Maximum sequence length for Jupiter

[Ags]

Assuming 150 mm aperture, what is the maximum length of time one can image Jupiter before motion blur sets in? I vaguely recall it is 3 minutes?

[ONIKKINEN]

Im no expert but have done many (mediocre) 10-15 minute animations of Jupiter and Jupiter moves shockinly much in just that short a time so probably a bit shorter than 3min is better.

This example here spans around 15 minutes:

These are with 60s recordings. Any longer than that and derotation of the video itself would probably be necessary and that seems like a lot of work.

[CraigT82]

The maths for this one isn’t difficult, as long as you know a few key parameters.

Jupiter is about 450,000km around its circumference and rotates once every 10 hours ish. So a spot on its equator moves at about 12.5km per second. At a distance of 600,000,000km, 12.5km per second gives an angular distance travelled of 0.0043” per second. 

With a 150mm scope you’ll be sampling at 0.25”pp let’s say. 

So it would take 58 seconds for a spot on the planets equator to move from one pixel to the next, but to actually be visible in the image as blur it’d probably need to move at least 3 pixels, so there you have it…. 3 minutes! 
 

[symmetal]

I would actually say that anything over 0.5 pixel movement during the video would cause actual noticeable blurring in the stacked image. 🤔

A quick way to calculate the max video duration without worrying about angular resolution etc. is to measure the diameter in pixels of your Jupiter image. Lets say it's 300 pixels.

Circumference of Jupiter would then be 300 * PI = 942 pixels.

In 10 hrs (36,000 seconds) a spot on Jupiter's equator would therefore move 942 pixels

It will move 0.5 pixels in 0.5 / 942 x 36000 seconds = 19.1 seconds.

So it's proportional to Jupiter's image pixel diameter and the amount of pixel movement you're willing to accept.

Alan

[Ags]

I think it must be somewhere between 3 minutes and 19 seconds. I shot 3 minute sequences tonight but I will also process short better parts of the runs separately to see less time leads to better images.

[Ags]

This was three minutes....

[symmetal]

As your image is 250 pixels across Ags then 23 secs gives 0.5 pixel movement. 😊

Alan

[neil phillips]

What most people dont consider is AS/3s ability to line up a 3 min sequence with minimal rotation occurring. There will be a very slight contrast loss. or smoothing effect. Which in the scheme of things is so minimal as to be easily sorted with a slight increase in sharpening. Dont believe me  here's the proof. 

Look for the sequence where a  3 min de rotated Jupiter image is blink compared to a straight AS/3 3 min sequence. The only difference is as mentioned. so minimal. that some people are worrying about slight contrast differences. When there's a million and one other things preventing them getting good images. 

 

How long is too long on Jupiter? (planetary-astronomy-and-imaging.com)

 

[neil phillips]

3 hours ago, Ags said:

This was three minutes....

Here's also 3 min no winjupos 

 

Edited September 4, 2022 by neil phillips

[Ags]

Did a reprocess....

[symmetal]

13 hours ago, neil phillips said:

What most people dont consider is AS/3s ability to line up a 3 min sequence with minimal rotation occurring.

I admit I didn't consider AS3's effective de-rotation up to a limit by its grid alignment stacking process. 😊

Alan

[neil phillips]

35 minutes ago, symmetal said:

I admit I didn't consider AS3's effective de-rotation up to a limit by its grid alignment stacking process. 😊

Alan

Without seeing the Chris Pellier blink test. I think most wouldn't consider 3 mins as something viable. Over time I think i can see its effects. As a slight softening. But only when someone has perfected other things in planetary imaging. Would it be worth worrying about. I do prefer 5 min de rotated vids for sure. Often i run off 3 min vids and when i think ive nailed focus and everything is looking stable. i start running off 5 min vids for winjupos. It works for me. But yes that blink test is a real surprise for sure

[CraigT82]

19 hours ago, symmetal said:

I would actually say that anything over 0.5 pixel movement during the video would cause actual noticeable blurring in the stacked image. 🤔

A quick way to calculate the max video duration without worrying about angular resolution etc. is to measure the diameter in pixels of your Jupiter image. Lets say it's 300 pixels.

Circumference of Jupiter would then be 300 * PI = 942 pixels.

In 10 hrs (36,000 seconds) a spot on Jupiter's equator would therefore move 942 pixels

It will move 0.5 pixels in 0.5 / 942 x 36000 seconds = 19.1 seconds.

So it's proportional to Jupiter's image pixel diameter and the amount of pixel movement you're willing to accept.

Alan

Maybe half a pixel if captured with perfect optics, perfect collimation, perfect focus and perfect seeing, but in real conditions I disagree. 
 

I agree that it’s down to the observer and what amount of rotation they’re willing to accept though, and of course the sampling resolution.

I personally use 60 second captures with my 12” newt and they work out well. I think next time I’m out I might do some testing along the lines of that Christopher Pellier did in Neil’s link and see what happens. If anything I think AS3 is even better than AS2 at dealing with rotation, both field and planetary. 

