Two Million Stars Database?

[BrainDead]

Could someone please tell me why the one or two million stars databases
are NOT compatible with versions of Celestia after 1.3.2?

Just curious to know if there is some way to convert these files to whatever
is needed for use in 1.4 Pre whatever.

Where might this information be found? Sure is a lot of EMPTY space in the
universe right now.

Thanks

[selden]

The author of those databases simply hasn't had the time to convert them. He's aware of the problem.

[BrainDead]

Okay,

Thanks for the information, Selden. I understand that that is a problem.

What I was trying to get to though, was what it might take to convert
the data myself. I looked through a couple dozen web pages trying to find
a file definition, or some such data, but could find nothing that makes any
sense to me.

Was just curious.

[selden]

The file format of the current 2 million db is the same as celestia's original stars.dat, which has been described somewhere here on the Forum. There's a program on SourceForge that converts the new stars.txt format into the new binary stars.dat format. So what's needed is a program to convert the old binary format into the new text format.

Actually, I suspect conversion directly from old binary to new binary formats would be easier, avoiding the binary to ascii translations. My understanding is that the primary difference between the two is the representation of the stars' locations, which is RA, Dec and parallax in the old, and Celestia's internal XYZ in the new.

[ajtribick]

Writing a utility to do this should be fairly simple.

The old stars.dat format is:

Code: Select all

  4 byte int   : number of records

  4 byte int   : HIP number
  4 byte int   : HD number (n.b. this is missing from the README.txt file)
  4 byte float : right ascension
  4 byte float : declination
  4 byte float : parallax
  2 byte int   : apparent magnitude
  2 byte int   : stellar class
  1 byte       : parallax error


The new stars.dat format is:

Code: Select all

  4 byte int   : number of records

  4 byte int   : HIP number
  4 byte float : x-coordinate
  4 byte float : y-coordinate
  4 byte float : z-coordinate
  2 byte int   : apparent magnitude
  2 byte int   : stellar class


The x,y and z coordinates are stored in light years. Celestia uses a conversion factor of 1 parsec = 3.26167 light years.

I'm not sure what the coordinate system convention used is though.

[BrainDead]

Writing a utility to do this should be fairly simple.

Everything is simple when you understand what you're doing.
It's the understanding part that's difficult...

I'm not sure what the coordinate system convention used is though.

That's okay... You've given me a far better understanding of the file layout
than I had before.

Many thanks for the explanation.

[granthutchison]

Folks, I converted it.
I've just got the 2 million stars.dat humming along nicely under 1.4.0pre6.

I noticed this thread when I dropped back to see if there was any feedback on the new numberedmoons.ssc, and it so happens I wrote myself a little utility last week that converts old-style stars.dat into new-style stars.txt, which I can then compile into new-style stars.dat using Chris's makestardb.exe.
So: I just ran it experimentally on Pascal's big stars.dat, and (after a lot of chugging and groaning to spit out a 100MB stars.txt file), it worked. Ta-da!

I guess now I just need to track down Pascal to make sure he's happy to have this released ...

Grant

[jestr]

Nice one Grant,cant wait to try it,Jestr

[BrainDead]

Folks, I converted it.
I've just got the 2 million stars.dat humming along nicely under 1.4.0pre6.

I guess now I just need to track down Pascal to make sure he's happy to have this released ...

You ARE the man, Grant. I can't imagine why Pascal would be unhappy, but
that's entirely your perogative. I'll simply wait here quietly until you're
ready to lay it on us.

Boy... THAT was easy...

See what happens when you poke your nose in places where it doesn't
belong? <smirk>

Seriously though. THANKS to you Sir.

[granthutchison]

I can't imagine why Pascal would be unhappy, but
that's entirely your perogative.

Well, it's kind of Pascal's prerogative, too. He did a hell of a lot of work on these data, and then I came along and did (literally) two minutes of tweaking my original program and about ten minutes of watching the screen and sipping coffee, and it was done. So I figure he gets the final say.

But anyway, you're right: I e-mailed him and he's happy. So I'm just sending him the processed files to look over, and then we'll be sorted.

Grant

[t00fri]

Hi Grant,

I forgot what the original catalog was for the 2m stars (TYCHO-2 [2000] 2.6m ?) and exactly what were the STRONG assumptions made for assigning distances to many of the 2m stars. In any case this looks like a very easy homework for me to write a little PERL script and extract precisely the wanted star data in the appropriate stars.txt format from the original catalogs in Strasbourg. At the same time we then get a documentation in human readable form in that script about the assumptions that led to the modelled distance assignments!

Then we are able to always start from a well-defined text star base that may be converted any time with Chris' tool. It might well be that one day we want to play a bit with those assumptions leading to the distances... Then it's only a few minute affair with a PERL script.

