can be added to the package.
MB

PS
If some people can help me in the compilation:
I had no problem with the previous 1.2.2 version if I remember well (before the introduction of the various (celengine,...) folders.
With 1.2.3 and 1.2.4:
after having introduced the correct "include" and "library" pathes the only "warning" at link is something like
possible conflict with LIBC use NOLIBRARY : library (or something like that.
At the exe, The "celestia" frame shows up but freezes imediately with the the previous backgroud inside the frame.
May be the exe does not find the OpenGL libs??? Thanks again for any help.

[t00fri]

I am the guy who makes this try to render eclipse. It is straightforward to compile under Linux, but since the "Windows version" is ahead of the "Linux version ", I got frustrated and despite several hours to try to compile with visual C++ (6.0) for windows, every .exe crash!!!. If some guy have time to do such compile, Then the .exe can be added to the package.
MB

PS
If some people can help me in the compilation:
I had no problem with the previous 1.2.2 version if I remember well (before the introduction of the various (celengine,...) folders.
With 1.2.3 and 1.2.4:
after having introduced the correct "include" and "library" pathes the only "warning" at link is something like
possible conflict with LIBC use NOLIBRARY : library (or something like that.
At the exe, The "celestia" frame shows up but freezes imediately with the the previous backgroud inside the frame.
May be the exe does not find the OpenGL libs??? Thanks again for any help.

I have compiled your patched render.cpp in the latest CVS version of 1.2.4. under Linux. I had no problems during the compilation, but the program performance was quite unstable near a moon eclipse.

So I skipped it again without studying your code in detail for reasons of time...

Bye Fridger

[MB]

Fridger,
Thanks for your comment,
I did not get any stability problem under linux (checked on two PC with nv GEFORCE 2 and 3, respectively ( Mandrake 8.0). However I used the downloaded version of celestia 1.2.4. and not the CVS one that I don't know how to use.
MB

[t00fri]

Fridger,
Thanks for your comment,
I did not get any stability problem under linux (checked on two PC with nv GEFORCE 2 and 3, respectively ( Mandrake 8.0). However I used the downloaded version of celestia 1.2.4. and not the CVS one that I don't know how to use.
MB

If I find a little more time, I may try to specifically reproduce the instability. What I did qualitatively was to move to the earth's surface and look towards the moon around a known total moon eclipse. I observed the appearance of a sharply limited, dark horizontal (shadow) band, spreading out over large parts of the sky shortly before the eclipse. While I could observe I nice appropriate yellow-brown colorization of the moon due to Umbra and penumbra effects, the display froze at some point which it normally never does on my environment;-)
Thus the issue moved down (somewhat;-)) on my priority list...


Bye Fridger

[t00fri]

I recompiled your render.cpp and went again to the total moon eclipse visible in Europe, dated January 21, 2000, penumbra entry start 2:02 UT.

This time I could not detect any fault in the program, the rendering looks quite natural, indeed.

I probably was too fast last time, due to lack of time. Sorry.

Bye Fridger

[MB]

To Mad Boris,
I finally succeed in compiling celestia for "Windows" ( final 124 version) with this rendering of the moon eclipse.
I just send it to Bruckner, in order that I put the new zipped file with the .exe in the "add-on" repository.
MB

[bruckner]

Just a brief note: it's up there, already... http://bruckner.homelinux.net/addons.html

Greetings!

Bruckner

[MB]

Warning!
I discovered after my last post yesterday that they should have some bug in my "windows" version. This bug, that does not show up in the "linux" version, appears as "yellow patches" at the outer border the penumbra. This is the reason why I did not see it in too quick check
I will look more when time permits.
SORRY!
MB

[Mad Boris]

Thanks for the patch, MB ! I think its great, the yellow areas are a fairly minor problem. There was a similar problem with the gas giants in the previous version of Celestia : zooming up close, a large blue area sometimes developed (not sure but I think this only occured during eclipses).

Would it be possible to randomize the brightness ? Eclipses can be quite bright to very dark, depending on the amount of dust in the atmosphere. Any way of making the brightness vary randomly within a specified range during the eclipse ?

[MB]

Answer to Mad Boris
I partially improved the eclipse rendering but not perfectly, now the yellow patches appear only "after" the eclipse. But I will still work on it before replacing it in the Bruckner repository.
BTW the error cames from the fact that contrarily to the 124 win.exe versions that got, with time,different labelling e.g. 124 pre1, 124 pre2, etc, the source code got the same name ....124.tar.gz and only differ by the date. I therefore mixed up two 124.tar.gz versions of different dates! I hope to have corrected this error but I am still not sure that the stable 124 Win version is built with the very same code sources as in the stable 124.tar.gz. release. Chris is it correct?
In any case I will look more carefully through the codes.
Concerning the light blue hint, I also find that, my code is probably too simple. Again I will look at that.
The last point concerns a random brigtness of the color. This seems probably impossible without a profond change of the celestia code that is outside my "amateur" programming technique. Actually the eclipse are hard coded in celestia and cannot be changed easily. This would be possible by creating the eclipse texture on fly as suggested by Chris. But again that is beyong my programming skill.
MB
MB

[MB]

The bug giving the yellow patches is now corrected. I send the modified celestia 1.2.4 windows version and code to "Bruckner" to put it in his repository. Thanks Bruckner again.
MB

[CelestiaFan]

What do I so? I am so lost on this program when it comes to eclipses, is there some add-on I should get to view eclipses under the Celestia 1.2.4? I can't even get a single moon shadow on any planet? Any suggestions, thanks!

[Guest]

What do I so? I am so lost on this program when it comes to eclipses, is there some add-on I should get to view eclipses under the Celestia 1.2.4? I can't even get a single moon shadow on any planet? Any suggestions, thanks!

With the following URL

http://www.celestiaproject.net/~claurel/eclipse/

you have at least some events that you should be able to reproduce.
----------------------------------
You must of course make sure that you place yourself at the right place. Otherwise you cannot see anything at the nominal time.
Also you must allow some time uncertainty, since the light travelling time is not yet included.
----------------------------------

You should also not forget to switch from local to universal time (UT =GMT) first!

Here is a sketch of how to proceed:

Turn the earth to the night side, push Y to lock it, beam yourself close to the surface by means of the 'Home' key watching the altimeter display while pushing.
Then push the '*' key to turn around and look for the sky by adjustments with the 'Up Down Left Right' keys. Then type in the
appropriate UT time into the time set dialog as well as the
planet you want to watch, hit return and type C to center it in the display. Then use the telescope mode (drag with shift Left mouse ) until you see the planet and the moons in high magnification.

Note: If you want to whatch solar or lunar eclipses you must really look from the /precise/ location at the /right/ time as given in respective tables (e.g. Sky & Teslecope). It should be obvious why this is so;-)

That's it

Good luck

Bye Fridger