About Miles data base

[Martin Dubs]

Hello Francois,

disregard my comment about the non dereddened specrtra and my mail. I got the spectra numbers somehow mixed-up, so I thought that the dereddened and non dereddened spectra for HD20041 looked very similar, but everything is in order as you showed above and I checked myself. Sorry about my mistake.
So everything is fine in this case. I agree with you to use only stars with E(B-V) = 0 for determining instrument response.

Martin

[Marco Leonardi]

Hello,

following the very important last posts by Christian and François related with the MILES catalogue de-redded spectra inconsistency, and thanks François to share the link where to download the original MILES spectras (before de-reddening), as promised today we have updated the excel file posted yesterday by me and Paolo Berardi.

SPEC_VIEW_V2.zip

(789.68 KiB) Downloaded 672 times

In the zipped enclosed file you will find:


1)
SPEC_VIEW (folder) with:

- file excel "MILES_SPEC.xls"
- folder EXPORT (empty)
- folder MILES_OBS (empty)


2)
Please download the 985 spectra ASCII (ZIPPED) at the following link:

http://quasar.teoth.it/html/varie/MILES_OBS.zip

unzip the single files s0001-s0985 (file extension .test.dat) in the previous empty folder MILES_OBS .

3)
Open the file "MILES_SPEC_V2" (there are two versions Excel 97-2003 and Excel 2010). You will see the table with the 985 stars list. You can select a star and click on "profile" button.

4)
If you click [ON] over the label "Synthetic de-reddening" a new spectra will be superimposed. This spectra has been de-reddned on the base of the color excess E(B-V) indicated on the cell J3. This is the new screen view:

Profile_test.gif (34.47 KiB) Viewed 13489 times

5)

The button[Esporta VSpec] will copy the fie MILES in the folder EXPORT (see point 1). Moreover a new file will be created for the inter-stellar extinction even if the "Synthetic de-reddening" is on OFF, this new file will have the extension "_dered.dat".

Finally the changes vs the previous excel file posted yesterday are the following:

- the name of the folder that contains the spectras
- the spectras
- the function ON/OFF that now will do a de-reddening while in the version before a reddening was made.

The MILES OBS spectra can be used for the calibration in intensity, selecting the specific star spectra in the catalogue. While the de-reddened spectra is just for didactic purpose in order to understand what can do the inter-stellar extinction.

Again thanks to all experts here on supporting us to develop this simple tool. If there is any problem please just ask.

Kind Regards,

Paolo and Marco

[Christian Buil]

I have a good new. By using ogininal MILES data (observed spectra - and thanks Palo and Marco for the quick and useful reaction!), the consistency of real observation reduction is significantly improved. I have intensively tested V5 ISIS database (natural reddened MILES spectra) on many "standard" reference stars (I have intensively observed these objects with LISA last weeks) and now results are really valuable for compute decent instrumental response.

I have extended subset MILES library accessible from ISIS to A, B and... F stars (for instrumental response evaluation). More stars availables on the sky and F stars are less reddened (nearest distance from you).

Christian

[Francois Teyssier]

A propos du cas de HD 28978.

Il y a manifestement une erreur dans la base Miles, qui n'est pas difficile à identifier.
Pour HD 28978, Miles donne un indice d'excés de couleur E(B-V) = 0.117
(le tableur de Marco et Paolo facilite la recherche de la valeur)

Or dans Simbad, on trouve
B = 5.738
V = 5.681
D'où B-V = 0.057

On vérifie que les valeurs publiées correspondent à cette valeur. Toujours dans Simbad, on trouve :
| V B-V U-B |Q/nbm| remarks | reference |
------------------------------------------------------------------
| 5.68 +0.05 +0.12 |4 | |1966CoLPL...4...99J|
| 5.63 +0.004 |? 1 | |1954AJ.....59..233M|
| 5.68 +0.05 +0.12 |4 | |1957ApJ...126..113J|
| 5.68 +0.05 +0.08 |> 4 | |1962MNSSA..21...20C|
|(J) 5.68 +0.05 +0.12 |4 | |1966CoLPL...4...99J|

Donc, pas de problème (léger écart avec la valeur publiée en 1954)

On peut également vérifier dans le Bright Star Catalog qui donne 0.05

Donc, on retient B-V = 0.05

Par ailleurs, HD 28978 est une A2V

Pour une étoile de type A2V, Pickles donne B-V = 0.02


On a donc :
(B-V)observé = 0.05
(B-V)réel = 0.02

D'où l'on calcule l'excés d'indice de couleur : E(B-V) = (B-V)observé - (B-V)réel = 0.05-0.02 = 0.03, donc négligeable.

E(B-V) est donc significativement différent de l'indice donné par Miles.

Voilà pourquoi le test de comparaison avec une A2V réalisé par Christian ne fonctionne pas : l'indice Miles est faux (et le spectre dérougie Miles est erroné)

On notera que la vérificationest facile à faire, puisque pour toutes les étoiles brillantes les magnitudes B et V sont connues avec précision.
En revanche, il peut avoir des erreurs concernant les types spectraux. Faire attention également aux étoiles doubles serrées.

Indices B-V fournis par Pickles :

BV_Pickles.PNG (4.39 KiB) Viewed 13477 times

François

About the case of HD 28978.