Bye Fridger

[selden]

Fridger,

You might want to take a look at Pascal's Web site at http://pascal.hartman.free.fr/newcat.html, which discusses (mostly in layman' terms, though) some of the statistical procedures used to statistically determine the stars' photometric distances. He's not a professional astronomer but did get some help from one. I suspect it's slightly more work than you're anticipating, but certainly not beyond your capabilities

See

[t00fri]

Fridger,

You might want to take a look at Pascal's Web site at http://pascal.hartman.free.fr/newcat.html, which discusses (mostly in layman' terms, though) some of the statistical procedures used to statistically determine the stars' photometric distances. He's not a professional astronomer but did get some help from one. I suspect it's slightly more work than you're anticipating, but certainly not beyond your capabilities

See

Selden,

thanks for the pointer. Since I have done lots of VERY similar considerations and written corresponding PERL scripts I don't think the task would be very much bigger than what I did before with my binary orbits or partly also the galaxy extractions.

The idea is always similar: One tries to estimate absMag using various color-luminosity relations . Once absMag is known and since appMag is measured the distance may be estimated. This is also used many times in my 700 line Perl program for galaxy extraction.

What I wanted to emphasize above is that additional transparency in the modelling used would be very important. We both remember the extensive discussions about VERY strong biases in Pascal's 2m dataset, I suppose.

Yes this will take some additional time which right now certainly is not in sight But I really don't think this task is in any way particular.

Bye Fridger

[alphap1us]

Hi,
Grant sent me both sizes of Pascal Hartmann's Star DB for Celestia 1.4.0 You can find them here:

http://celestiamotherlode.net/catalog/s ... 1126545717

Cheers,
Joe

[BrainDead]

Grant sent me both sizes of Pascal Hartmann's Star DB for Celestia 1.4.0

Thanks very much Joe. I've already downloaded the updated version.
If I haven't said it enough, thanks also for providing your great service via
the MotherLode.

Much appreciated.

[t00fri]

The All-Sky Compiled Catalogue (ASCC-2.5) of 2.5m stars, on which Pascal Hartman's analysis and Grant's recent conversion was based has had a substantial error revision. My question would whether Grant/Pascal use already the new corrected data that appeared in 2003:

Bye Fridger

Code: Select all

History:
  * 21-Aug-2002 (date of the first release)

  * 18-Jun-2003: Version 'A'
    The 1st version of the ASCC-2.5 (21-Aug-2002) contains some errors:

    -  3982 false stars were detected at Dec = 0 ... +1 degrees. In
       the 2nd version of the ASCC-2.5 these stars were excluded from
       the file ccp00.dat. The corrected file ccp00.dat supersedes
       the previous file ccp00.dat and contains 88230 stars.

    -  for 3948 stars at Dec = 0 ... -1 degrees the PPM data were
       missed during the catalogue compilation. In the 2nd version
       the missed data were added and corresponding stellar data were
       recompiled. The corrected file ccm00.dat supersedes the previous
       file ccm00.dat and contains the same number of stars.

    -  34 stars with PPM-only data were missed in zone Dec = 0 ... -1.
       They were included in a new file of additional catalogue
       objects ccadd.dat.

    Besides, 3284 components of multiple systems from over the sky
    were included in the catalogue. These stars and their compiled
    stellar data are placed in the file ccadd.dat.

    The false and corrected stars are listed in files fals.dat
    and corr.dat, respectively.


[Dollan]

Is there any "intermediate" star add-on suitable for 1.4.0 pre6?

I downloaded this data base, but at the present time it is simply too large for my system to handle (I'm working on that, though).

...John...

[BrainDead]

Is there any "intermediate" star add-on suitable for 1.4.0 pre6?

John,

Are you saying that you can't use the ONE million star database either?

My system, a very old Gateway, 256MB RAM, no graphics card, and a slow
Intel Celeron 1400 MHz CPU chugs for a little while when it starts Celestia,
but the full 2 million stars database causes no problems on this machine.

Just FYI.

[t00fri]

Is there any "intermediate" star add-on suitable for 1.4.0 pre6?

John,

Are you saying that you can't use the ONE million star database either?

My system, a very old Gateway, 256MB RAM, no graphics card, and a slow
Intel Celeron 1400 MHz CPU chugs for a little while when it starts Celestia,
but the full 2 million stars database causes no problems on this machine.

Just FYI.

Well I do have a fast machine with 3GB of RAM. Still it was comforting to see that with the 2m stars and our new Celestia-FT1 version, there was NO visible slowdown as compared to the default stars.

Yet e.g. Andromeda (M 31) looks much more beautiful with hundreds of faint TYCHO stars embedded into the galaxy...

Bye Fridger

[Dollan]

Yeah, my system bogs down big time. The primary problem, though, is I'm running a temporary machine with XP on it, and the cmputer really wasn't designed for the OS. In a few days, though, I'll be getting a new hard drive for a different machine that *can* handle XP. Once that happens, I'll transfer the vid card over from my temp machine, and we'll try it out then.

...John...