Re: stellar subtraction method -- _really_ hard

[Glenn Gombert <gleng@infinet.com>]

Michael et all,

        According to Emmanual Bertin the new version of SExtractor (due out
in March) should do most of the things necessary to support this type of
calcualtion. (See Below) at least is should be interesting to experiment
with :-)

"Concerning the discovery of variable stars, you might be aware of the new
method
implemented by one of my colleagues, Christophe Alard:
http://xxx.lanl.gov/abs/astro-ph/9712287. However I don't think the software is
available yet. With this method you just need to use the dual-image mode in
SExtractor."


Glenn G. 

At 07:15 PM 1/11/98 -0500, you wrote:
>
>  I know a bit about the method proposed by Christophe Alard
>and Robert Lupton for subtracting the PSF of bright stars so 
>that one can measure the properties of faint stars nearby.
>I would claim that it is _much_ more detailed than we ought to
>consider at the moment.  This method requires exquisite measurement
>of the PSF, at all points across a field (which takes a good bit of
>computing power and a good bit of nice, clean data) .... and then
>a whole lot of computing power to apply to the crowded sections
>of the image.  I suspect strongly that
>
>     a. TASS images have a PSF which varies pretty strongly
>          across the frame
>     b. the TASS PSF may not be adequately sampled for this method
>     c. the computing power may exceed our current resources
>     d. only a small section of the sky contains stars so crowded
>          that this method would yield results significantly different
>          than simple aperture photometry
>
>  I don't mean to stop anyone who is interested in learning about
>the method, or even writing the code to try it out on TASS images.
>However, I do think that, at the moment, this might not be the
>best thing one could do for TASS.
>
>  But, of course, we're all here to have fun -- so if getting into
>the details of such method is your way of kicking back after a
>hard day at work, then go for it!
>
>                      Michael Richmond
>
>
Glenn Gombert <gleng@infinet.com>