Removal or Hiding Dust Donuts

[Startrek]

I’ve had my new ZWOASI2600MC for 2 weeks now and have imaged mostly Nebula ( 90% of image field covered with nebulous etc... )
Last night I imaged M30 globular cluster ( first image with a majority of interstellar space in the image ) and after processing noticed 2 large dust donuts in the off axis area
So imaged M30 again ( same exposure time and camera orientation , camera temp etc.. ) this time without the Baader coma corrector and after processing noticed the original dust donuts have disappeared but now 2 new fainter dust donuts appeared in the opposite area of the image but not in same the rotation field , these fainter ones are closer to the centre if field

Very strange ??

Any ideas why or possible causes

Also is there a module in version 1.5 I can use to mask or hide these damn dust donuts

NB I don’t take flats

Thanks in advance !

[admin]

Unfortunately, this is one of the reasons why flats are virtually a must (aside from uneven lighting and.or vignetting).

It sounds like a simple case of dust particles moving and/or new debris settling somewhere else.

There are some ways you can attempt to remove dust bunnies using the Heal module, but fixing things in post will always be sub-optimal.

To use the Heal module for this, what you will want to do is put the the bunny in a mask, then run the Heal module and set the 'New must be darker than' value to a percentage that approach the brightness of background level (it may be a little bit higher). You can do this by eye. I.e. if the mean background level brightness is about 10% of full intensity (white) then set it to 15% or so.

You'll see the Heal module fill in a plausible background. The 'New must be darker than' value is simply there to make sure it is not adding 'plausible' stars as well. If you see it do that, just lower the value.

Once you're happy with the 'plausible' background, it's time to add back in any stars that were healed out. For this, launch the 'Layer' module, click Undo->Fg (which puts the original image in the foreground) and set the 'Layer Mode' to 'Lighten'. The composite will now be the result of choosing the lightest pixel of the two images for every pixel.

Hope this helps!

[Startrek]

Ivo
Thanks for your advice
I tried the Heal module and then Layer
As you mentioned it is not a natural fix but does obscure the remnants of dust donuts to some degree although the darker background behind the stars is still noticeable
Started learning how to take flats with an A3 LED sketching panel with adjustable illumination and having some reasonable success, definitely removes vignetting and helps remove the dust donuts but not totally

Getting somewhere now

Thanks again !!

Martin

[thomehm]

" put the the bunny in a mask, then run the Heal module "

I am trying to deal with dust donuts also. I click on Mask, set to lasso, I click on Keep. But Heal is greyed out. Actually, I have never seen Heal not greyed out. What am I missing?

Dana

[Burly]

Heal will be available after tracking off .

Dave

[thomehm]

Heal will be available after tracking off .

Dave

Now I am more confused. The first thing I wanted to do was to eliminate the dust donuts. Why would they make it a requirement to turn tracking off in order to use Heal? Or should I use a different tool to eliminate dust donuts?

Dana

[admin]

Heal will be available after tracking off .

Dave

Now I am more confused. The first thing I wanted to do was to eliminate the dust donuts. Why would they make it a requirement to turn tracking off in order to use Heal? Or should I use a different tool to eliminate dust donuts?

Dana

Hi Dana,

Signal evolution Tracking monitors how your signal evolves from its linear state all the way to a finished image.
Some modules insert signal that did not exist before or otherwise interfere with the ability to mathematically "reconstruct" (e.g. go forward and backward in time) the signal at any point in time. The Heal module completely makes up signal and hence any notion of tracking its "evolution" goes out the window; the engine cannot make mathematical sense of how the signal evolved and "got there" from the source dataset.

So, to use these modules, you either must turn Tracking off, or - when importin your image - indicate your image is from a non-linear sRGB source and choose "Don't activate Tracking".

I hope this helps!

[thomehm]

OK, I think I understand that. Now, the second part of my question was, should I use Heal, or Wipe, or something else to work on dust donuts and vignetting?

Dana

[admin]

OK, I think I understand that. Now, the second part of my question was, should I use Heal, or Wipe, or something else to work on dust donuts and vignetting?

Dana

The real answer would be "neither"; these issues are solved by calibration flats (they are really not optional). However, if flats are unavailable and/or calibration was not possible or successful, you can use the Wipe module's Vignetting preset to mitigate vignetting. Be mindful, however, to mask out any dark anomalies (see here) such as dust donuts or specks, as otherwise these anomalies will cause Wipe to back off.

As for dust donuts and other imperfections, these are quite hard to address in post-processing and should really be addressed by calibration with flats, so they don't happen in the first place. The method discussed in earlier posts may ameliorate them in some instances - particularly when they exists in rather even/"empty" areas, but your mileage may vary.

Hope this helps!

[thomehm]

I am using flats. I fed them into DSS. They did not correctly null my dust donuts. I made the flats, in the daytime, using a piece of white acrylic, which fits over the end of my EdgeHD8. Same principle as the stretched t-shirt. The key term here is calibration. I have read instructions advising matching Camera exposure speed, temperature, and gain. But this is useless advise if the flats were made in the daytime. Maybe this works with a flat box, I wouldn't know. I assume that the key is to get the exposure and contrast right in the flat. I haven't seen any advise, or software, that would help me get the flats right. This is not something where it just works.

Dana