Normalization in ISIS

[Keith Graham]

Hi All,

I see that the final image(s) produced by ISIS appear to be normalized. I see the Y scale centers around the number 1, and I also see in the log that there is a line indicating the normalized area. I have seen other spectra where the Y axis is in ADUs. So my questions are these:

1. Is the final spectrum actually normalized ( Since there is a Scaling Profile Spectral region in the Settings page, I am assuming it is).
2. If the spectrum is normalized, can you please tell me why this is preferable to outputting a spectrum that is not normalized? Are there advantages or disadvantages to normalize or not normalize?
3. I ask the number 2 question because there is a button on the profile page to Normalize the spectrum. If ISIS does normalize the final spectrum, why is there a Normalize button.
4. Is there a way in ISIS to choose whether to normalize or not normalize?

Cheers,

Keith Graham

[Christian Buil]

Keith,

ISIS use defined lambda range in settings tab for compute the normalization (scaling is a better expression).

The utility of Normalize button is the capacity to change the normalization zone (or scaling zone) of a given spectrum, after
processing. Remenber, the scaling operation is essentiel for compare spectra from different source. For this, original ADU value is
of poor interest and complicate comparison between spectra. The scaling solve the comparison problem by computing relartive intensities.

Some intermediate spectra returned by ISIS (raw or partially processed) are not normalized for giving possibilty to check original data and instrument efficiency. In the present ISIS
version, it is not possible to normalized status for final result... and I am not very favorable (it is horrific for me to see spectrum
profile with intensites like 456086 ADU ... a very large number, not easy to manipulate, and also... try to compare the profile with another profile with intensities of 5233 ADU at the same wavelelength....).

Christian

[Christian Buil]

You can download lastest ISIS release (V5.1.2). The version improve the processing speed for echelle spectrograph data.
http://www.astrosurf.com/buil/isis/new/release.html. Some bug also fixed.

La version 5.1.2 améliorement grandement la vitesse de traitement eShel (jusqu'à un facteur 3, ce qui accroit la "productivité" de
manière significative). C'est majeur pour les utilisateurs de spectrographes échelle.
Voir http://www.astrosurf.com/buil/isis/new/release.html. Quelques bugs ont été aussi dénichés et corrigés.

Christian B

[Robin Leadbeater]

Keith,

ISIS use defined lambda range in settings tab for compute the normalization (scaling is a better expression).

The utility of Normalize button is the capacity to change the normalization zone (or scaling zone) of a given spectrum, after
processing. Remenber, the scaling operation is essentiel for compare spectra from different source. For this, original ADU value is
of poor interest and complicate comparison between spectra. The scaling solve the comparison problem by computing relartive intensities.

Some intermediate spectra returned by ISIS (raw or partially processed) are not normalized for giving possibilty to check original data and instrument efficiency. In the present ISIS
version, it is not possible to normalized status for final result... and I am not very favorable (it is horrific for me to see spectrum
profile with intensites like 456086 ADU ... a very large number, not easy to manipulate, and also... try to compare the profile with another profile with intensities of 5233 ADU at the same wavelelength....).

Christian

Hello Christian,

I would prefer this to be a user choice. For example with slitless spectrographs we have the possibility of producing absolute flux calibrated spectra. I am currently trying to reduce runs of several hundred spectra taken at 25 sec intervals using a Star Analyser looking for very fast changes for the T -Tauri campaign. As well looking for changes in the spectrum, I need to know how the flux is changing relative to the comparison star in the same field. Software like ISIS which allows batch processing is the only possibility to do this.

Cheers
Robin

[Christian Buil]

You can now use the very new version 5.1.3. I have added a small tip for produce unnormalized spectra by an easy way. See desciption here:

http://www.astrosurf.com/buil/isis/new/release.html

Read also my comment about normalization purpose...

Christian B