Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

For Planetary: one long capture OR multiple shorter ones?

  • Please log in to reply
24 replies to this topic

#1 JayMo7

JayMo7

    Sputnik

  • -----
  • topic starter
  • Posts: 33
  • Joined: 05 Aug 2020
  • Loc: Montreal, Canada

Posted 10 August 2020 - 05:47 PM

The past few days I've been working with PIPP, AutoStakkert and RegiStax on the Jupiter data capture video that Steve (WarmWeatherGuy) provided me. This was a first for me with these softwares so my initial learning curve has been steep, but now I can find my way around for the basic functions.

 

PIPP is much more than a file converter. It can centre the object of interest in each frame, change the aspect ratio, extract portion of a file, etc. Not too difficult to use and handy!

 

I found AutoStakkert more difficult to use. I watched tutorials especially the ones by Steve available at http://planetaryimagingtutorials.com/. I kept it simple and finally made it to the end.

 

I found RegiStax even more difficult and not particularly user friendly for someone new to astro imaging. I watched even more tutorials than for AutoStakkert to the point of suffering from acute Information Overload! So again I kept it simple and made it to the end. WOW! I was blown away by the finished image. Either it was a beginner's luck or the raw data was so good that even a beginner could do it or AS and RS are so well automated that they compensate for the user's missteps or...

 

Steve's (avi) video is 120 seconds long for 12 000 frames. I decided to use PIPP to extract the first 20 sec (2000 frames). I captured a frame from the original 120 sec file and one from the 20 sec extract file, located à 00:00:10 in both cases. When viewed side by side they looked identical, switching from one to the other, full-screen image didn't change. PIPP really extracted the exact same frames.

 

So I processed the 20 sec file with AS and RS using the same settings I used for the 120 sec original file and compared both finished image. I could not visually see a difference. That surprised me because I thought that capturing 120 sec of raw data would deliver a visibly better signal/noise ratio than a 20 sec capture. If it isn't delivering a better finished image than why would people invest serious money to buy a gem with rock solid tracking to make longer planetary captures...

 

Am I missing something?


  • daringly likes this

#2 KiwiRay

KiwiRay

    Apollo

  • -----
  • Posts: 1,083
  • Joined: 09 Apr 2015
  • Loc: Seattle, USA

Posted 10 August 2020 - 06:05 PM

One comment: no one buys a "gem with rock solid tracking" to make longer planetary captures.  The stacking software centres each frame, and can deal with planets wandering all over the region of interest.  That's why inexpensive alt-az mounts are fine for planetary, and some people even track by hand.  Your main question is valid, though - I also would expect better signal to noise with more frames.


Edited by KiwiRay, 10 August 2020 - 06:57 PM.

  • daringly likes this

#3 ButterFly

ButterFly

    Apollo

  • *****
  • Posts: 1,231
  • Joined: 07 Sep 2018

Posted 10 August 2020 - 06:17 PM

Jupiter rotates very fast.  Capture as many frames as you can in the shortest amount of time.  Under 90s is a good rule of thumb.  If you can capture 10K in under 60s, even better.



#4 KiwiRay

KiwiRay

    Apollo

  • -----
  • Posts: 1,083
  • Joined: 09 Apr 2015
  • Loc: Seattle, USA

Posted 10 August 2020 - 06:19 PM

Jupiter rotates very fast.  Capture as many frames as you can in the shortest amount of time.  Under 90s is a good rule of thumb.  If you can capture 10K in under 60s, even better.

A myth, often repeated, but disproved on a daily basis by the people posting quality Jupiter images captured using videos of up to three minutes. 


  • eros312, RedLionNJ, Lacaille and 2 others like this

#5 CPellier

CPellier

    Apollo

  • -----
  • Posts: 1,484
  • Joined: 07 Aug 2010

Posted 11 August 2020 - 03:06 AM

The full story is that today's modern softwares like Autostakkert can partially correct the rotation of planetary globes thanks to the multipoint alignment. However it has a limit, and even if no softening is visible on a 3 mn video Jupiter image, this does not mean that there is not any. The only way to be sure that no problem will occur is to use the derotations methods of WinJupos, either by derotating several short videos (of less than 3 mn!) or one longer of at least 5 to 6 mn.