[CraigT82]

1 hour ago, neil phillips said:

Without seeing the Chris Pellier blink test. I think most wouldn't consider 3 mins as something viable. Over time I think i can see its effects. As a slight softening. But only when someone has perfected other things in planetary imaging. Would it be worth worrying about. I do prefer 5 min de rotated vids for sure. Often i run off 3 min vids and when i think ive nailed focus and everything is looking stable. i start running off 5 min vids for winjupos. It works for me. But yes that blink test is a real surprise for sure

Looking at his 3 min test between AS2 only and winjup derotation I must admit I can’t really see any difference between the two blinked images, not on my phone anyway. The 4 minute test at the bottom of the page has obvious differences though. You don’t really see any of the specialist planetary guys running 3 min captures but, remember the OP asked about a 6 inch scope and I think 2-3 minutes will be fine for that (in my opinion).

[neil phillips]

51 minutes ago, CraigT82 said:

Looking at his 3 min test between AS2 only and winjup derotation I must admit I can’t really see any difference between the two blinked images, not on my phone anyway. The 4 minute test at the bottom of the page has obvious differences though. You don’t really see any of the specialist planetary guys running 3 min captures but, remember the OP asked about a 6 inch scope and I think 2-3 minutes will be fine for that (in my opinion).

Yes the 6" question i agree. I think the question itself is quite interesting also. My 10" i can see a slight effect. But not enough to have nightmares about. I am looking on a 30" 4k monitor. And the blink test is more obvious Craig. Just a slight softening really. Chris P says 2 mins is absolutely fine. If the serious planetary guys are not running that. Then it seems a little too purist perhaps. The good thing about running 3 min captures Craig is i can get to have a look at a capture fast. and it will tell me a lot with out any messing about. and dont forget 3 min captures can also be de rotated if someone felt it needed tightening up. Ive seen the effect video de rotation has on the features. It does de smear them enough to be a viable work flow in my opinion. Though when this planetary season is up. i could in theory chop up lots of two min captures and do image de rotation. plenty of time for that later. At the moment i am just trying to go into capture mode. not done anything special yet. But i live in hope and there is time. 

Edited September 4, 2022 by neil phillips

[CraigT82]

That’s a good point… you could just run off one long 30 min capture and chop it up as necessary in PIPP

[neil phillips]

1 minute ago, CraigT82 said:

That’s a good point… you could just run off one long 30 min capture and chop it up as necessary in PIPP

Yes though personally its a little risky putting all your eggs in one basket. i tend to refocus after about 10 15 mins to make sure its tight. often its not as tight as it could be. Though if one was certain yeah for sure chopping means you can run any amount of time. 

[knobby]

Maybe another way to approach this Is ... how many good frames do you need to stack to get a good image deep enough to sharpen ?

If for arguments sake 1500 frames gives a good enough result, why go further ?

[neil phillips]

1 hour ago, knobby said:

Maybe another way to approach this Is ... how many good frames do you need to stack to get a good image deep enough to sharpen ?

If for arguments sake 1500 frames gives a good enough result, why go further ?

That is a fair point. But in reality its not the whole story. Because the more frames captured. Enables better cherry picking. Think of it this way, lets say 4 runs are captured 2 mins each.

And only the top 25% is used on each capture. One run would be noisy despite the fact you have a sharper top 25% of the capture. Now do this 4x. using winjupos you have effectively used the same amount of frames from all of a 2 minuet capture. But now with a top 25% quality. So more frames increases the quality. Because you can cherry pick from each capture

[Mandy D]

On 03/09/2022 at 23:46, symmetal said:

I would actually say that anything over 0.5 pixel movement during the video would cause actual noticeable blurring in the stacked image. 🤔

A quick way to calculate the max video duration without worrying about angular resolution etc. is to measure the diameter in pixels of your Jupiter image. Lets say it's 300 pixels.

Circumference of Jupiter would then be 300 * PI = 942 pixels.

In 10 hrs (36,000 seconds) a spot on Jupiter's equator would therefore move 942 pixels

It will move 0.5 pixels in 0.5 / 942 x 36000 seconds = 19.1 seconds.

So it's proportional to Jupiter's image pixel diameter and the amount of pixel movement you're willing to accept.

Alan

Would it not take half of one rotation to visually move completely across Jupiter's equator, or have I missed the point, somehow?

[knobby]

I'm going to have to learn WinJupos me thinks.

[symmetal]

1 hour ago, Mandy D said:

Would it not take half of one rotation to visually move completely across Jupiter's equator, or have I missed the point, somehow?

Your statement is true, but doesn;t apply to the purpose of the discussion as to what amount of pixel movement due to planet rotation would be acceptable before image blurring is noticeable in the final stacked image. 😊

It;s equivalent to how much camera shake can you accept when taking a normal terrestrial image with a camera before the picture starts looking blurred. My saying half a pixels worth of camera shake during the exposure before noticeable blurring occurs assumes a sharp image subject to begin with.

Planetary images are blurred to begin with, due mainly to atmospheric conditions so the question is how much extra blurring due to planet rotation is acceptable. The initial figure given of around 3 pixels by CraigT82 is a reasonable figure it appears, and Autostakkert can align surface features quite well too, if they have moved a few pixels during the video.

My calculation using the image pixel size of the planet is valid, and if you input 3 pixels instead of 0.5 pixels then you get 2 mins video duration which is similar to what is commonly used. 🙂

Hope that helps Mandy D. 😃

Alan