There is clearly an error in the database Miles, which is not difficult to identify.
For HD 28978, Miles gives a hint of color excess E (BV) = 0117
(spreadsheet Marco and Paolo facilitates the search for value)

But in Simbad, we find
B = 5738
V = 5681
Hence B-V = 0057

We check that the published values ​​correspond to this value. Also in Simbad are:
| V BV UB | Q / nbm | remarks | reference |
-------------------------------------------------- ----------------
| +0.05 +0.12 5.68 | 4 | | 1966CoLPL ... 4 ... 99J |
| 5.63 0004 |? 1 | | 1954AJ ..... 59 .. 233m |
| +0.05 +0.12 5.68 | 4 | | 1957ApJ ... 126 .. 113J |
| +0.05 +0.08 5.68 |> 4 | | 1962MNSSA .. 21 ... 20C |
| (J) 5.68 +0.05 +0.12 | 4 | | 1966CoLPL ... 4 ... 99J |

So, no problem (slightly different from the value published in 1954)

One can also check in the Bright Star Catalog, which gives 0.05

So we retain B-V = 0.05

Moreover, HD 28978 is an A2V

For a star of type A2V, Pickles gives BV = 0.02


We have:
(B-V) = 0.05 observed
(B-V) = 0.02 real

Hence we calculate the excess color index E (BV) = (BV) observed - (BV) = 0.05-0.02 = 0.03 real, negligible.

E (BV) is significantly different from the index given by Miles.

That is why the comparison test with an A2V directed by Christian does not work: the index is wrong Miles (Miles stopped ringing and the spectrum is wrong)


Note that the check is easy to do, since all the bright stars of the B and V magnitudes are known with precision.
However, there may be errors on the spectral types. Pay attention also to the tight double stars.

B-V indices provided by Pickles:
[attachment = 0] BV_Pickles.PNG [/ attachment]

[Martin Dubs]

Hello,
Robin recently asked the question in spectro-l if metallicity has an effect on the continuum, or in other words if the spectral type and luminosity class are sufficient for the selection of a calibration reference star. My answer I post here because some illustrations help the discussion. Since the Miles library has selected stars in particular with different metallicity it can be used to answer this question.
I already had prepared earlier a table of the catalog stars sorted after spectral class copied from
http://miles.iac.es/pages/stellar-libra ... alogue.php
I looked at a spectral class where I found several spectra of the same luminosity class (A1V), suitable for spectra normalization and checked if I find a correlation between metallicity (Fe/H) and effectice temperature of the fitting Planck slope:

A1VMiles.PNG (7.65 KiB) Viewed 13366 times

As a correlation seems to be there I plotted the two extreme spectra, corresponding to:

A1VMilesstars.PNG (6.42 KiB) Viewed 13366 times

HD60179 is better known as Castor A. By coincidence I had used it earlier as a reference for my echelle.

A1comparison.png (4.62 KiB) Viewed 13366 times

I also checked that reddening (Eb-v) was small.
Other spectral classes show a similar correlation (B2IV and B8). Metallicity seems to have a quite large effect on the continuum shape, although it has to be noted that the example shown here are two extreme cases. Does anybody know more about this?

Regards, Martin

[Robin Leadbeater]

Hello Martin,

Thanks for doing the leg work So for accurate instrument response correction, just chosing a spectrum of the same spectral type and luminosity class is clearly not sufficient for best accuracy as both interstellar extinction and metallicity will affect the result. As far as I can see, the only guaranteed method is to compare the measured spectrum with one of that star from a reliable source, eg the non dereddened Miles spectra

Cheers
Robin

[Ken Harrison]

I'm trying to use Marco&Poulo's spreadsheet for the Miles database...
I can't seem to be able to down load the test.dat files from their nominated site - and the other sites have the database in test.fits format...
What's the answer???

[Martin Dubs]

Hello Ken,

look up the star under
http://miles.iac.es/pages/stellar-libra ... alogue.php
choose your star and right click on the green arrow. Save the file under the star number, with the ending .fit:
e.g. HD00004_s0004.fit. You can then open the profile in ISIS or other programs. Note that these profiles are dereddened, as discussed above.
Of course you can also load the reddened (this is what you need) in ISIS - Profile - Database, but this gives only a selection of the whole library.
See also the earlier post of Marco

Sun Jan 29, 2012 9:28 pm
Hello,

following the very important last posts by Christian and François related with the MILES catalogue de-redded spectra inconsistency, and thanks François to share the link where to download the original MILES spectras (before de-reddening), as promised today we have updated the excel file posted yesterday by me and Paolo Berardi.
....

Paolo and Marco

Similarly you can also access the full NOAO database at
http://www.noao.edu/cflib/V1/TEXT/
after checking at http://www.noao.edu/cflib/V1/table4.pdf if the star is in the database.
I made an EXCEL sheet where I can look up a suitable star from the data base, selecting only bright stars e.g. or north of a certain declination etc.:

cflibdb_NOAO.zip

(197.66 KiB) Downloaded 631 times

Save the star with the ending .txt as this is a text file.
Before you can load it in VSpec or ISIS you have to remove the header with a text editor and save it as *.dat
Hope that helps, maybe there is an easier way.

Regards, Martin

[Andrew Smith]

I'm trying to use Marco&Poulo's spreadsheet for the Miles database...
I can't seem to be able to down load the test.dat files from their nominated site - and the other sites have the database in test.fits format...
What's the answer???

Hi Ken - I just managed to do it ok - Regards Andrew

[Ken Harrison]

Thanks for the help and input....
I found in my case that I had to visit the quasar site and re-set it to "Web" by pressing the button on the top right hand side.
When this was done and the full path of the file entered, it started downloading.