How to despeckle in PI?

[SkyHiker]

You should do a poll with 4 choices.
1. No Darks, Bias, DarkFlats
2. Darks only
3. Darks and Bias
4. Darks and Dark Flats.

I'm putting my money on #4 by a landslide.

Any scenario without flats, optionally combined with dark flats, are useless, so 2 and 3 are pointless. Darks and dark flats can compensate for bias so 1 is also pointless, assuming there are hot pixels. Based on the DSS workflow documentation, it would be:

a) Bias and flats (only valid with negligible dark current)
b) Bias, darks and flats
c) Bias, darks and dark flats
d) Darks and dark flats

Scenarios b, c and d should be generally valid, a is invalid if there is significant dark current. I don't know how PI supports stacking but according to Luc Coiffier b, c and d are the only generally valid scenarios.

Everybody else can't be wrong.

Not everybody agrees, according to the link I provided. You may very well be right, and that may be a fluke, who knows. Like I mentioned, most people who make the comments either don't have an ASI2600 or haven't tried the comparison using an ASI2600. Most respondents to a poll will belong to that group. The only way to find out without relying on general comments is to try it with an ASI2600.

[AstroBee]

Of course I wasn't dismissing flats. My poll suggestions were in regards to darks, dark flats and bias files, not the exclusion of flats. I didn't mention lights either. I just assumed we understood that.

Good luck.

[SkyHiker]

Just try to pass on my experience with Cosmetic Correction.

I had really good success with WBPP until one dataset. I had hot pixels scattered throughout the stretched masterlight.

I submitted a thread to the PI forum.
https://pixinsight.com/forum/index.php? ... sue.16560/

Take care and good luck!

Hey Jim, I tried this with various parameter settings, and got it working in 2 steps. Doing the CC once as part of WBPP with automatic hot pixels checked removed the hot pixels but left some warm pixels. I tried it with a hot sigma of 3 and 0, and the results are the same. Checking CFA is definitely needed. I was able to get rid of the warm pixels by applying CC once more on the stacked/gradient reduced image by applying CC again with a hot sigma of 0 (3 did not work). Here are the results (the background color of the first one is different because the background gradient removal was done manually, the other two automatic),

No CC applied

CC in WBPP only

CC in WBPP and CC on stacked image hot sigma=0

I read up on hot pixels to get a better understanding of it. I found this excellent description on the Hubble website. It talks about hot and warm pixels, both stable and unstable. For unstable hot/warm pixels, a library won't work well but automatic CC most likely will.

I also learned that most hot pixels are caused by dark current, which is why darks are very helpful for removing them. I always thought that hot pixels were removed by including bias frames, but I was wrong. They can be in bias frames, and there are several articles online stating this, but they are often talking about darks because they assume the same exposure time as the lights.

Given that automatic CC can remove the hot pixels, I will stick to just bias and flat frames for calibration until proven wrong. The difference in dark current between the ASI1600 and ASI2600 is a factor 10 (from the dark current curves on the ZWO website). This is supported by many ASI2600 users who successfully imaged without darks seeing no or hardly noticeable differences, including users on TSS who did the comparison.

Thanks for your advice!

[Juno16]

Excellent Henk!

The speckles certainly put you thorough the paces!

Glad that you found a CC solution that works for you.

I wonder if CC would be less complicated if darks were included in calibration?

Either way, you have a workflow now in PI that takes care of the issue and that’s a very good thing.

Thanks for rge follow up and the link to the Hubble site!

[SkyHiker]

I wonder if CC would be less complicated if darks were included in calibration?

I think it would. But who wants to deal with darks if you don't have to? You have to create dark libraries for a specific temperature and exposure time, that sounds limiting to me.

Check this Sky and Telescope article on the future with sensors that don't have dark noise (2/3 down). The author says he already used a camera that allows it (I wonder which one, an ASI2600 maybe?), and says that hot pixels can be removed by a rejection algorithm - as shown here. A factor 10 down compared to an ASI1600 is not to be ignored, not by me at least.

BTW I did not dither - my images so far were not good enough to even worry about that - and I have never used the cooler for cooling. I turn it on because otherwise Ekos can't recognize the camera. There was no need for cooling is because I never saw signs of dark current like you see with most DSLRs. If dark current is responsible for hot pixels it would help. I don't know for sure if hot pixels are caused by dark current, or by defective pixels with abnormally high gain, which would lead to the same effect, I think. See aforementioned HST article. I will start dithering and cooling.

[Juno16]

Pretty cool stuff Henk.

If I had to guess, I would say that the sensors with little to no dark noise started with the 533/2600 series of sony sensors and all that have introduced since then.
Dithering certainly should help with hot pixels (back to the rejection algorithm).
I’ll keep using darks because it is my workflow. Creating a dark library is no biggie for me. I put the camera in a closet and have a sequence in N.I.N.A. to produce 25 dark files at 0C and -10c for exposures of 60, 90, 120, 180, and 240 seconds. I Teamviewer it to my phone so I can keep up with things. My current set is 1 1/2 years old and I figure that I’ll make a set every few years.

All very interesting stuff. Thanks for sharing your findings!

[Jockinireland]

Just to add a few bits from my experience with the asi 2600mc

I use darks, flat darks and flats.

My darks all show hot pixels when stretched. The longer the exposure the worse they get


When I calibrate with darks and cc I never have hot pixels in my final image

I have tried without darks and found I then get hot pixels showing through.

Without dithering hot pixels become more of a problem. That's what walking noise is.

I have read many comments saying darks are not needed and that statement is usually answered by several statements to the contrary, usually always referring to hot pixels.

As always in AP, ask three astrophotographers and you'll get 4 opinions. So do what works for you.

Using flats and dark flats is easy. I use the ccd flats tool in apt to define that flats exposure and cloning that, throwing on the lens cap and taking the dark flats adds less than 5 minutes to my end of session routine.

Using darks is no chore My dark library is 18 months old and consists of master darks for the various setting plans I use. So it requires 1 file to be placed in WBPP. As I said above, without them I get hot pixels in my final image.