Here is one example I made many years ago, this is an animation of a 3 mn video of Jupiter processed with Autostakkert, with and without derotation. Although both results are good, the one without derotation is slightly soft. You would not see the softening on this frame alone.

juprotate3mnWJAS.gif

 

Full article here : how long is too long on Jupiter?


  • eros312, roelb, DMach and 1 other like this

#6 JayMo7

JayMo7

    Sputnik

  • -----
  • topic starter
  • Posts: 33
  • Joined: 05 Aug 2020
  • Loc: Montreal, Canada

Posted 11 August 2020 - 08:09 AM

Merci M. Pellier. Je viens de lire l'article sur votre site. Intéressant et informatif. Je vais parcourir tout le site.

 

From what I can see Jupiter is somewhat special if not just because of it relatively fast rotation.

 

My original question, "For Planetary: one long capture or multiple shorter ones?" was more general and intended for any planetary objects.

 

For example: normally, would 1 capture of (say) 120 sec of a planetary object typically deliver a visibly better image than grouping 6 captures of 20 sec each of the same object from the same equipment under the same conditions?


  • CPellier likes this

#7 Dunkstar

Dunkstar

    Viking 1

  • -----
  • Posts: 671
  • Joined: 26 Mar 2012
  • Loc: Brisbane, Australia

Posted 11 August 2020 - 08:30 AM

If you take one long capture, you can always split it into smaller chunks with a tool such as PIPP.

 

I tend to keep my captures below 60 seconds, but that’s just practice from using a mono camera.


  • CPellier likes this

#8 CPellier

CPellier

    Apollo

  • -----
  • Posts: 1,484
  • Joined: 07 Aug 2010

Posted 11 August 2020 - 09:14 AM

I like the answer from Dunkstar, which is probably the best, just take one long video (6 mn at least) and split it with PIPP.

I'm personnally into very long video but usually observers don't bother wait for such a long time to get the image (it can take long to WinJupos). Derotating a few images from short videos works very well too.


  • KTAZ likes this

#9 KTAZ

KTAZ

    Viking 1

  • *****
  • Posts: 892
  • Joined: 09 Apr 2020
  • Loc: Scottsdale, AZ

Posted 11 August 2020 - 09:23 AM

I like the answer from Dunkstar, which is probably the best, just take one long video (6 mn at least) and split it with PIPP.

I'm personnally into very long video but usually observers don't bother wait for such a long time to get the image (it can take long to WinJupos). Derotating a few images from short videos works very well too.

I’ve read your article previously. Thanks for that effort!



#10 JayMo7

JayMo7

    Sputnik

  • -----
  • topic starter
  • Posts: 33
  • Joined: 05 Aug 2020
  • Loc: Montreal, Canada

Posted 11 August 2020 - 09:35 AM

If you take one long capture, you can always split it into smaller chunks with a tool such as PIPP.

 

I tend to keep my captures below 60 seconds, but that’s just practice from using a mono camera.

 

I like the answer from Dunkstar, which is probably the best, just take one long video (6 mn at least) and split it with PIPP.

I'm personnally into very long video but usually observers don't bother wait for such a long time to get the image (it can take long to WinJupos). Derotating a few images from short videos works very well too.

Dunkstar and CPellier :

I don't understand what value will be gained from splitting with PIPP a longer video capture into smaller chunks. Will it bring in any more pixel information than what's in the longer capture?  I doubt it would.

On the other hand, if someone's equipment can not deliver quality capture of longer than (say) 20 sec because of being on a alt-az mount, would 6 consecutive 20 sec quality captures stacked together at processing time deliver a quality image equivalent to a single 120 sec quality capture from the same tube on a gem, under the same conditions?



