|
OK, my take on the terminology (And workflow) and trying to to make too much war and peace.
"Integration" is commonly used to describe the stacking of multiple light frames, but it is actually more like averaging. Integration builds values up and up over time, the actual exposures are an integration. The longer the exposure or the more photons arriving in a given time, the larger the pixel value and the bigger the difference to random or repeating noise, which will occur less often so does not integrate up as strongly. I think "Stacking" is better terminology.
"Calibration" is done before stacking, and what the calibration frames do depends on what type they are.
Dark, flat and bias frames. You stack these to make master darks, master bias and master flat frames (Include dark-flats here, they are nothing different to a normal dark frame, just the exposures match your fixed length flats)
Master dark, master dark-flat and master bias frames calibrate lights or flats by subtraction. They are subtracted pixel by pixel from light or flat frames. Master darks and dark-flats subtract fixed noise (Read noise) hot pixels and dark current (All predictable and repeatable) Master bias frames subtract only the fixed noise.
A master flat is made up of calibrated flats (With a master dark flat or master bias, depending on camera and preference) divided in to your light frames. Say the centre of the master flat frame has a brightness of 100 and the corners have a brightness of 50, that is 0.5 times the illumination in the middle. The corners would be divided by 0.5 to make them as bright as the middle.
After the calibration frames (Which are applied to each light) you should have lights which don't have read noise (On average) Don't have hot pixels, don't have any dark current and the illumination should be flat from centre to edge. THEN they can be registered and stacked to average out the shot noise and reject outliers!
|