Processing a lot versus a few spectra

[Olivier GARDE]

Is it possible to have 3-4 raw spectra to see what it looks like because by only seeing the curves, it is difficult to understand what is wrong.

And cosmic ray correction is only to apply if you have cosmic ray present in your raw spectra.

[David Cejudo]

Here are four raw images.
https://www.dropbox.com/sh/8zsdy7i5v4k9 ... PSwba?dl=0

The ones ending in 20 and 22 give good results.
The ones ending in 33 and 38 are the problematic ones.

I could not attach them here so i am including a link to my Dropbox where the images are for anybody to have a look.

Thank you very much for your interest. I am struggling to learn!

[Robin Leadbeater]

Optimal binning is checked.

Hi David,

My recommendation was to run with it unchecked. Because it works by deciding which rows to use, it can become confused, particularly in spectra which are not oversampled. I recommend unchecking optimal binning for normal use. I find it is mainly useful for faint spectra against a bright sky background where it allows only the optimum rows to be included to maximise SNR. I use it regularly for faint supernovae for example but I always compare with and without to see if it is really improving the SNR in the spectra and to make sure it is not doing anything silly like this. (Fortunately it is really quick and easy to rerun in ISIS, even some time later by saving the XML file)

Cheers
Robin

[Robin Leadbeater]

If you do use optimal binning, you need to set the rejection coefficient correctly to avoid this problem. (Note there are two settings for rejecting cosmics, the rejection coefficient in the general tab which works with the optimum binning and the filter value in the settings tab which works with the cosmic filter)

See item 5 here
http://www.astrosurf.com/buil/isis/guid ... to1_en.htm

Cheers
Robin

[David Cejudo]

Hello.

I unchecked the optimal binning and now it works perfectly even processing all the 170 spectra.
Checking optimal binning and setting a rejection office above 30 also works, but unchecking it results look better!


Thanks Robin.