Getting the most from a smallish scope with Jupiter

[chiltonstar]

Advice needed please!

How do I improve my 'umble attempts at Jupiter with a 180 Mak, for example during the GRS transit on 29/09/2022. I always seem to end up in a battle between noise and detail....

Chris

Edited October 5, 2022 by chiltonstar

[neil phillips]

How long are your captures, Chris. what exposure frame rate ect ? 

[chiltonstar]

I collected a sequence of 5ms/6000 frame avi's, processed each in AS, then registered and stacked 6 of these (6 mins) in Registax and PS.

Chris

[vlaiv]

I would not call 180mm smallish scope

Out of those 6000 in each run, how many did actually end up in stack? Was it always the same number?

If not - try doing whole process again but this time making sure each of those 6 stacks has same number of subs in them (say 10% or however you see fit).

If you use different number of subs in each stack - then each stack will have different SNR and just regularly stacking images with different SNR produces sub optimal results.

Best way to explain this without getting into complex math is to observe what happens with diagonal of square vs rectangle.

With square diagonal is noticeably longer than either of sides (since sides are equal we can say just one side):

But as you make more difference between two sides - diagonal tends to bring less "benefit" - or rather won't be as long compared to longer side as with square case:

In above image I added circle so you can see how small difference is between diagonal and longer side.

Alternatively - join AVI into one long movie (using PIPP) and then stack that (or maybe first derotate it in WinJupos if recording is longer than say ~5 minutes between start of first movie and end of last one).

This is better method as stacking software will then pick subs according to their quality and you might end up with 90% of subs from one movie if they are better than all the rest.

 

 

[chiltonstar]

Thanks for this @Vlav.

Each avi was treated the same - 50% used.

I tried Pipp with these files, but it was slightly less sharp. With Winjupos, good register but noisier.

Chris

[vlaiv]

8 minutes ago, chiltonstar said:

Each avi was treated the same - 50% used.

That is quite a lot, not sure where the noise is coming from then.

What camera and settings?

[chiltonstar]

ASI120 MC, gain at max ....May explain some of the noise, but I can't get down to 5ms otherwise.

Chris

[johnturley]

11 minutes ago, chiltonstar said:

ASI120 MC, gain at max ....May explain some of the noise, but I can't get down to 5ms otherwise.

Chris

That's surprising, I usually operate at around a Gain of about 300, and an exposure time of around 5ms, when imaging Jupiter with my ASI 462 through my Esprit 150.

John

Edited October 5, 2022 by johnturley

[neil phillips]

23 minutes ago, chiltonstar said:

ASI120 MC, gain at max ....May explain some of the noise, but I can't get down to 5ms otherwise.

Chris

personally, I would never max out the camera. It's likely contributing to the noise once sharpened. I am always unsure how to advise in cases like yours Because I will always find a way to keep noise down no matter the equipment. I did it recently with a 4.5" scope. I do know My 7.3 CC And 462 C camera. I can get noise free images. With an image scale far larger than the posted one. There are numerous ways to control noise. from capture and stacking procedures. To post processing choices. But your camera is less sensitive. So Getting a more sensitive camera would likely help. If your finding it difficult to manipulate everything else in the direction you want. 

Edited October 5, 2022 by neil phillips

[chiltonstar]

39 minutes ago, johnturley said:

That's surprising, I usually operate at around a Gain of about 300, and an exposure time of around 5ms, when imaging Jupiter with my ASI 462 through my Esprit 150.

John

The 180 is very slow, only f15, probably nearer to f20 if you take the adc and flip mirror into account.

Chris

[johnturley]

24 minutes ago, chiltonstar said:

The 180 is very slow, only f15, probably nearer to f20 if you take the adc and flip mirror into account.

Chris

I usually use a 2.5 x Powermate with my Esprit 150 (f7), so similar sort of focal ratio at f17.5 

John 

[chiltonstar]

10 minutes ago, johnturley said:

I usually use a 2.5 x Powermate with my Esprit 150 (f7), so similar sort of focal ratio at f17.5 

John 

Camera then maybe.

Chris

[vlaiv]

29 minutes ago, chiltonstar said:

Camera then maybe.

Chris

Could be down to processing as well.

Maybe try to limit how much you push the data - find that place where noise starts to show too much and then back off a bit?

You could also post raw stack so people can process it with their workflow and confirm if data is really noisy or maybe it is just processing.

[CraigT82]

The 462 is more sensitive than the 120, though if you take the pixel size difference into account probably not that much of a gap. Maybe transparency was poor for that capture? 
 

Edited October 6, 2022 by CraigT82

[chiltonstar]

17 hours ago, vlaiv said:

Could be down to processing as well.

Maybe try to limit how much you push the data - find that place where noise starts to show too much and then back off a bit?

You could also post raw stack so people can process it with their workflow and confirm if data is really noisy or maybe it is just processing.

Good idea - maybe me just overdoing it a bit on a file that isn't sharp enough. This is the stacked, unsharpened file.

Chris

Edited October 6, 2022 by chiltonstar

[vlaiv]

1 hour ago, chiltonstar said:

Good idea - maybe me just overdoing it a bit on a file that isn't sharp enough. This is the stacked, unsharpened file.

