identify for multi aperture data
Philip Massey wrote on Aug 08, 2008
Hi, Frank et al---
After a decade or so of not using multi-slits, I find myself at Magellan using a multislit mask. Let's say for example that there are 21 objects. I've used apall to define the 21 apertures and done the spectra extraction, and I've used appall with reference defined to extract the comps.
But, I'm stymied. If I run identify it is willing to do a single line in the multispec image (first, middle, etc), but what I need it do is perform a wavelength of each of these. Because they are done through a slitelet mask, the central wavelengths
vary drastically from one aperture to the next.
Is my best bet just to use scopy to turn these into one-d spectra? Or am I missing a simple way to do this?
thanks,
phil
Sad to think that 10-15 years ago I could have answered this for someone else....
After a decade or so of not using multi-slits, I find myself at Magellan using a multislit mask. Let's say for example that there are 21 objects. I've used apall to define the 21 apertures and done the spectra extraction, and I've used appall with reference defined to extract the comps.
But, I'm stymied. If I run identify it is willing to do a single line in the multispec image (first, middle, etc), but what I need it do is perform a wavelength of each of these. Because they are done through a slitelet mask, the central wavelengths
vary drastically from one aperture to the next.
Is my best bet just to use scopy to turn these into one-d spectra? Or am I missing a simple way to do this?
thanks,
phil
Sad to think that 10-15 years ago I could have answered this for someone else....
Francisco Valdes wrote on Aug 08, 2008
Hi Phil,
I also feel like it has been ages since I've dealt with this type of question. I had always hoped to add some special software and package for multislit.
Anyway, the way I would suggest is you keep the multispec format. When you do identify on one, the middle or first for instance, then you use the keys to move to the next spectrum (j and k as I recall). The lines and solution from the last one will still be present. If you recognize the shift you could do a a recenter (probably recenter all). At the end that shifted into the region you will be missing lines. You can add them with 'm' or probably better would be a new find lines. After getting the lines you would, of course, do a new fit. After you've done one comp tediously, any other comps will be a matter of reidentify (for the same mask).
I hope this slightly sketching description is helpful.
Regards,
Frank
I also feel like it has been ages since I've dealt with this type of question. I had always hoped to add some special software and package for multislit.
Anyway, the way I would suggest is you keep the multispec format. When you do identify on one, the middle or first for instance, then you use the keys to move to the next spectrum (j and k as I recall). The lines and solution from the last one will still be present. If you recognize the shift you could do a a recenter (probably recenter all). At the end that shifted into the region you will be missing lines. You can add them with 'm' or probably better would be a new find lines. After getting the lines you would, of course, do a new fit. After you've done one comp tediously, any other comps will be a matter of reidentify (for the same mask).
I hope this slightly sketching description is helpful.
Regards,
Frank
Last post on Aug 08, 2008