Thanks for that great idea too!

Extract 1 layer, then keep only the residual (residual). Then extract 2 layers and do the same (residual1). Then do the same with 3 layers (residual2) and 4 layers (residual3). I hope that this answers your question.

@@craigcalvert7978 thank you! Yes, it was the order of creation/extraction that I was unclear on. 1,2,3,4 got it! Really appreciate your efforts, this is a very cool technique!

Finally worked out how to create residual residual1, 2, 3 and its brilliant so thankyou! TBH I'm using this on top of blurxterminator But your method is great as both a standalone AND an an additional tool and you have made it freely available. It also does not really add much noise which is a big plus point for sharpening

Thanks for your comment. On a different video, someone asked if I could "share" my Process Container. I had to figure out how to do this and to design one that most users could benefit from using. I created this 'generic' version so that it would work on many different objects. Perhaps it is too generic? And it obviously doesn't cover all scenarios (e.g., if using masks). For those, like yourself, who know how to build and use process containers, it makes sense to build your own (or to modify this one) so that it is specific to your need.

Does PI give you an error message, perhaps that a file does not exist? Make sure there are no other files named 'original' open in PI, perhaps in a different workspace. All four residuals have to be in your workspace and labeled correctly (residual, residual1, residual2, and residual3), the first line in the container (ImageIdentifier) must have the green check next to it, and at least one of the other lines in the container (PixelMath) also must have a green check next to it.

@@craigcalvert7978 oh thats my mistake, I did not create the 4 residual(1-2-3-4) I directly transferred it to thr RGB image, Thx Craig 👍

I’ll try to make a quick video on how to do it. I’ll post it today or tomorrow.

Certainly before adding back the stars. I have a OSC camera so I work in RGB space. But the individual channels are still gray scale. I have been playing around with sharpening selective gray scales and not others, for example, R but not B and G. In this case you obviously sharpen before combining. However, I also experimented with sharpening the individual channels and then combining them, vs. sharpening the RGB (combined) image. They were identical. So it appears that you can sharpen at any step AFTER noise reduction but BEFORE adding back the stars.

@@craigcalvert7978 makes sense. I have a OSC camera (ZWO ASI294MC PRO) also. Thanks for the quick reply and looking forward to trying this out!

I don’t think that it matters as long as you reduce noise in the image before sharpening.

It's easy to see if there are any "artificial" structures. Just compare it to the unsharpened image😉

If the noise or mottling was in either the original image or an extracted residual, I suggest that you clean it up before sharpening. Sharpening can make it much worse. You can try to reduce noise with one of the noise reduction methods. If there is a small amount of mottling you can try to clean it up using CloneStamp. If there is a lot of mottling you might have to darken the background using HistogramTransformation or CurvesTransformation (preferably do it in the original image before extracting residuals). if the mottling is localized you can try masking (inverse) the region using the Game script and then just work on that region. I've run into this same issue many times Good luck!

@@craigcalvert7978 Thanks. I was commenting specifically on what looks to be left over artifacts at 18:40. Or is that just the KZbin compression making it look like noise?

@@feraxks You have good eyes! I went back and looked at the actual images and did see mottling in the original image that carried through the process. Because there are also locations on the image where it is clean, I used CloneStamp to remove this mottling. To make sure that I got all of it, I used the HistogramTansformation to brighten the image, but only in the preview screen (I didn't apply it to the image). This accentuated the mottles in preview. I then looked at the preview screen as I cleaned the image with CloneStamp until all of the mottling was removed. Thanks for pointing this out. Fortunately, it was an easy, 1-minute fix.

@@craigcalvert7978 Oh cool. I was worries that the mottling artifacts were from the sharpening process. Glad to know that's not the case. But that does emphasize what you said in your first response -- make sure noise reduction as taken place before you get started. Appreciate the responses and the interesting technique.

Thanks, and I totally agree that MT is doing most of the rounding, but it can also eliminate/reduce stars. When I was experimenting I found that increasing the intensity/size with CurvesTransformation helped to limit this reduction. So I found it useful to initiate the process using the Advsnced Star Mask. Script.