New Celestia-1.4.0pre-FT1.2 Version for Download

[brokfn]

Hi,

I've got the same problem as Kolano : an excessive brightness of the galaxies when an add-on is present. In my case I only had the M1 add-on installed (Don't remember whose is it). The center of the galaxies is then so bright that it appears to be white. When I uninstall the add-on, everything is ok.

Thanks Fridger and Toti for the wonderful job you've done.
That shows that incredibly young people can manage to do valuable developpements.

Bye, Brokfn

[Boux]

OK, I have just rebuilt a very basic Linux system.
Only Glut build is available (by the way I love its simplicity, pure cli launch and stuff, hehe...).
Rendering is gorgeous and fps figures have never been so high
..... buuuuut(t).... there are a couple of bugs.
1- I get a console error that says "Error opening solar system catalog". Everything is loading and showing up properly where expected but it is not possible to label planets and moons, and the right-click context menus are lost (satellites, alternate surfaces, etc.). Keyboard shortcuts for them don't work either.
2 - Binary cmod models don't show up as previously stated about my previous kde build.
Hummmm, as this is a very basic barebone system, I am puzzled.
Would my graphics card have something to do with this (NVidia 6800 Ultra)?
The drivers are the latest 7676 and Xorg is 6.9.xx.
3 - Console log also says:
libpng warning: Incomplete compressed datastream in iCCP chunk
libpng warning: Profile size field missing from iCCP chunk
No other error messages. Console log says that all NV extensions are nicely loaded.
I would add: "no problem, no fun"

[selden]

An Nvidia 6800 should be fine, except for possible bugs in the driver you're using. However, that shouldn't be affecting how binary CMOD files are handled.

I fear it'll take a real debugging session to determine what's failing: putting a breakpoint where the file is being read and looking at the results in memory.

Both the ASCII and binary CMOD formats should be translating into the same internal format, but it's obvious that they aren't.

Have you tried running cmodfix on the ASCII version of a file and comparing its output with a binary CMOD file produced on another system?

[t00fri]

...
Thanks Fridger and Toti for the wonderful job you've done.
That shows that incredibly young people can manage to do valuable developpements.

Bye, Brokfn

Thanks for the "flowers", but what what do you mean with incredibly young people ??

I am 15 aleady, soon becoming a real man (!) and also I have to do something about my growing beard once every week already!

Bye Fridger

[t00fri]

OK, I have just rebuilt a very basic Linux system.
Only Glut build is available (by the way I love its simplicity, pure cli launch and stuff, hehe...).
Rendering is gorgeous and fps figures have never been so high
..... buuuuut(t).... there are a couple of bugs.
1- I get a console error that says "Error opening solar system catalog". Everything is loading and showing up properly where expected but it is not possible to label planets and moons, and the right-click context menus are lost (satellites, alternate surfaces, etc.). Keyboard shortcuts for them don't work either.
2 - Binary cmod models don't show up as previously stated about my previous kde build.
Hummmm, as this is a very basic barebone system, I am puzzled.
Would my graphics card have something to do with this (NVidia 6800 Ultra)?
The drivers are the latest 7676 and Xorg is 6.9.xx.
3 - Console log also says:
libpng warning: Incomplete compressed datastream in iCCP chunk
libpng warning: Profile size field missing from iCCP chunk
No other error messages. Console log says that all NV extensions are nicely loaded.
I would add: "no problem, no fun"

Hi Boux,

wait a moment...

The glut-only build is NOT supported anymore since long.

Certainly, the distributed FT1.2 version --if correctly installed--

1) does NOT produce a console error for ALL supported Linux flavors (KDE, gnome & gtk). I did run /every/ single one right before the release with the final code.

2) On my system ascii CMODS and binary ones, converted with cmodfix.exe work equally well.

3) I get no console.log warning as you indicated.

Sorry that's all I can say.

Bye Fridger

[t00fri]

You can email the corrected .ssc to the Celestia development mailing list (or PM Fridger or myself) along with a description of your modifications, data sources, etc.

Hope this helps.

OK, though I already emailed the corrected albedos for the nine planets to the developer list but nobody seemed to notice.

I think I'll PM the final corrected ssc to you or Fridger, it's more likely to be incorporated into a new version. And I'll add the moons too.

Malenfant,

I did notice your Albedo discussion (also in the developer list). Sorry, I was just too busy with preparing the release of FT1.2. I also think we should sync the values with the official ones . I would love to hear Grant's opinion beforehand, though.

Please refer to the official sources for all values if you send them. Implementing them into solarsys.ssc is a matter of minutes. A few-line PERL script does that job without touching any of your numbers by hand . Please double check the values...

