In photography in general and astronomy in particular, calculating the optimum amount of time that the shutter should be open and film exposed or (for digital imaging) the amount of time over which the CCD or MAMA detector should integrate can often be very simple. Many astronomical observatories can have very complicated setups and so the ExposureTimeCalculator for the HubbleSpaceTelescope? is a very detailed application.
The result of the ExposureTimeCalculator is the signal-to-noise ratio: SNR = signal / noise
Since all of the HubbleSpaceTelescope? detectors are digital, signal and noise are all measured in either "counts" or "photons" (yes, some of the optics are sensitive enough to detect individual photons).
There are three components of the calculation: source characteristics, throughput of the instrument components being used, noise sources.
Source characteristics and throughput are expressed as sampled curves. Source counts for a particular observation are the integration of the area under the curve (source) * (throughput).
The three main sources of noise are: source counts (I'm not a physicist, I'm a programmer, I can't explain the science behind this) (cosmic rays, which can be defined in this case as photons eminating from a source other than the object being viewed, and hence constitutes noise), background sky (earthshine, light in the zodiacal plane), instrument heat (for NICMOS) and detector noise. Noise is the square-root of the sum of the counts from the various noise sources.
More later.
Updated Nov 15 - happy now?
-- DonaldMcLean
Last edited May 2003. It's now June, nearly July, 2004. Does this count as later?