I've been noticing that I am not able to completely calibrate the fixed pattern noise from my images. For reference, I am processing through
PixInsight. Frames are collected using a a ZWO 183MM Pro sensor. My flat fields are taken using a Spike-a Flat Fielder. I generate a master
flat per filter using sixty exposures at the same gain and offset values as the light frames, themselves. The histogram of a single flat field
looks reasonable to me. I also apply the necessary Bias master and Dark frame master when calibrating the raw flat field frames off of the camera.
Looking at an individual, calibrated light frame, it's very difficult to detect any residual FPN. But if I aggregate many calibrated
light frames, e.g. two hundred, the FPN becomes perceptible.
I basically am following the procedures outlined in Charles Bracken's book, "The Deep-Sky Imaging Primer, 2nd Ed." Since I am
working with CMOS, I never select the Optimize option at any point in either generating the flat field master, or the calibrated
target lights.
Here is an overly stretched example of the Leo Triplet with about 400 calibrated light frames to illustrate my point.
Edited by AXAF, 04 January 2020 - 03:01 PM.