Pixinsight - Cosmetic correction

[Spookyer]

Hello, I am trying to use cosmetic correction in pixinsight with my DSLR.

With hot pixels I am having trouble seeing how much I should move the sigma setting, the image looks like a lot of colour noise up close and it is difficult to make out the hot pixels and how far I should move the slider.

With cold pixels it is reporting over 1 million cold pixels even with the sigma slider in the default position way over to the right? It doesn't then go up until the sigma slider is set nearly all the way to the left.

This is nothing like Harrys tutorials, any suggestions as to what I am doing wrong?

Brett

[RickS]

Are you using a Master Dark or Auto Detect, Brett? I've never used CosCorrect with colour images but it has always behaved reasonably for me with mono subs.

[White Rabbit]

Yes you need to use a master dark, you can use a single dark but a master works better.

You need to create a preview to work with as it will not show up on the original. Alt+n, to create a preview, then drag a box around the area that has some hot and cold pixles, select the preview. Enable real time preview on the cosmetic correction panel and your away.
I must admit that the auto detect works really well. Every time I've used it I do both manual and auto just to see the difference and auto is just as good as doing it manually.

Cheers.

[RickS]

You don't need a master dark to do Auto Detect. It just compares each pixel to its neighbours and corrects any which vary more than the specified number of deviation units.

[niharika]

Hi Rick, I need to clarify something. Do you see an increase in FWHM if you apply cosmetic correction? I am getting an increased FWHM according to ccdinspector when applying cosmetic correction compared to not applying cosmetic correction. FWHM does not get effected if I use Defect List for columns only but will get effected once I select 'Use Auto Detect'; hot sigma 3; cold sigma 3. Is it expected?

[RickS]

No, I wouldn't expect that. I'd use the real time preview on CosCorrect to check out what is happening to the stars and adjust the sigma parameters if needed. It could also be that CCDI's FWHM calculation isn't terribly robust. You could see what the FWHMEccentricity script thinks of your before and after versions.

Cheers,
Rick.