[vsnet-chat 2119] Re: the right integration time

[aah@nofs.navy.mil]

Gianluca Masi wrote:
>In order to optimize to S/N ratio and time resolution, how can one
>determine the right integration time for her/his target (not too short,
>not too long)? How the background brightness influences this (light
>pollution, bright moon...)?
  If you want to do it theoretically, you have to look at the
signal/noise equation and add in the noise from the CCD readout,
dark current, sky brightness, etc.  Look at the noise articles by
Mike Newberry in CCDA, or in one of our book appendices.
  It all boils down to two questions:  what kind of signal/noise
do you want in an individual measure (one percent accuracy? ten percent?)
and what kind of time resolution would you like to have?  You may
find that you can't have both, depending on the brightness of the
object and your telescope/CCD/observing conditions.  Unless I am
applying for time at a national observatory, I usually determine
these quantities empirically by either experience or by performing
a test run and varying parameters to determine what works best.
While accurate photometry seems like a 'black art', it really is
just common sense.  Do the best job of calibrating your data (which,
of course, means you have to understand what effects each calibration
step might have in the processing); try to optimize your telescope
and CCD camera; find the optimal, most efficient
method of observing through trial and error; and use the best
reduction software you can to extract the most information out
of your images.
Arne

VSNET Home Page

Return to Daisaku Nogami

---

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