Dark calibration and optimization questions

[magnusl]

Hi!

Working with a new DSLR (a 1000D, new for me) and running into pixels being clipped to zero, has triggered a few thoughts on my calibration process. I have 2 questions regarding dark calibration and optimization. First calibraion:

In his book, "Inside Pixinsight", Keller recommends integrating darks without calibrating the individual dark subs with master bias. Then the Calibrate box in the master dark section is ticked in the calibration of lights (p. 16).

I've followed a slightly different procedure: I calibrate individual dark subs with my master superbias, then integrate (same procedure as Keller describes for integration), and then in the calibration of light subs NOT ticking the calibration box.

Now is there a real difference in these processes? if it is just a matter of subtraction, they should be identical (if my algebra works...). Or am I missing something else?

For optimization: I tend not to tick the optimization box, as my light subs normally are taken with the same exposure time as the dark subs. And then I found that in my master dark, there is actually no information about exposure time - no FITS keyword about that is preserved. How then can optimization work? And should I use it, even when having the same exposure time?

I've uploaded a master dark, master flat, an uncalibrated light sub (fit) and a calibrated sub (xisf) in this folder:

https://www.dropbox.com/sh/fs744ujp8oed79d/AACfnztnt-E98N8km8Tv2dUBa?dl=0

Any comments on this is warmly welcome.

Best,

Magnus

[bulrichl]

Hi Magnus,

In his book, "Inside Pixinsight", Keller recommends integrating darks without calibrating the individual dark subs with master bias. Then the Calibrate box in the master dark section is ticked in the calibration of lights (p. 16).

Regarding your MasterDark:
You should build a new, not precalibrated MasterDark using what Warren Keller recommends: no precalibration of the dark frames, no precalibration of the MasterDark. Simply integrate the dark frames using the parameters that he specifies. Then there are two options for the light frame calibration: either to use dark frame optimization or not.

Since the proper use of PixInsight's ImageCalibration of light frames with a MasterDark was very recently discussed in this forum and I am not in the mood for repeating it, please read this thread: https://pixinsight.com/forum/index.php?topic=11950.0

For optimization: I tend not to tick the optimization box, as my light subs normally are taken with the same exposure time as the dark subs. And then I found that in my master dark, there is actually no information about exposure time - no FITS keyword about that is preserved. How then can optimization work? And should I use it, even when having the same exposure time?

PixInsight doesn't use exposure time for dark frame optimization at all, see https://pixinsight.com/forum/index.php?topic=8839.0

---

Moreover, there are some pecularities with your MasterDark and MasterFlat: In the FITSHeader of your MasterDark, these values are strange:

ImageIntegration.totalRejectionLow: 49.669%
ImageIntegration.totalRejectionHigh: 2.499%

That is not what you want; we use rejection in integration of dark frames only in order to eliminate artifacts like cosmic ray hits, so the rejection rate should be VERY low. I don't think that the high rejection rate is due to the used rejection method and parameters (linear fit, lfit_low=5.0, lfit_high=2.5), though the recommended method for the integration of dark frames is windsorized sigma clipping. Presumably there is something wrong with your dark frames. Take a close look at your dark frames and compare them. Do this before you build a new MasterDark.


Regarding your MasterFlat: the medians of the "channels" of the CFA image are extremely different:

   Median
R:   259
G:  2244
B:  2163

I guess you used skylight flats; the problem here is, that there is almost no red signal. The usage of this MasterFlat complicates the judgement of the calibration process. So I suggest first to omit the MasterFlat, until you are sure that the calibration with the MasterDark works correctly.

Bernd

[magnusl]

Hi!

Thanks, very useful!!

Best,

Magnus