Jupiter, 19th Sept am.

[neil phillips]

5 hours ago, sorrimen said:

Bit late to the party, but I’ve often found that with more frames it may look softer pre sharpening. When you take it into registax however, you can push it much more without having to denoise in response, ending up with the better image. I’m not super experienced, so Neil and co.‘s advice is more important, but this may at least explain why you were seeing the difference pre registax. 

You are correct. Although the number of frames needed will vary on many factors such as scope size. transparency at the time. Exposure and gain choice. I Certainly have found more frames can help with pushing the image more during sharpening. But of course, this only works to a point because if too many bad frames are added in., you're trying to push further something that's already blurrier by definition. So, the art is figuring how many frames can you stack without too much blur occurring. So that sharpening can be pushed.

The consistency of a capture run. Is the be all and end all of how many good frames can go into a stack.

If seeing is great I have had 75% usable Because the ser file is virtually the same throughout the entire run. So, it makes sense the majority of the stack will be usable. When of course seeing is more disturbed. Thats when cutting down the frames to a point where you are adding not too many bad frames into the mix Becomes important. 

Because of the bad weather I am presently running a 5 min capture. that I am de rotating in winjupos. with around 62000 frames in it. Of course, a small stack size can be done to good effect. What it won't do is allow a lot of hard intense sharpening. Before the tell-tale mottling starts appearing. I will post the image in a bit and log how many I use from the long run. 

Edited September 24, 2022 by neil phillips

[Laurieast]

@symmetal Thanks for the excellent data!

1. 10% stack of 600
2. 40% stack of 2400
3. 50% stack of 3000
4. 50% stack of 3000 slightly (12%) AI sharpened/denoised.
5. 50% stack of 3000 slightly (12%) AI sharpened/denoised, and warmer colour balance.
All Registax RGB Balance and Wavelets. Brightness up 10 in PS. 

Edited September 26, 2022 by Laurieast

[neil phillips]

1 hour ago, Laurieast said:

@symmetal Thanks for the excellent data!

10% stack of 600
40% stack of 2400
50% stack of 3000
50% stack of 3000 slightly (12%) AI sharpened/denoised.
Registax RGB Balance and Wavelets. Brightness up 10 in PS. 

Hi Lauireast  was the sharpening the same on all 4 stacks ? 

[Laurieast]

3 hours ago, neil phillips said:

Hi Lauireast  was the sharpening the same on all 4 stacks ? 

All were identical default wavelets.
Default wavelets Layer 6@100 5@70 4@50 3@30 2@3.0 1 off/unchecked.

On the last one, the 50% of 3000,  I posted that one without and then with the Image Analyser AI sharpening. 

They all look slightly blue to me, but I did not want to process them in any way other than the RGB and wavelets. 

Edited September 24, 2022 by Laurieast

[vlaiv]

Very nice data and some very nice renditions.

Here is my quick process:

[neil phillips]

1 hour ago, neil phillips said:

Hi Lauireast  was the sharpening the same on all 4 stacks ? 

There is the problem their is little point adding more frames into a stack. and not increasing the sharpening. The whole point of adding more frames from my perspective. Is to enable more sharpening. You will just get a smoother image. That may well look less sharp. without the added sharpening. 

Edited September 24, 2022 by neil phillips

[Laurieast]

26 minutes ago, neil phillips said:

There is the problem there is little point adding more frames into a stack. and not increasing the sharpening. The whole point of adding more frames from my perspective. Is to enable more sharpening. You will just get a smoother image. That may well look less sharp. without the added sharpening. 

I started with the 50% of 3000 got the wavelets that looked best with that, and then applied those to the others with less frames. 

Not from the lowest resolution stack up, but other way round.

I suspect I'm missing, not understanding your point here somehow.

Edited September 24, 2022 by Laurieast

[neil phillips]

10 minutes ago, Laurieast said:

I started with the 50% of 3000 got the wavelets that looked best with that, and then applied those to the others with less frames. 

Not from the lowest resolution stack up, but other way round.

Ok that does help because you applied sharpening correctly for the larger stack agreed. But as a test. The problem will now be reversed. With the smaller stack becoming noisy. Where less sharpening may look more correct.  The point is sharpening will have to be proportional to stack size. If the point of the test was to show more noise will happen in a smaller stack size. then that's a success of course. But we already know that.  You can see the problems with these kinds of tests. Its why I don't bother.  Sometimes I will compare stack sizes. But once i am sure one image is looking good. I just apply sharpening by eye on that.  Variable sharpening for variable stack sizes. Starts to become a bit subjective. Thats hard to control for any kind of testing.