Calibration...Once Again

[sreilly]

I use the separate Image Calibration process to calibrate all my images. My process follows those outlined to make my master calibration frames following this tutorial http://www.pixinsight.com/tutorials/master-frames/index.html . I have just recently tried making SuperBias masters as I usually take 200 bias frames. To calibrate my light frames I use my master dark, usually at least 30 frames combine to make the master with the same time and temperatures as the light frames. At this time of year or the last few months, -15 for 900 and 1800 seconds respectively. My flats are done using a FlatMan XL panel and the auto flat script in ACP. Due to the fact I use a 3" Pyxis rotator, I take 40 flats, 20 at the given PA and 20 rotated 180 degrees (E/W), resulting in a master flat consisting of 20 frames each. Using my master darks and flats I calibrate the images using Image Calibration process as pictured below. The question is if this is the prescribed method? The reason for the West side flat being used for the East side images is because of method of pierside differences between the program taking the images and ACP's interpretation of the same. Evidently this hasn't become an ASCOM standard yet and the imaging train is not perfectly centered thus the need for E/W flats. Bottom line is there a better way to calibrate and get better results? And please don't suggest the Batch Preprocessing script as I much prefer to control the entire process and check each image after processing for defects before combining.

Thanks,

Steve

[sreilly]

I should have mentioned that these lights are 1800 second -15 degree images hence the proper dark.

[mschuster]

Steve, normally when doing master dark frame optimization, both a master bias and master dark are specified. This would typically be done both for flat and light calibrations. You did not specify a master bias, and I am not sure what the dark frame optimize option does with no bias specified, maybe nothing at all?

My own preference for long darks that are both time and temperature matched, is to disable master dark optimize, not specify a master bias, and just have the master dark simply subtracted from the flat or light frame. My detector on 2400s darks has a lot of annoying warm pixels and trying to scale them the same way as normal pixels as well as trying to measure a useful noise metric just does not make sense IMO.

Thanks,
Mike

[sreilly]

Mike,

I always use matched darks except with the short flat field images. But I've found that sometimes leaving dark optimized does help somewhat although I've never quantified that.

Steve