fxcor PANIC

flame · Joined: 22 Mar 2006 Posts: 3 Location: Padua

I work with iraf 2.12.2a-export on Mandriva 2006.
When I run the task fxcor and I use a square Fourier filter on data the following errors appear on fxcor display:

http://www.geocities.com/bohzzz/fxcor_f.jpg
http://www.geocities.com/bohzzz/fxcor_p.jpg

In particular the first error happens when I enter in spectrum mode ('s') and then in fourier mode ('f'). As you can see the display is uncomplete! There isn't the fourier spectrum of the object!
The second error happens when I try to enter in spectrum mode ('s') and then I try to display the prepared spectra prior to correlation ('p')

Both errors are irreversible and cause the crash of the cl session.

I tried to change the spectrum that I use, to change the cut on and cut off on filtpars but the result is the same. I have this problem only with this type of filter.

Help me please...

fitz · Posted: Wed Mar 22, 2006 6:18 pm Post subject:

Please post a link to the object/reference spectrum (FITS file) as well as your filtpars parameters and any fxcor parameters. I can't reproduce the error using an artificial spectrum. You might also try setting the 'pixcorr' parameter to see if that makes any difference, it could be that something in the header is causing the problem and turning this parameter on will ignore the wavelength solution and correlate just the pixels.

-Mike

flame · Joined: 22 Mar 2006 Posts: 3 Location: Padua

Hi!

Here is the link with two example of different spectrum (giraffe spectrum and uves spectrum) and the fxcor, filpars and continpars parameters that I used.
I tried to change "pixcorr" parameter that initially is set to "no" but the results are the same, panic.. panic panic.... Sad

Filtpars parameters are relative to giraffe data, while for uves data cut-off is set to 70.

ftp://ftp.pd.astro.it/pub/paladini

I tried to reduce the same data with Noao Digital Unix/IRAF revision 2.11.3, Export version of DUNX/IRAF V.2.11 on DIGITAL UNIX V4.0 and it works well, but I cannot always use this computer.

Thanks a lot,
Claudia

fitz · Posted: Fri Mar 24, 2006 4:16 pm Post subject:

Thanks for the data and params, I'm able to easily reproduce the problem. I'll get back to you with a fix as soon as I can.

-Mike