How many darks do you use for your cooled camera?

No Darks, 40 Bias.

Using a QHY600 mono.

CS 
Chris

I use a masterDark file created from 25 darks from a MasterDark Library I created for temp, gain, exposure and bin settings.
I use FLat frames -25
I use Flat-Dark frames -25
I do not use Bias Frames.

If you use a CMOS sensor and use Flat-Darks, do not use Bias Frames.

20 darks
20 biases (or dark flats, depends on the camera)
20 flats

I have a darks library that contains darks with different exposure lengths. Once every 6 months I make new onces.

For any  of the 3 cooled cameras I've owned I've always used 20 darks and 100 bias to make up my dark library. Then 20 flats each morning after.

I don't use calibration frames, only dithering with ASI 2600MM which is cooled in the range -15C to -20C depending on the ambient temperature.

Usually 20 darks, 30 flats, 30 biases (CCD) or 30 flat-darks (CMOS).

I do 32 darks for my ASI 2600 mc pro. It hits a great threshold of size/benefit for me. I do 32 flats and around 100 bias when I do bias.

Be aware that darks and bias introduce noise. As a consequence you need to take many of them if you do not want to destroy information by calibration. Just making 10 Bias or 10 darks will signficantly decrease your S/N ratio. For bias it is recommended to take at least 100, in Pixinsight I use the superbias process that simulates an insane number of bias in order to provide a superbias. But you need already a good number of bias to do this. Darks can be around 40-50, I saw a significant difference in my subs when calibrating just with 10 darks or with 40-50. I also depends on your camera model of course: for short exposures with cameras which have a low dark current you might skip Darks! For Flats where the signal is much much higer (about 50-65% of the histogramm) you might just take 15 or 20.
My dark library is done with the highest exposure time of my subs (480 mm at present) and in Pixinsight I can optimize (downscale) the calibration for lower exposure times. However you cannot upscale this, for instance calibrating 600s subs with 240s darks is not appropriate and can deliver bad extrapolations.

20 for darks, dark-flats, and flats. You don't really need more.

https://www.astrobin.com/forum/c/astrophotography/deep-sky/lets-discuss-about-dark-bias-dark-flats/?page=1


see John Hayes post. Yes, more darks is better, but you very rapidly get to a point where gains are meaningless. 16 is his recommendation as a reasonable number. I use 25  with a 294MM and 533MM. I can tell you that the amp glow on the 294MM is not any better corrected with 100 darks than with 25. I’ve actually used 100 darks at one point and it was not meaningfully different than 25.

Hi,
It will depend on the camera of course, and even with the same model it's strongly recommended to do some test due to the fabrication process of the chip.
In my case, with an ASI 2600 MC, I use 40 darks (and 40 bias, and 40 flats and 40 dark flats). Maybe it's not necessary, but the WBPP is doing well with all that data.

I did a video on my YouTube Channel about that (it's in Spanish, maybe you can use the automatic translator subtitles) and I measured the ratio signal-noise.

https://youtu.be/3IbWIjjo6ig?si=78ryZxm8ioGSWCRH

Clear skies!

Patrick Stevenson:
I use a ZWO ASI 533 Pro OSC.  Last year I did a number of comparison tests both using and not using calibration frames.  The result is that I don't use calibration frames at all.  The flats clean up dirty optics and vignetting.  I keep my optics clean and there are a number of programs (pixinsight) that remove vignetting so I don't need flats.  The new CMOS cameras do not need darks since they were used to clean up CCD camera images; so I don't need darks or Bias.  You can check my images at "Patrick Stevenson" to see the results.  This image was done a few weeks ago and show the results pretty well.  I still use WBPP to stack my images.

Hi Patrick,
I want to point out that you are short changing yourself by not calibrating.  First off, flat calibration does 3 things:

1) As you pointed out, it removes the effect of mechanical vignetting from edges that intercept parts of the off-axis marginal ray bundles and from dust, digs, and coating flaws in the optics.
2) Flats also remove radiometric cos^4 irradiance fall off in the field.  This is a symptom of ALL imaging systems and it's due to simple geometry.
3) Finally, flats calibrate out the effects of PRNU, which is the variation of responsivity of the pixels across the sensor.  Because of the read-out structure of CMOS sensors, this may not be an insignificant effect but it will depend on what sensor you use.  You can easily see FPN (fixed pattern noise), which is what PRNU causes by zooming way in on one of your master flats.  All that graininess is FPN and that's what flats help to remove.

It is simply not true that "the new CMOS cameras do not need darks"!  If you closely examine a master dark, you'll quickly see that CMOS sensors clearly show dark signal.  The most obvious component of the dark current shows up as hot and warm pixels across the sensor.   If  you don't remove it, you are relying on dithering along with the statistical stacking filters to eliminate the small scale component of the dark signal.  Remember that dark current comes from the photosensor--not the read out method.  One big reason that dark current appears to be so much lower in CMOS sensors is that they typically have much smaller pixels than CCD sensors.

You can often get away without removing bias current so long as the chip is properly trimmed for zero offsets.  Including bias in the calibration data makes sure that any offsets are properly removed to insure accurate results.  Taking bias data is trivial and I can't understand why so many folks are so eager to toss it out.

The only reason that you might not have seen a difference between your calibrated and uncalibrated results might be because your threshold of detecting a difference might have been set too high.

John