AutoStakkert -what do the boxes mean?

[Merlin66 (Ken)]

I've been using AS3! for the past few years and always wondered why there were four selection (# frames or % age frames) boxes?
Did you know??

Well it appears that it is to allow you to select up to four different stacking options at the one time!
Nothing to do with the 25% lines shown on the quality graph.

Now I just have to find out what moving the green line cursor does on the quality graph.....

[Startrek (Martin)]

Ken
I posted the same question a month ago
Definitely not a straightforward tool , the quality graph
It’s a bit vague isn’t it

[Merlin66 (Ken)]

Martin,
Know I understand that each box is independent that certainly makes things a bit easier to understand.
I had assumed that dragging the green cursor across the quality graph that it would accept the position as a selection.....this doesn’t seem to be the case.
Onwards and Upwards.

[Startrek (Martin)]

The method I’ve been adopting lately is where the graph intersects the 50% or halfway line, I select the green line at this point and use that for my quality batch of frames (which essentially the best 30 to 40 % I assume )
It’s usually around 30 to 40% and only places this value in the left hand box , the other 3 boxes are left blank
I saw this on a few YouTube clips
It’s been working ok for me lately
I’ve tried loading 4 boxes with various percentages but don’t see any difference
You have experiment with this program

[Merlin66 (Ken)]

Different boxes and different %ages give a series of different images.
I’ve tried 5% v’s 10% and up to 50% settings.
When I compare the images produced I’m not sure there is a great difference.

[sil (Steve)]

Originally Posted by Merlin66

Different boxes and different %ages give a series of different images.
I’ve tried 5% v’s 10% and up to 50% settings.
When I compare the images produced I’m not sure there is a great difference.

yes I find the same usually. But yes thats what the boxes are for, so you can create a number of images using different quality constraints in one go. I don't think my captures are close enough with enough detail for differences to become obvious. I think my focus is out more than the difference between best 5% vs 50% of frames.

I don't process further since AS! uses wavelets for sharpening already so I don't like pushing it hard with more in another program. However many people do and many have better gear than me (better magnification and more detail captures etc) so I suggest you test for yourself.

Test by setting 25 and 75% as output stacks as TIF files. Takes these two saved files into your photo software of choice and load one on top of the other as a separate layer in the image (which one doesn't matter). Now you can do a difference between them, a method to do the same is to Invert the colours of the top layer image and then set its opacity to 50%. Now flatten the image, so its one layer with the changes above applied. It should look basically an even grey at this point. you can play with levels and contrast to exagerrate any image in the greyness. this image is the difference between the two source images. if the two are identical you wont see anything in the grey image, its all grey (test by doing this process with any image duplicating it to a second layer.)

However if you are seeing the outline of your planet or distinct rings for saturn or bands on jupiter or and surface detail at all then this represents signal that is lost (ie detail) between the 25% and 75% stacks. It may be sharp or it may be faint/indistinct, or not there at all. But its an indication there IS a real difference between using the 25% and 75% stack options and use that to inform your workflow decisions.

you can also try running frames through winjupos to derotate then stack source and derotated separately to see if there is a detail difference to make it worth bothering with.

[Stonius (Markus)]

I often find no difference between 50 & 10%, but a marked improvement with 5%, or even 2%.

[sil (Steve)]

Originally Posted by Stonius

I often find no difference between 50 & 10%, but a marked improvement with 5%, or even 2%.

yep, if you use the image as is from AS! it's hard to pick a difference. but if you further process the differences both good (signal) and bad (noise) will be made more obvious. 2% should be noticeable as you are only stacking the few very best quality frames. I'd advise to use a % that gives you at least 100 frames to combine however so at least noise is well suppressed as are any possible blocks present in your source file from the file type and/or any compression present. That will at least give you a file you can adjust contrast and saturation a bit on to taste.

everybody captures different durations so advice on percentage isn't appropriate but i think aiming for a minimum of 100 frames in the stack is a good guide. just my opinion, i havent tested but please feel free to disagree.

part of my process is to run my planetary SER captures through PIPP first to crop to about 400x400 pixels centering the planet and exporting to SER about 2,000-5,000 frames (my captures often are above 10k+ frames) and ordered by quality. My thinking here is since i am using an AltAz rig with dubious tracking the planet drifts in my source captures and likely rolls too slightly. So centering in pipp makes up for bad or no tracking. order by quality should kill off the really bad frames. now the output file i drop into autostakkerthas the frames well aligned making it easier for AS! to fine tune with its alignment and its quality method on top should be using the most perfect frames to work with. So hopefully it makes less errors with picking alignment points and gives me the very very best result possible with my source capture. Until AS4! comes along which blows our minds.