Bye Fridger

[t00fri]

Hi,

I've got the same problem as Kolano : an excessive brightness of the galaxies when an add-on is present. In my case I only had the M1 add-on installed (Don't remember whose is it). The center of the galaxies is then so bright that it appears to be white. When I uninstall the add-on, everything is ok.

Thanks Fridger and Toti for the wonderful job you've done.
That shows that incredibly young people can manage to do valuable developpements.

Bye, Brokfn

To become more serious again:

It would be great if we could manage to discuss these matters on a "slightly" more solid basis than

M1 add-on installed (Don't remember whose is it).

What I did instead (which you also could easily have done) was to download the M1-add-on (Crab nebula) by Killeen from the Motherlode. So there is a reference.

On my Linux system at least, everything looks nice. M1 displays beautifully and all my galaxies are illuminated as they should be. NO changes after toggling the 'U' key (galaxy rendering on|off) .

Did you perhaps refer to the galaxy brightening after typing GoTO ('G')??? Because when you get really close to a galaxy, its illumination should /tremendously/ increase! In Celestia-FT1.2 it now increases stronger than in FT1.1 but the increase is still far too small, because Monitors just could not handle this. If we look at a typical galaxy from 10 parsecs distance in reality, its total luminosity is more than 11 orders of magnitude =100000000000 larger than what we can see from Earth!

Another possibility I might guess about would be for GALAXY add-ons! Since in this case, behind the add-on /our rendered galaxy/ is located in addition, that galaxy obviously appears much brighter, if the galaxy render flag is ON!

Please, if anyone could manage to produce a /precise/ description of the effect using a reproducable add-on etc, this could save us a LOT of time.

Thanks,
Fridger

PS:

Here is a quantitative illustration of the GALAXY add-on M81 brightening if galaxy rendering is ON (lower image), since our M81 is sitting slightly in front of the add-on.

Is it that sort of effect some people see? If yes, then please eliminate the respective galaxy from deepsky.dsc!



Bye Fridger

[t00fri]

As a small intermediate summary:

So far I can only reproduce 1 bug and that both in Windows and in Linux (reported concisely by Frank (fsgregs)):

If you load a nebula such as the Eagle nebula or Ring nebula as an extras add-on and view it from a distance, it will be visible only if it is reasonably centered in your view. If, however, you swing your view left or right using the left-mouse-drag or the arrow keys, the nebula disappears ... within I'd say ... 10 degrees left or right of center. It's as if you have moved the nebula behind an invisible cloud.

That bug was already there in FT1.1, but no-one spotted it

I suppose I know what the reason is...
It actually does NOT happen with GALAXY add-ons (like M81...). So far, we have simply not looked much into the rendering of nebulae. This will come in FT 2.0...


Bye Fridger

PS:

A little later: YES, indeed my suspicion was right. THAT bug is fixed now. It was just again related to the /inappropriate/ size of a numerical parameter for the case of NEBULAE. Since for galaxies that parameter was in the correct range, GALAXY add-ons would not exhibit that bug.

For people who are able to compile FT1.2 themselves, here is the simple FIX:

In render.cpp, just replace in lines 6567 and 6602 the command

Code: Select all

if (dsoRadius < 1000.0)


by

Code: Select all

if (dsoRadius < 10.0)


That's it!

[Malenfant]

I did notice your Albedo discussion (also in the developer list). Sorry, I was just too busy with preparing the release of FT1.2. I also think we should sync the values with the official ones . I would love to hear Grant's opinion beforehand, though.

Please refer to the official sources for all values if you send them. Implementing them into solarsys.ssc is a matter of minutes. A few-line PERL script does that job without touching any of your numbers by hand . Please double check the values...

Bye Fridger

Will do. I'm getting the numbers from the USNO (US Naval Observatory) Astronomical Almanac 2006 so they're about as up to date as they can get (and a papers published by James Hilton who is the guy who calculates all these values there). AFAIK this almanac is the official source for all magnitude calculations that you see in other almanacs and magazines, so it's a reliable source.

Expect this by the end of the week probably.

[t00fri]

Hey Boux,

just as a little "present" , I now built also the glut version of FT1.2, JUST for you...

Result: NO problems whatsoever. Straight compile and running without any error messages or warnings.

Cheers,
Fridger

[t00fri]

Looks good so far! thanks for the work done!

one thing I did notice, however, that when looking at a tiny DSC object with a radius of 170 ly from a distance (used to imitate a globular cluster), I notice that it seems to disappear gradually-- blob-by-blob.

other than that, it's all good!

PlutonianEmpire,

what you observe might well be happening. But you see, e.g. for the critical octree culling and other rendering issues the galaxy parameters are necessarily required to be within certain /generic/ ranges. So as soon as you exceed these galaxy-typical boundaries, many unexpected things could happen!

