FWHM spectral resolution
Ivan Ramirez wrote on Sep 16, 2007
Hello Everybody,
I have ultra high resolution spectra (R~200k) from several nights and different runs which, most likely, have resolution variations in wavelength and time that I need to take into account in my analysis. I was wondering if there is a task in IRAF that can compute a sort of mean FWHM from the arc spectra. I now I can do this manually using implot, splot, etc. but I need to do this for hundreds of files so anything automated would be very helpful. Another question: I've seen people suggesting to measure FWHM in the original 2d spectra but wouldn't it be more realistic to measure the FWHM in the extracted 1d spectra? In principle you should get the same answer but if the instrument is not very well calibrated or the resolution you're trying to achieve is very high perhaps the FWHM of the extrated spectra would be larger (?).
Any help would be greatly appreciated,
-Ivan
I have ultra high resolution spectra (R~200k) from several nights and different runs which, most likely, have resolution variations in wavelength and time that I need to take into account in my analysis. I was wondering if there is a task in IRAF that can compute a sort of mean FWHM from the arc spectra. I now I can do this manually using implot, splot, etc. but I need to do this for hundreds of files so anything automated would be very helpful. Another question: I've seen people suggesting to measure FWHM in the original 2d spectra but wouldn't it be more realistic to measure the FWHM in the extracted 1d spectra? In principle you should get the same answer but if the instrument is not very well calibrated or the resolution you're trying to achieve is very high perhaps the FWHM of the extrated spectra would be larger (?).
Any help would be greatly appreciated,
-Ivan
Michael Weber wrote on Sep 16, 2007
You could autocorrelate the arc spectra (using fxcor). The FWHM of the correlation peak gives you the resolution (or 1/sqrt(2) of it?).
I remember a task on the kpno telecopes do do that on 2d images, which was very useful for focussing the spectrograph, but once you have the final 1d spectra, I agree with you that using 1d makes more sense.
I remember a task on the kpno telecopes do do that on 2d images, which was very useful for focussing the spectrograph, but once you have the final 1d spectra, I agree with you that using 1d makes more sense.
Ivan Ramirez wrote on Sep 16, 2007
Thanks for the help :D That is exactly what I've been looking for. And yes, the FWHM of the autocorrelation is sqrt(2) times the FWHM of the input data.
Last post on Sep 16, 2007