iraf.net

Dear allsorry for bothering you again^^
I have a synthetic spectra and I wanna compare it with observed spectra from echelle data.
I already transform observed spectra into text/ascii file using wspectext, and select a certain range of wavelength.
how I justified that the synthetic spectra is match enough with observed spectra?
what task I can use to do that in IRAF?thanks
sorry for that 'newbie' questionregards
PJ

This is difficult to answer without knowing what you mean exactly by "compare the spectra". Tasks like SARITH can be used to difference the spectra (i.e. subtract one from the other) and the result might give you an idea how well they match, but won't give you an answer like "these are 93% identical". Similarly, if you want to know whether e.g. the line strengths or widths are roughly the same you'll need to use tasks to analyse the individual lines and not simply a list of the pixels/wavelength. A task like SLIST can print information about the dispersion, other tasks like SCOPY can be used to rebin/extract data to a common wavelength range and sampling.If you can be more specific about what you're after I might have other suggestions.-Mike

Dear Mike,1. I would like to see how good a particular line in a star spectrum eg HeII fit with that of a synthetic spectra. Visually I can overplot using splot but I would like to know if Iraf has a task to quantify the fit.2. Using the hst_calib.ctools.tomultispec task on a bintable gives me the OIF format, that is .imh and .pix. Even though it woks, I would like to convert it to the .fits format which I am more familiar with. What task can I use?Thank you.Yours sincerely,
Siti

Hello,Concerning the format all you need to do is specify .fits as an extension when you create the file with tomultispec to rspectext. If you want to work primarily with fits format then in your login.cl file (or interactively for a one-time change) you do[code:1:bd9161a73f]
set imtype ="fits"
[/code:1:bd9161a73f]As to the goodness of fit this is a statistical question. But one way is to do a subtraction, I would suggest the task "skytweak" to minimize the rms of the region with the line by adjusting the wavelength and flux scales, and then measure the rms in the region. To convert this to a goodness of fit, chi squared, or a likelihood measure you would need to characterize the noise in your spectrum, probably from the rms is a clean continuum region. It it the comparison of the rms in the data (assuming there is no noise in the synthetic spectrum) against the rms after the subtraction that tells the goodness of the fit.Frank Valdes

Dear Frank,Thanks for your suggestion. I am experimenting with "skytweak" now.I have a question that is not related to Iraf but to my user's account. I did't receive notification in my email when someone sends me a private message, but only when someone replies to my post in the forum. Could you help me with this?Thanks.Your sincerely,Siti

Siti,I checked your account and all the notifications (e.g. new private msg, reply, etc) are enabled.-Mike

Dear Mike,Thanks for your attention.Yours sincerely,Siti

Hi Frank,I've tried the task which you suggest to me, here what I was doing:
1. I corrected my observed spectra from doppler shift using DOPCOR (i have a question about this in separate thread)
2. using SKYTWEAK, I did minimization of rms by adjusting flux and wavelength
3. i measured the rms of this 'tweaked' spectra from clean continuum, consider this as rms1
4. I subtract my 'tweaked' spectra with synthetic spectra
5. I measured the rms of subtraction result, consider this as rms2
6. then rms1/rms2 can be considered as goodness of fit (?)any comment please? did I do right?
thanksPJ

Dear PJ,I thought when we apply "skytweak" and adjust the wavelength and flux scales of the synthetic spectra, the outcome is automatically the subtracted spectra with the minimum rms. Maybe the experts can help clarify? Thank you.Yours sincerely,
Siti