Cheers,
Fridger

[Toti]

As a small intermediate summary:

So far I can only reproduce 1 bug and that both in Windows and in Linux (reported concisely by Frank (fsgregs)):

If you load a nebula such as the Eagle nebula or Ring nebula as an extras add-on and view it from a distance, it will be visible only if it is reasonably centered in your view. If, however, you swing your view left or right using the left-mouse-drag or the arrow keys, the nebula disappears ... within I'd say ... 10 degrees left or right of center. It's as if you have moved the nebula behind an invisible cloud.

That bug was already there in FT1.1, but no-one spotted it

I suppose I know what the reason is...
It actually does NOT happen with GALAXY add-ons (like M81...). So far, we have simply not looked much into the rendering of nebulae. This will come in FT 2.0...


Bye Fridger

It was a clipping artifact caused by the huge scale difference between galaxies and nebulae.
I just fixed it by adjusting the near plane distance for small objects.

EDIT: Fridger, I just saw your post addendum with your (different) fix.

[t00fri]

As a small intermediate summary:

So far I can only reproduce 1 bug and that both in Windows and in Linux (reported concisely by Frank (fsgregs)):

If you load a nebula such as the Eagle nebula or Ring nebula as an extras add-on and view it from a distance, it will be visible only if it is reasonably centered in your view. If, however, you swing your view left or right using the left-mouse-drag or the arrow keys, the nebula disappears ... within I'd say ... 10 degrees left or right of center. It's as if you have moved the nebula behind an invisible cloud.

That bug was already there in FT1.1, but no-one spotted it

I suppose I know what the reason is...
It actually does NOT happen with GALAXY add-ons (like M81...). So far, we have simply not looked much into the rendering of nebulae. This will come in FT 2.0...


Bye Fridger

It was a clipping artifact caused by the huge scale difference between galaxies and nebulae.
I just fixed it by adjusting the near plane distance for small objects.

Hi Toti,

did you not see the rest of my respective message (that you quoted)? I added it slightly later, but certainly before you came into the forum just a moment ago...Are you replying offline?

Cheers,
Fridger

[Toti]

Hi Toti,

did you not see the rest of my respective message (that you quotet)? (I added it slightly later, but certainly before you came into the forum...

No I didn't refresh this thread's tab. I saw your message when I looked for new posts in the previous page (with my post just submitted). I quoted by copying/pasting before you added the news.

.Are you replying offline?

Yes, you know how expensive dialup is around here

[t00fri]

Hi Toti,

did you not see the rest of my respective message (that you quotet)? (I added it slightly later, but certainly before you came into the forum...

No I didn't refresh this thread's tab. I saw your message when I looked for new posts in the previous page (with my post just submitted). I quoted by copying/pasting before you added the news.

In any case, so far we (implicitly) considered distance scales of the order as relevant for galaxies. Now with FT2.0 ahead of us, we'll have to account also for the much smaller ones related to clusters and nebulae . This will need some "practicing" I am afraid...

Bye Fridger

[t00fri]

.Are you replying offline?

Yes, you know how expensive dialup is around here

Yes since my stay in Rio de Janeiro last week I do know
In my hotel one hour of internet connection essentially at modem speed was 20 R which is ~ 8 Euro or 10 $!

This is precisely what I pay here for 1 month of flatrate for a 6000 Kbit (!) DSL line...

[PlutonianEmpire]

one thing I did notice, however, that when looking at a tiny DSC object with a radius of 170 ly from a distance (used to imitate a globular cluster), I notice that it seems to disappear gradually-- blob-by-blob.

This is how the LOD works, but the transition should be pretty subtle. In fact I can't see what you describe. Could you please post the .dsc ?

It happens when I point at it, and then gradually move the view away.

Code: Select all

Galaxy "Pluto Globular"
{
        Type  "E0"
        RA           17.00000000
        Dec         110.00000000
        Distance      20000.000
        Radius         170.000
        AbsMag         -8.000
        Axis    [  0.000   0.000   0.000  ]
        Angle      0.0000
}

[Toti]

In my hotel one hour of internet connection essentially at modem speed was 20 R which is ~ 8 Euro or 10 $!

u$s 10 for a whole month of 6000 Kbit DSL line (flatrate) connection is a true bargain!

[Toti]

Plutonian Empire,

Sorry, I can't reproduce what you describe. There must be something wrong with your .dsc entries. Could you please post the entire file content? Do you have other .dsc files in your extras/ folder?

[suwalski]

Great work Toti, Fridger.

I have one issue, though I'm not certain where the problem lies.

In Windows, I get the Galaxies as expected.

Under Linux, same computer, I don't see the galaxies. This is with Galaxy Rendering on (U). If I use ')' to brighten all the way up, I get the purple galaxies I had before, with no Milky Way. It doesn't seem to be affected by Render Path.

I'm not discounting a problem in my GTK interface, but I was wondering if it's just me.