[vsnet-chat 5863] Re: GSC vs Tyc2

[Arne Henden <aah@nofs.navy.mil>]

Out of curiosity, did you compare your results against
my single-night field photometry file for V4742 Sgr?

  Bt/Vt -> R is less robust than Bt/Vt -> V, so I can
believe larger error for that transformation.  This is especially
true for V4742 Sgr since the field is heavily reddened.  Also,
if you used Tycho stars from 8.8 to 10.5, you need to remember that
the errors at Vt=10.5 are quite large, and due to star densities,
there are more of these stars in your field than the brighter ones,
thereby dominating the statistics.
> 
> My conclusion was that the Tycho values were good enough for visual
> observing but of no value at all for any precision CCD or pe photometry.
> This wasn't unexpected as the Tycho system has the defects of the old
> photovisual and photographic magnitude systems. A third bandpass in the U
> region might have helped resolve some of the problems.
> Basically, if you are trying to determine V, V-R or V-I there is no
> substitude for standard stars measured with these filters.
> 
   For V, I think using an ensemble of Tycho stars (not just one!) can
often get pretty close to the Johnson V system.  There will be fields,
such as reddened ones, where there will be a systematic offset.  For
Rc/Ic, I agree that the best approach is true standards in the right
bandpass, but even here you can often come close if you restrict the
color range of Tycho stars that are used.  The trick is to use *many*
stars, not just one, so that the individual transformation errors
get averaged out.

> I notice that a number of people use CCDs with a V filter and publish (and
> presumably submit these somewhere) without transformation. The V filter I
> use has a 7.5% slope against V-R and I presume that others are similar. So
> values determined without transformation corrections are probably 3-10% out.
> Add to this the uncertainties introduced by using poor standards and it's
> doubtrful that there is any value in the measures. Bad observations are
> extremely frustrating and any method which encourages these is not good.
> 
   Using a V filter without transformation is certainly better than trying
to match unfiltered observations!  Usually there is just a zeropoint
offset between two V-filter observers that can be removed easily.  The
second order effects, such as differential color extinction or color change
in the variable, remain second-order and small.  Of course, the best photometry
generally requires two filters (or knowledge of the star colors) to
place the results directly on the standard system.  In all cases of
doing photometry, you should be careful in how you describe your
measurement.  If untransformed V, you need to indicate that somehow
along with what comparison star you were using (and the magnitude you
used for that comparison star).  If unfiltered, you need to tell people
what "system" you placed your results onto, whether V-band, R-band or
whatever.  The more description, the more useful your results become.
Arne

VSNET Home Page

Return to Daisaku Nogami

---

vsnet-adm@kusastro.kyoto-u.ac.jp