ISIS ou VSPEC calibration inverse Angstrom vers pixel

[Hubert Boussier]

Bonjour,

Dans ISIS à l'issue de l'étape de calibration, le logiciel génère une loi de calibration qui permet de déterminer la longueur d'onde correspondant à une ordonnée X en pixel.
Existe-t-il dans ISIS ou Vspec un moyen simple (hors la résolution de l'équation de calibration du troisième ou quatrième degré) permettant de connaître la valeur de X lorsqu'on connait la longueur d'onde.

Utilité : Sur un spectre dont le S/B est faible, lorsqu'une raie semble douteuse, aller voir sur les images du spectre s'il n'y a pas à l'ordonnée X un point chaud, un cosmique ou autre chose afin de déterminer si la raie que l'on voit est vraie ou bien si c'est un artefact.

Merci d'avance de votre aide

[Paolo Berardi]

Hello Hubert, the need to check in the low SNR spectra for artifacts is the same I have. It would be very useful to have a "dynamic" relationship between x-coordinate on the 2D image and wavelength or vice versa. For example I'm thinking to a lambda value related to the x coordinate of mouse cursor when it hovers over 2d spectrum image (i.e. in the first tab of ISIS that must have a polynomial loaded).

When I need, I use a spreadsheet filled with polynomial formula and parameters read in ISIS (dispersion window). I put x-coordinate (where I suspect an artifact) and read the corresponding wavelength. But it is not very comfortable...

Ciao
Paolo

[Francois Teyssier]

Bonjour Hubert,

As-tu essayé dans ISIS de décocher la case 'Spectral Calibration' dans l'Onglet 2. General ?
Cela te donne un spectre non calibré en longueur d'onde. Tu dois ainsi retrouver l'abcisse de la raie douteuse pour aller vérifier sur l'image.

François

[Paolo Berardi]

Hi Francois, I think that it's perfect for Hubert. You just have to play with profile thresholds and zoom to identify the spectral feature because the low-res profile shape is different from that corrected by response (it should depend on flat lamp type). I prefer to scan the 2D spectrum (I do it before I saw the profile) searching for any visible artifacts, then check if they have impact on the profile.

Ciao
Paolo

[Hubert Boussier]

Hi Paolo and François,

Paolo summarized very well what I was looking for i.e. "have a "dynamic" relationship between x-coordinate on the 2D image and wavelength or vice versa". I am afraid that this tool is not implemented in ISIS or VSPEC.

François proposal is usable but it do not answer to my wishes because doing this you need to reprocess again all the set of raw images without spectral calibration in order to pick the abscissa of the anomaly ; but sometime I have not all the files that are necessary to reprocess the set of raw images immediately available.

So, I developed two small tools to make the link lambda toward x abscissa. I give this tools here, may be it could be useful for somebody :

1/Visually with the picture of the Ar-Ne spectrum on which quite all the spectrum lines are identified, I compare the calibration Ar-Ne image to this picture and try to see on which x coordinate zone is located the lambda I am looking for.

http://www.astrosurf.com/astrild/Spectr ... _Ar-Ne.jpg


2/ I use the "solve" function of EXCEL, I did a spreadsheet with a small table in which I copy the polynomial calibration coefficients and then the solve function will directly calculate what is the x value corresponding to a given lambda.
http://www.astrosurf.com/astrild/Lambda ... cisse.xlsx


All the best