Something odd in starsdb2.0A & C

[granthutchison]

I downloaded both of these files (first C, then A) and there seems to be something very odd about the nearby stars in Celestia once either of these databases is installed. Local space fills up with a lot of luminosity-class III and V stars with unnaturally low absolute magnitudes - almost 500 of them within ten light-years of the Sun, crowding out all our real near neighbours.
I'm guessing there's some error in the data extraction process underlying this.

Grant

[Rigel]

I am not convinced it is an error of the extraction procedure.

Take a look at my web site to see how stars were extracted from the ASCC catalog and which parameters were taken in account.

You can also query yourself the Hipparcos and Tycho catalogs at :


http://cdsweb.u-strasbg.fr/viz-bin/VizieR?-source=I/239

and if you make a query with Par > 326 mas (stars within 10 ly of the Sun), you will get 7 stars from the Hipparcos catalog and more than 500 stars from the Tycho catalog. You will also see that the parallax error value in the Tycho catalog is often great.

Best regards

Pascal

[granthutchison]

I am not convinced it is an error of the extraction procedure.

Fair enough - wouldn't be the first time I've guessed wrong .
But what do you think has gone wrong? These 500 stars don't seem to be real stars - I think we would have noticed before now if there was, for instance, a B giant within 10 light years, even if it did have a (completely unphysical) magnitude of 12. Is there any sort of error-rejection you could build into your extraction process, to filter things like this out?
I certainly find such a cluttered local space so unrealistic that I've reverted back to the default Celestia stars.dat meantime.

Grant

[Rigel]

I will see if I can improve the extraction procedure for stars close to the sun.

Pascal

[granthutchison]

I will see if I can improve the extraction procedure for stars close to the sun.

I've been wondering if a look-up table could be used to compare the calculated absolute magnitude from Tycho with the theoretical absolute magnitude appropriate for the spectral and luminosity class given. If the calculated and theoretical absolute magnitudes are too far apart, then there is presumably something wrong with that entry, either in its parallax or its spectral/luminosity class - from my memory of eyeballing the list of the 500 nearest stars in Celestia using your database, I'd guess almost all the extraneous stars within 10ly would be trimmed out by that sort of cross-check.

Grant

[Rigel]

This is already done as you can see through the details of the extraction procedure explained here :

http://perso.wanadoo.fr/celestia.stars/newcat.html

I will take some time tomorrow to see if something is going wrong with nearby stars...

[granthutchison]

This is already done as you can see through the details of the extraction procedure explained here :

Oops. I'm sorry - I had read the extraction information, but I've obviously misunderstood it. I thought you were just using spectral/luminosity class and apparent magnitude to generate new distances, when Tycho didn't supply a parallax, rather than as a check on parallaxes already provided in the database.

Grant