Spectral 2.0 - Document update

[Petr Skoda]

Thanks Igor for help

On Wed, 10 Jun 2015, Igor Chilingarian wrote:

> Hi Mark,
>
> This type of normalization is often used in stellar astronomy in order to
> measure equivalent width of spectra lines. And as far as I know, even now
> most astronomers working in the field prefer to define it manually, i.e.
> subjectively.

To be precise the equivalent widths was the first motivation.
But currently the most interesting reasons for using continuum 
normalization is the need to investigate (even tiny) changes in spectral 
line profiles on high res spectra - e.g. asteroseismology (non-radial 
pulsations, stellar flares , stellar spots ...)

and the most fancy need is the machine learning -
if you want put spectrum into the machine learning procedures you must 
normalize them to have most points at the same scale (same amplitude)
- basically most stars have spectrum between 0.0 and 1.0 - emission stars 
may have peaks higher than 1 (even 100!) - this makes them easily 
identifiable by machine learning codes ....

Some spectra processing - e.g. Fourier disentangling of binary stars, SVD 
methods,
removal of tellurics by HITRAN model of our atmosphere or microturbulence 
analysis using Fourier transform are extremely sensitive to the smoothness 
of the continuum (and proper placement).
I do not want to open physical discussion - late type stars, novae, 
symbiotic variables etc ... do not allow to see the continuum et all ...
All is just "virtual" fit here, needs theoretical models ...



*************************************************************************
*  Petr Skoda                         Tel   : (323) 649201, l. 361      *
*  Stelarni oddeleni                          (323) 620361              *
*  Astronomicky ustav AVCR, v.v.i     Fax   : (323) 620250              *
*  251 65 Ondrejov                    e-mail: skoda at sunstel.asu.cas.cz  *
*  Ceska republika                                                      *
*************************************************************************