[vsnet-chat 4280] Visual observation precisio: Was vsnet chat topics

[Thom Gandet <tgandet@mindspring.com>]

Taichi Kato wrote:
> 
> Re: [vsnet-chat 4265] Re: vsnet chat topics
> 
> > When trying to interpret a star's behavior based on visual data,
> > there is inevitably going to be uncertainty involved. There is
> > usually about a magnitude of scatter in visual observations, and I
> > find it not at all unusual to have a mix of "fainter than"
> > observations in the midst of the positive data. That is fine.
> 
>     Even CCD observations may suffer the same kind of bias.  As may be
> well known, reporting only positive detections (or successful software
> measurements) can lead to overestimation of the true stellar magnitude.
> This is particularly true in eclipsing CVs, which can easily reach
> very low light levels at eclipse centers.  The same effect is present
> in virtually any measurement; the effect is simply most prominent in
> visual photometry.
> 
>     [There are a lot of other subtleties in CCD photometry of faint
> sources.  I omit them here for simplicity.]
> 
> Regards,
> Taichi Kato

   Excellent points.  Visual magnitude estimates are the only photo-
metric measurements made with uncalibrated instruments - the
individual human eye.  Forgive me for emphasizing what most everyone
here knows, but CCD and PEP observations can be (and ought to be)
transformed to a standard system using observations of standard stars,
made with standard filter sets and (more or less) standard detectors.  
Corrections for differential extinction, etc., must also be done
carefully.  Measurement of the images with photometric software is
another source of bias that is unique to CCD photometry, although
there are analogous sources of error in photographic photometry.
Such observational and reduction techniques can achieve very precise
and accurate photometry, even of faint objects.  The search for
eclipses by putative exosolar planets depends on achieving that
level of precision.

> There is usually about a magnitude of scatter in visual observations,
> and I find it not at all unusual to have a mix of "fainter than"
> observations in the midst of the positive data. That is fine.

   Does the 1-mag scatter cited refer to the scatter between different
observers?  The standard deviation for the "average observer" is
around 0.2 magnitude I recall, and I have seen what appear to be
consistent zero-point differences between different visual observers.
It sill amazes me that some observers are able to estimate to a
precision of less than 0.1 magnitude.  I don't consider myself a
particularly good estimator, but my observations of standard are
very consistent (around +/- 0.03 mag., standard deviation) over
relatively short time intervals during one night.

   Not meaning to start a heated controversy by this, but some
consider, and I think rightly so, that Cousins did what is probably
the most precise and accurate PEP in the world - and he observed
from a light-polluted sight.  Arlo Landolt's observations also
constitute perhaps the largest body of precise and accurate PEP
measurements.  They used careful observational and transformation 
techniques.  Others have also accomplished similar levels of
precision and accuracy, both CCD and PEP, using similar rigorously
applied, more or less standardized, techniques.

   Analogous transformation equations for visual observations would
undoubtedly vary from observer to observer and night to night (and,
in my own case, seemingly from hour to hour!).  There may be
value in transforming visual observations in a similar way.  They
_could_ be corrected for differential extinction, based on
observations of standard stars (perhaps using the M-67 project
methodology, or some variant of it) at various zenith distances
during the night.

   That doesn't necessarily require a lot of additional observing
time, but perhaps too much for what is an unclear gained from it
all.  Experiments with my own observations have not been very
conclusive - nor very extensive either, I admit!


Regards,
Thom Gandet


Taichi Kato wrote:
> 
> Re: [vsnet-chat 4265] Re: vsnet chat topics
> 
> > When trying to interpret a star's behavior based on visual data,
> > there is inevitably going to be uncertainty involved. There is
> > usually about a magnitude of scatter in visual observations, and I
> > find it not at all unusual to have a mix of "fainter than"
> > observations in the midst of the positive data. That is fine.
> 
>     Even CCD observations may suffer the same kind of bias.  As may be
> well known, reporting only positive detections (or successful software
> measurements) can lead to overestimation of the true stellar magnitude.
> This is particularly true in eclipsing CVs, which can easily reach
> very low light levels at eclipse centers.  The same effect is present
> in virtually any measurement; the effect is simply most prominent in
> visual photometry.
> 
>     [There are a lot of other subtleties in CCD photometry of faint
> sources.  I omit them here for simplicity.]
> 
> Regards,
> Taichi Kato

-- 
************************************************************************
                      Lizard Hollow Observatory
                      Thomas L. Gandet, Director
                            PO Box 77021                   
                      Tucson, AZ 85703-7021  USA
************************************************************************

VSNET Home Page

Return to Daisaku Nogami

---

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