#11 CPellier

CPellier

    Apollo

  • -----
  • Posts: 1,484
  • Joined: 07 Aug 2010

Posted 11 August 2020 - 10:39 AM

Dunkstar and CPellier :

I don't understand what value will be gained from splitting with PIPP a longer video capture into smaller chunks. Will it bring in any more pixel information than what's in the longer capture?  I doubt it would.

On the other hand, if someone's equipment can not deliver quality capture of longer than (say) 20 sec because of being on a alt-az mount, would 6 consecutive 20 sec quality captures stacked together at processing time deliver a quality image equivalent to a single 120 sec quality capture from the same tube on a gem, under the same conditions?

Oh it's not to gain anything: the purpose was to have both solutions available, if you take a very long video, you can split it if you wish too.

Thanks to WinJupos, you can make very long video even if you have an alt-az mount (this is my case! I'm using a Flextube 305 goto). I have made videos up to 30 mn (CH4 imaging) without any problem. But you must use Winjupos.


  • JayMo7 likes this

#12 JayMo7

JayMo7

    Sputnik

  • -----
  • topic starter
  • Posts: 33
  • Joined: 05 Aug 2020
  • Loc: Montreal, Canada

Posted 11 August 2020 - 10:57 AM

Oh it's not to gain anything: the purpose was to have both solutions available, if you take a very long video, you can split it if you wish too.

Thanks to WinJupos, you can make very long video even if you have an alt-az mount (this is my case! I'm using a Flextube 305 goto). I have made videos up to 30 mn (CH4 imaging) without any problem. But you must use Winjupos.

That I understand!  wink.gif

Thanks 

 

PS-  CPellier: I just downloaded Winjupos and watched your tutorial on astrosurf.com on it and noticed that Winjupos is for all planets including the Moon, for derotating but also for correcting the FOV rotation from alt-az mounts. VERY INTERESTING... BTW does it give very good results for detailed surfaces like on the Moon ?


Edited by JayMo7, 11 August 2020 - 11:59 AM.


#13 ToxMan

ToxMan

    Skylab

  • *****
  • Posts: 4,212
  • Joined: 23 Jan 2011
  • Loc: Tucson, Arizona, USA

Posted 11 August 2020 - 03:52 PM

I take a "blended" approach with Jupiter using monochrome camera and filters.

1 minute per filter...accounting for time between to switch filters and re-focus, about 15 seconds, it is 3 and half minutes to collect RGB data. It is not long enough "smear" the data, and if I want to go longer, say 90 second captures, WinJUPOS can de-rotate.

The limit to the length of the video is really the file size and abilities to process in post.

It's been a long time, Astrovienna showed it was possible to de-rotate long video without having to chop it up.

In addition, Darryl has shown it is possible to combine multiple stills from multiple videos, in a what I would dub, "a super de-rotation scheme." I think he covered a 22 minute spread into a single image.

Edited by ToxMan, 11 August 2020 - 04:00 PM.


#14 ButterFly

ButterFly

    Apollo

  • *****
  • Posts: 1,231
  • Joined: 07 Sep 2018

Posted 11 August 2020 - 05:39 PM

Of course, you don't have to use all the frames from that 6 min long video either, or even the 20s one.  Try comparing various 20s extracts of the original video to each other to see the effects of seeing variations.  Both the software you mentioned allows you to choose the best X percent or the best Y frames from whatever you input.  I haven't seen a mention of how many frames from each you are actually comparing.  If you just happened to compared a really good 20s sample to the entire 120s, it's not surprising.



#15 JayMo7

JayMo7

    Sputnik

  • -----
  • topic starter
  • Posts: 33
  • Joined: 05 Aug 2020
  • Loc: Montreal, Canada

Posted 11 August 2020 - 07:12 PM

(...) I haven't seen a mention of how many frames from each you are actually comparing.  If you just happened to compared a really good 20s sample to the entire 120s, it's not surprising.

Both the 120 sec (12000 total frames) and the 20 sec (2000 total frames) videos were processed through AutoStackkert and RegiStax with exactly the same settings, including a 25% frame percentage to stack option.



