Re: [guide-user] ppmxl in Guide 8/9 update - post 3

metpaper@Safe-mail.net Mar 4, 2012

I thought I could get away with not using the declimit command with the full dataset, as a test on a few degrees of declination when I first set it up worked at full speed, but it turns out that the tdf as is might cause noticeable plotting slowdown without use of declimit when using tens of degrees of declinatin, even for fields well away from those that there are data for.

Basically I was trying to get around having to key in tons of declimit commands like declimit 5.00 to 5.25, declimit 5.25 to 5.50, for each data file, but fortunately it seems sticking to ten degree ranges works well enough. That is, for each tdf segment (except the very last, which is for the vizier downloaded data which will end up randomly distributed) in ppmxl.tdf just after the command

mag 145 5

and just before the command

go_to_top

enter

declimit 50 60

such that it reads

mag 145 5
declimit 50 60
go_to_top

for all the tdfs for the files n50a.dat to n59d.dat inclusive (the first one, n50a.dat, is the baseline tdf and will not have the go_to_top command, just the mag 145 5 command).

then the same for n60a.dat to n69d.dat, except now it will be

mag 145 5
declimit 60 70
go_to_top

then 70 to 80 for the n70a.dat to n79d.dat segments of the tdf, and finally 80 to 90 for n80a.dat to n89d.dat. The last TDF segment remember is for vizier download data, so doesn't need this, the one for file ppmxl.dat.

Similarly if you extend the tdf to all sky using these declimit ten degree strips with relevant settings it should work fine. Notice it is always "smallest" number first, so if you're setting up Southern declinations the range has to have the most negative number first, eg for declination South 30 to 40, ie -30 to -40, you actually write "declimit -40 to -30" else it won't work (if I remember rightly).

These ten degree strip amendments, which are applied with the same value for bunches of 40 files, seems to work fine on testing with no noticeable slowdown either in ppmxl data range or beyond it when fields with ppmxl.tdf either toggled on or toggled off.

Cheers

John

-------- Original Message --------

> NOW, the main reason for this update though is the above file is also for running PPMXL from hard disk. The data files for PPMXL can be downloaded here :-
>
> ftp://cdsarc.u-strasbg.fr/pub/cats/I/317/dat/
>
> the tdf is expecting them to live in guide\ppmxl\ subdirectory, the tdf of course goes into the Guide directory, and contains links to the ppmxl data as well as the vizier download...
>
> But not fully. For various reasons I use UCAC2, not 3, and CMC14. These end at about 50 degrees North declination. Previously I used UCAC3, once it became available, North of there. UCAC3 takes a lot of cleaning and special care, such as Bill has taken with the Guide 9 DVDRom version.
>
> I've got fed up with it. PPMXL is also "better" than USNO B1.0 as it has USNO B1.0 object linkage via 2MASS (ties it down better, gives 2MASS J and K magnitudes as bonus info) where stars are bright enough, and makes some attempt to merge the false multiple entries into one object at the faint end.
>
> So, the tdf in the above zip file only works for +50 degrees to the North Celestial Pole. This was a download of only a few Gigs, and takes up 13.4 Gb of disk space. UCAC2 and CMC14 look after most of what else I need, with PPMXL from vizier should I need to go fainter (I also have USNO A2.0 on hard disk).
>
> Anyone preferring full PPMXL to full USNO B1.0 on their hard drive can use the above. If the ppmxl files are placed elsewhere rather than in the guide directory, the lines beginning with the word "file" need to point to them, have the pathname to them.
>
> And the above tdf can be extended all sky by just adding the entries for each additional PPMXL dat file. These can be entered at the end, but the end tdf segment is for the vizier download of ppmxl, so you may want to include them just before that one in order for ease of finding things, though if one segment works, they all should, as long as you include all the file names and just once for each.
>
> Full download is ostensibly 39 Gb utilising 89 Gb of hard disk space. Given some drives nowadays, and their speeds, this isn't too evilly bad. Guide can cope with big datasets very well anyway, and has done so for a long time, as it takes note of where you are looking and only looks through data relevant to that point. Having said that I have not bothered using the declimit command here as things are fast enough without it.
>
> I've used automag filtering you see.
>
> Right click and More Info and outlinking are just as with the vizier version.
>
> Cheers