Adding MultiSpectrum to SpectrumDM (v1.3??)

[Petr Skoda]

Hi Gregory and Markus ,

When you start work on multispectrum DM, please consider also the echelle 
spectrum .

In fact what you propose is exactly how the echelle spectra are currently 
presented - one line is the wavelength and then multiple lines are just 
giving flux, flux variance, some flags of good/bad pixels and in case of 
echelle also blazed flux, unblazed and normalized

see may slides 
https://wiki.ivoa.net/internal/IVOA/InterOpJun2026Apps/spectra-revisited-final.pdf

slide 48  - simple example

some spectrographs (e.g. ESPRESSO) have even 2 wavelengths (air and vacuum 
) in primary header or first row

but the principal is same.

I you case you will have multiple spectra in the given piece of sky - but 
of different object - in echelle case all those spectra in multispectra 
are bound together - different orders
but they have separate wavelength range for each order, you have in fact 
all spectra in the same range but as I understand you want to pack to one 
spectrum the wavelength anyway - so it doesnt matter




And if you look at my slide 52 the prepared EPRV format (being prepared 
under the lead of JPL and Univ of Arizona) has the same - but the 
wavelength is one of such Image rows/extensions as well as variance, flux, 
blaze etc...

So in principle we could us the multispec DM for echelle easily.

Keep me in loop when you start the work on this !

Best regards,

Petr Skoda


On Wed, 17 Jun 2026, Markus Demleitner via dm wrote:

> Date: Wed, 17 Jun 2026 14:31:13
> From: Markus Demleitner via dm <dm at ivoa.net>
> Reply-To: Markus Demleitner <msdemlei at ari.uni-heidelberg.de>
> To: "dm at ivoa.net" <dm at ivoa.net>
> Subject: Re: Adding MultiSpectrum to SpectrumDM (v1.3??)
> 
> Dear Grégory,
>
> On Sat, Jun 13, 2026 at 01:51:24AM +0000, Dubois-Felsmann, Gregory P. wrote:
>> What is the next reasonable step to take?  Submit a PR to the
>> standard repo?
>
> Well, that would be the ideal way.  But really, you could also just
> file an issue with ±standard-suitable text and leave it to the editor
> to place it if you are unsure.
>
>> I'm happy to work on it but I feel a little out of my depth in
>> being sure that I've understood the full gestalt of the existing
>> document.  Would we need to figure out how to support MultiSpectrum
>> in the FITS serialization as well, in order to proceed?
>
> If you don't want to figure out the FITS part, I think the WG can
> help out.
>
> IIRC you also use SpectrumDM utypes to annotate spectral/flux columns
> coming out of in-database spectra.  I think that is a very good
> thing, too, and the SpectrumDM update would be a good opportunity to
> add a (non-normative, for now) appendix saying something about this
> kind of usage.  I'd volunteer to write this if other people also
> think that's a worthwhile thing.


I fully support the direct "rendering' of spectra from table - in fact 
this is needed for Parquet usage  anyway .


>
> Thanks,
>
>           Markus
>
>

*************************************************************************
*  Petr Skoda                         Phone : +420-323-649201, ext. 361 *
*  Stellar Department                         +420-323-620361           *
*  Astronomical Institute CAS         Fax   : +420-323-620250           *
*  251 65 Ondrejov                    e-mail: skoda at asu.cas.cz          *
*  Czech Republic                             skoda at sunstel.asu.cas.cz  *
*************************************************************************