Re: Dumb question

[Tom Droege <tdroege2@earthlink.net>]

First, on reread, I don't find the following particularly tactful.  I am 
just trying to discuss this subject and to learn what I might try to do 
with the data I am taking.  I hope no offense is taken.

I think this is all a question of error.  99% of the light flux for what 
error?  For what absolute photometric error?  I am still struggling to 
decide what I should be trying to do.  Meanwhile, I am keeping the raw 
data, and I hope enough calibration stuff so that I can later reprocess the 
data to get the lowest possible error.

My real goal is to get something useful done in my lifetime.  I define 
lifetime as ten years (if I am lucky).

So I think in terms of what I might get done in ten years.  This is not 
really much different from a professional who is lucky if he can keep 
funding for a project like this for ten years.  ;^)

I am in a suburban location with lots of light pollution.  I don't think it 
is possible to do the best absolute photometry here.  So I plan to do what 
I can and to try to characterize the error.  I think I will be quite 
content to get absolute photometry of 0.05 to 0.1 mag.   When I read 
experts comments on the quality of the currently available catalogs of the 
size I expect to accumulate (1 billion measurements of 5 million stars) I 
think that this might be useful.  Experts might comment.

My plan is then to do what is possible here.  I have processed some data 
sets with different apertures.  I find that Michael's choice of 4 pixel 
radius aperture photometry to give near optimum results with my data.

I would be very interested in arguments as to why I should do something 
else.  I stand ready to support anyone who wants to do work on this with 
suitable data sets.

>The
>quest for accuracy in deriving magnitudes to a standard system is what
>absolute photometry is all about.

While I too wish to strive for the best possible accuracy, I also wish to 
strive to get something done.  There is some compromise to be reached 
between generating a lot of data of poor quality, and a small amount of 
data of high quality.  I must say that my bias is to try to produce a large 
catalog of pretty good relative photometric accuracy with a knowledge of 
the absolute error.

>I intend to advise anyone wishing to use the
>spreadsheet to reduce CCD images to make sure that they choose a photometric
>aperture that includes 98-99 percent of the light flux from nominal point
>sources not only for the entire area covered by the CCD image but also from
>image to image used in the reduction.

Why 98-99%?  Why not 99.99%? This must imply some anticipated absolute 
error.  In fact, as the aperture increases the collected light increases, 
and the noise increases.  At some point the signal goes into the noise and 
no further benefit of increased aperture is obtained.  Seems to me that 
such a statement should include "the aperture is selected to reduce the 
error to a minimum with the observed sky brightness".   Possibly this 
anticipates only using data from the best photometric nights.  I don't have 
many of those.  I always have a lot of light pollution.  So the choice 
depends on what one is trying to do.  I think this points out the 
difference between making the best measurements on a single star and 
measuring all the stars in the northern sky between mag 8 and mag 14 in 
many epochs.   That is what I am trying to do.

Tom Droege




At 02:42 PM 12/13/02 +0000, you wrote:
>Chris,
>
>Good point.
>
>On this subject, you may be interested to know I am putting together a
>complex Excel spreadsheet (on behalf of the Variable Star Section of the
>British Astronomical Association) designed for absolute (as well as
>differential) photometry.  I keep coming up with the realisation that
>aperture methodologies designed for differential work are not readily
>transportable to absolute photometry especially where not all standard stars
>or reference stars are on the same single CCD image.  My conclusion is that
>it is actually necessary to utilise apertures that include about 99% or more
>of the light flux  from each source.  The reason is that several things
>conspire to change the PSF of images both across the image (e.g. coma and
>other optical aberrations) but also from image to image (e.g. tracking and
>differences in 'seeing' at different zenith distances, etc.).  I know that
>this sacrifices some potential improvement in precision because of the
>'over-large' area of sky background included in the software aperture
>centered on the star or asteroid but it does this with a concomitant
>improvement in photometric accuracy not achievable in any other way.  The
>quest for accuracy in deriving magnitudes to a standard system is what
>absolute photometry is all about.
>
>Using 'cut-down' apertures or PSF fitting come into their own for
>target objects that are relatively faint where the precision is especially
>limited by photon statistics.  However, for relatively bright objects
>(viz. filter photometry of objects having V<11 or 12 using a 100 mm diameter
>collecting aperture), I intend to advise anyone wishing to use the
>spreadsheet to reduce CCD images to make sure that they choose a photometric
>aperture that includes 98-99 percent of the light flux from nominal point
>sources not only for the entire area covered by the CCD image but also from
>image to image used in the reduction.
>
>Cheers,
>Richard Miles
>
>----- Original Message -----
>From: Chris Lloyd <cl@ast.star.rl.ac.uk>
>To: <tass@listserv.wwa.com>
>Sent: Friday, December 13, 2002 11:25 AM
>Subject: Dumb question
>
>
> > Forgive me if this is a dumb question, but I suppose the aperture used for
> > the photometry is large enough to encompass the whole image at the centre
> > and edges of the field. I was prompted to ask by the comments in a recent
> > posting about the shape of the images at the edge of the field. Having
> > looked at TN70 I can see where the figure of 7 pixels comes from but
> > depending on how the aperture is centred it may be possible to lose the
> > faint extremities. I wondered if any experiments had been done to compare
> > photometry across the field using different aperture sizes. If not then
> > perhaps DS24 is where to do it.
> >
> > Apologies for any dumbness...
> >
> > Chris
> >
> >
> >
> >