Chris

First of all - I think your processing skills are extraordinary.

I've never worked with data that behaves like this and with standard tools - I can't get anywhere near what you got.

As if all the sliders run out of space - I want to sharpen more but there is only so much a slider will go to the right

My recommendation would be to restack the data, but this time using less frames from each recording.

Here is what I believe is happening.

Number of stacked subs is a tradeoff.

Stack too few subs and you will have noisy result. Stack many subs and you will have good SNR

Stack too few subs and you'll be able to sharpen result properly (as these subs are already not too blurred - high quality), stack too many results and no amount of normal sharpening will get you there.

This means that there is "tipping point" - once you stack more subs than this point, you enter a "race" - add more (low quality) subs - can sharpen more because of better SNR but also need to sharpen even more because you added blurred subs.

And if you try to sharpen enough you will certainly hit the noise. I think your data is past this tipping point and this is the reason for the noise - if you want it to look as sharp as in your image - you need crazy amount of sharpening and SNR is simply not up for it.

Try using less but more quality subs. AS!3 has 4 fields where you can enter percentage - try to make each movie with range of top percentages used - like 10, 20, 30, 40 and see what gives you best results.

Also - try doing what I've suggested - joining video in one and performing derotation (if you have significant time gap between any two runs - derotate to same time each movie individually before joining, else join and derotate) and then stack.

This will select good frames across the whole run - not only in each segment (it is a bit like taking a top athlete from each country and declaring that these are best athletes in the world in a given sport - it might not be true, one country might have 3 such athletes while some countries would have athletes that would otherwise not qualify - but are still the best in their country).

[Ian McCallum]

2 hours ago, chiltonstar said:

Good idea - maybe me just overdoing it a bit on a file that isn't sharp enough. This is the stacked, unsharpened file.

Chris

I just re sharpened your original, using RegiStax.

[Laurieast]

@chiltonstar Did you have sharpened on in AS3 ? 

 

Edited October 6, 2022 by Laurieast

[chiltonstar]

1 hour ago, Laurieast said:

@chiltonstar Did you have sharpened on in AS3 ? 

 

Yes, but I used the non-sharpened file (not the .conv)

Chris

[neil phillips]

Either seeing was bad. Or your slightly out of focus, or both?  I think Chris ?

As Vlaiv said it doesn't respond as a good tight capture should. Probably for reasons mentioned. No wonder you're getting noise when you're hitting it with such extreme sharpening. Which is evident in the detail you have managed to extract from a blurry a capture. And how unnatural and spotty its appearance is texture wise.

Solution better seeing. Better capture. And or more in focus.

Good quality captures actually require very little sharpening. And with more sharpening a wealth of detail comes out, with minimal noise. Thats my current thoughts from the raw you uploaded. Unless I am missing something? 

[chiltonstar]

1 hour ago, neil phillips said:

Either seeing was bad. Or your slightly out of focus, or both?  I think Chris ?

As Vlaiv said it doesn't respond as a good tight capture should. Probably for reasons mentioned. No wonder you're getting noise when you're hitting it with such extreme sharpening. Which is evident in the detail you have managed to extract from a blurry a capture. And how unnatural and spotty its appearance is texture wise.

Solution better seeing. Better capture. And or more in focus.

Good quality captures actually require very little sharpening. And with more sharpening a wealth of detail comes out, with minimal noise. Thats my current thoughts from the raw you uploaded. Unless I am missing something? 

Thanks for this Neil. I think the focus was good as detail comes out with over-sharpening like the eye of the GRS, but the seeing was "unexceptional". That was the point of the post really, how to get the best out of a capture in less than (ie typical) conditions.

Chris

[neil phillips]

9 minutes ago, chiltonstar said:

Thanks for this Neil. I think the focus was good as detail comes out with over-sharpening like the eye of the GRS, but the seeing was "unexceptional". That was the point of the post really, how to get the best out of a capture in less than (ie typical) conditions.

Chris

With seeing that bad.  (Because you say focus was good ) There would Only be so much you can do with those types of captures.  You probably pulled out as much as is possible. But trying to do so without unwanted artifacts is a bit unlikely unfortunately. 

  If you did put a alignment box around the moon? when you processed the capture (Doesn't look like you did)

then I still think it has all the hall marks of being slightly out of focus.  Big and bloated appearance. Not tight and defined. If I was capturing the GRS and I slightly de tuned the focus, lots of sharpening would still show the eye of the GRS. Thats not a great way of determining focus Chris. I could test this with my CC

Interestingly I remember this night and at my location (locations can still vary mind) seeing wasn't great. But wasn't awful. If you redo the capture, put a box around the moon and if the moon is still bloated like that. focus should be considered When I did wavelets on your image it didn't respond as I would expect. I Dunno ?

 

[chiltonstar]

Thanks for all the advice.

Trying a few days later with better seeing certainly gave a better result and I was less tempted to over-sharpen. Interestingly, I found that some of the artefacts in the original image were introduced by rotating the image in PS Elements, particularly any sharpening operation after rotation. It is awkward for me to use the camera in the correct orientation as the cable pokes up under my chin using the flip mirror. I need to find a way round that. This is the new image (unrotated ) on the 6th Oct.

Chris

 

Edited October 8, 2022 by chiltonstar