[Aavso-photometry] Re: Calculating mag errors for filter bands from color indices

[Wolfgang Renz]

Hello

Arne thanks for your response.

Using differently assembled formulas one has usually the choice to calc
one mag and some color indices OR mags for all bands directly. The
second way nowadays just takes little more computer processing time
(thats probably the historical cause, beside the use of the color indices
by the astrophysicists, for doing it that way).

Its right that one needs one color index for a simple transform. So the
errors of two instrumental mags influence every transformed mag and
color index.
But the point that caused my question was due to the fact of having to use
the SDSS r' mags and TWO color indices to get the z' and u' mags. This
for sure introduces enlarged errors and error estimates for the z' and u'
mags that are basically and essentially not neccessary. In general two
transforms using neighbouring bands should give worse mags and error
estimates than a single transform skipping the band in between. And
having to combine a mag and two already "worse" color indices and
error estimates gives even more worse mags and error estimates.

Doing this might be OK with (V-Ic) and (Rc-Ic) color indices as the V and
Rc band are close together and overlap considerably to give low (V-Rc)
errors (for non-very-red and non-blue stars). But the SDSS bands are well
separated to not show this issue.
So I was wondering why they choose this way to present their measurements
for the SDSS Standard Star System. As Standard Stars are of much higher
importance for photometrist than for astrophysicists I would have expected
that they do it the way thats best for photometry to create secondary
standards and to make all-sky photometry.

Wouldn't it have been better to give e.g. u'g'r'i'z' directly or (u'-r')(g'-r')r'(r'-i')(r'-z')
instead of (u'-g')(g'-r')r'(r'-i')(i'-z') for a Standard Star System ?

Clear skies
  Wolfgang

-- 
Wolfgang Renz, Karlsruhe, Germany
Rz.BAV = WRe.vsnet = RWG.AAVSO



----- Original Message ----- 
From: "Arne Henden" <aah at nofs.navy.mil>
To: "AAVSO-PHOTOMETRY" <aavso-photometry at mira.aavso.org>
Sent: Wednesday, December 29, 2004 8:49 PM
Subject: Re: [Aavso-photometry] Re: Calculating mag errors for filter bands from color indices


> Wolfgang Renz wrote:
> > I know that the non-photometry astronomers prefere color indices
> > to ease their work.
> > Calculating a color index and a rms of a color index is a pretty simple
> > task. But keeping combined error estimates as small as possible is
> > not.
> > Shouldn't the photometrists use the original values with the smaller
> > error estimates instead of color indices with already combined error
> > estimates ?

> If your goal is to report separate magnitudes for all filters
> (B,V,Rc,Ic for example), then yes, calculating each magnitude
> from its transformation will yield the lowest errors.  However,
> in practice, you usually have to calculate all of the color
> indices anyway since they are used in subsequent magnitude-
> transformation equations, such as
>       V-v = coef*(B-V) + zeropoint
> which requires knowledge of (B-V) before you can get V.  Therefore,
> the usual practice is to determine all of the possible color indices
> along with one of the magnitudes, usually V.  When you have, say, 4 filters'
> worth of information, there are many ways the data can be combined,
> and people usually take the most common way.  There is nothing to
> prevent you from calculating your values in an alternative way if
> you feel your results will be improved.
> Arne