Problems with Identify for spectral wavelength calibration
sazajac wrote on Jan 30, 2011
Hello,
I have been having a couple of problems with the Identify routine in IRAF (I've written a lot to try to explain the problem I'm having, but my key questions are in bold):
-when I change the routine parameters using "epar identify" from the command line, when I try to change the fitting function and the order of that fitting function, the new parameters that I set are not the ones that Identify uses- for example, if the default parameters are function: spline, order: 3, if I change them in "epar" to function: legendre, order: 4, and then save the new parameters. the next time I run Identify, even though the parameter file says that it should be using a 4th order legendre function to fit the points, the fitting function and order are still listed as spline and order 3. Any thoughts on why the parameters are not updating? Any suggestion for how to get Identify to recognize what is it its own parameter file? I've noticed that if I try running Identify six or seven times, eventually the parameters will update to the values that I had specified, but it does not happen reliably or predictably, so I remain very confused.
-when I finally do get Identify into the desired function and order settings, I want to proceed with the calibration of my spectrum. The spectrum I am calibrating is a reduced image with defined skylines in a certain wavelength range. Within Identify, I use the Mark routine to specify the wavelengths of the peaks (place cursor in center of a peak, press "m," enter wavelength, press enter, repeat.) I usually mark 6-8 peaks in this way. I then want to fit a function to this points so that I have an mathematical correspondence between pixel number and wavelength, so I use the Fit routine: after my peaks have been identified, press "f" to fit a function to the points. After pressing "f," I next see graph displaying the residuals between my user-defined points, and the function-fit points. These residuals are typically 0.4-0.01 Angstroms. I usually them press "q" to quit the routine and then the interactive Identify program. However, I would really like to view a graph of the function and how it is fitting my points. Is there any way that I can do this?
Any feedback, ideas or suggestions would be very much appreciated!
-Stephanie
I have been having a couple of problems with the Identify routine in IRAF (I've written a lot to try to explain the problem I'm having, but my key questions are in bold):
-when I change the routine parameters using "epar identify" from the command line, when I try to change the fitting function and the order of that fitting function, the new parameters that I set are not the ones that Identify uses- for example, if the default parameters are function: spline, order: 3, if I change them in "epar" to function: legendre, order: 4, and then save the new parameters. the next time I run Identify, even though the parameter file says that it should be using a 4th order legendre function to fit the points, the fitting function and order are still listed as spline and order 3. Any thoughts on why the parameters are not updating? Any suggestion for how to get Identify to recognize what is it its own parameter file? I've noticed that if I try running Identify six or seven times, eventually the parameters will update to the values that I had specified, but it does not happen reliably or predictably, so I remain very confused.
-when I finally do get Identify into the desired function and order settings, I want to proceed with the calibration of my spectrum. The spectrum I am calibrating is a reduced image with defined skylines in a certain wavelength range. Within Identify, I use the Mark routine to specify the wavelengths of the peaks (place cursor in center of a peak, press "m," enter wavelength, press enter, repeat.) I usually mark 6-8 peaks in this way. I then want to fit a function to this points so that I have an mathematical correspondence between pixel number and wavelength, so I use the Fit routine: after my peaks have been identified, press "f" to fit a function to the points. After pressing "f," I next see graph displaying the residuals between my user-defined points, and the function-fit points. These residuals are typically 0.4-0.01 Angstroms. I usually them press "q" to quit the routine and then the interactive Identify program. However, I would really like to view a graph of the function and how it is fitting my points. Is there any way that I can do this?
Any feedback, ideas or suggestions would be very much appreciated!
-Stephanie
jkurk wrote on Jan 30, 2011
- concerning the first issue, I am unsure, but it might be that this happens if you apply identify on a spectrum already identified: IRAF might read the database it created for this spectrum and reuse the parameters it saved at that time. Try to apply identify on a new spectrum, or delete the database file (database/idfilename)
- the second issue is easier: after pressing f for fit, press h, j, k, and l to see different plots, including the one you desire.
I hope this helps,
Jaron
- the second issue is easier: after pressing f for fit, press h, j, k, and l to see different plots, including the one you desire.
I hope this helps,
Jaron
sazajac wrote on Jan 30, 2011
Thanks so much for your comments!
One question though: Does pressing h, j, k, or l change the nature of the data at all?
One question though: Does pressing h, j, k, or l change the nature of the data at all?
jkurk wrote on Jan 30, 2011
No, as far as I know, it does not change either the data or the fit (parameters); it just shows the fit (and data) in different ways.
Last post on Jan 30, 2011