#16 Dunkstar

Dunkstar

    Viking 1

  • -----
  • Posts: 671
  • Joined: 26 Mar 2012
  • Loc: Brisbane, Australia

Posted 11 August 2020 - 10:27 PM

A lot depends on how good or bad the seeing was.

 

If the seeing was good, I’d expect to see better results from a stack of 3000 frames vs 500 frames. BUT you would both be able to AND would need to push the 3000 stack harder to get more out of it. With better SNR you should expect to see less noise spoiling the image when pushing harder.

 

Conversely if the seeing was poor, it’s possible that there are only 10 or so good frames per capture and beyond that you are diluting the quality of the result with all of the other so-so frames.

 

Back on the subject of capture length...when I’ve taken shorter captures with a short pause in between, I’ve often wondered “what if?”, meaning what if there were darn good frames in that time and I missed them? With my OSC I’ve started gunning for 10,000 frames at a time, which works out at about 60 seconds for me, but I might try longer...



#17 ButterFly

ButterFly

    Apollo

  • *****
  • Posts: 1,231
  • Joined: 07 Sep 2018

Posted 12 August 2020 - 12:01 AM

Both the 120 sec (12000 total frames) and the 20 sec (2000 total frames) videos were processed through AutoStackkert and RegiStax with exactly the same settings, including a 25% frame percentage to stack option.

Try another 20s subset with 25% as well.  Compare that to the original 20s subset.  The best 25% during a time of bad seeing is worse than the best 25% during a time of good seeing.  A blanket percentage of frames may not be the best comparison between a subset and the whole thing when the subset just happened to have very good seeing.  If you find other 20s subsets that give you good seeing, by all means, combine those lucky subsets together.


Edited by ButterFly, 12 August 2020 - 12:01 AM.


#18 CPellier

CPellier

    Apollo

  • -----
  • Posts: 1,484
  • Joined: 07 Aug 2010

Posted 12 August 2020 - 04:10 AM

That I understand!  wink.gif

Thanks 

 

PS-  CPellier: I just downloaded Winjupos and watched your tutorial on astrosurf.com on it and noticed that Winjupos is for all planets including the Moon, for derotating but also for correcting the FOV rotation from alt-az mounts. VERY INTERESTING... BTW does it give very good results for detailed surfaces like on the Moon ?

Hi, I have never tried to derotate images of the Moon. It looks to be possible... but as I do not image the Moon, I can't tell you whether or not it is interesting to do so ! 

Except maybe for correcting the altaz field rotation. It looks simple to do it by image derotation. However it must be extremely complicated to use if the field is high res and does not contain either the limb or the terminator...



#19 Kokatha man

Kokatha man

    Voyager 1

  • *****
  • Posts: 14,492
  • Joined: 13 Sep 2009
  • Loc: "cooker-ta man" downunda...

Posted 12 August 2020 - 04:29 AM

I take a "blended" approach with Jupiter using monochrome camera and filters.

1 minute per filter...accounting for time between to switch filters and re-focus, about 15 seconds, it is 3 and half minutes to collect RGB data. It is not long enough "smear" the data, and if I want to go longer, say 90 second captures, WinJUPOS can de-rotate.

The limit to the length of the video is really the file size and abilities to process in post.

It's been a long time, Astrovienna showed it was possible to de-rotate long video without having to chop it up.

In addition, Darryl has shown it is possible to combine multiple stills from multiple videos, in a what I would dub, "a super de-rotation scheme." I think he covered a 22 minute spread into a single image.

...I think you're referring to this Paul...actually the longest was nearly 46 minutes. wink.gif

 

https://momilika.net...2-Square.png      Image can be clicked on for full-sizing...


  • eros312 likes this

#20 ToxMan

ToxMan

    Skylab

  • *****
  • Posts: 4,212
  • Joined: 23 Jan 2011
  • Loc: Tucson, Arizona, USA

Posted 12 August 2020 - 02:26 PM

Indeed, Darryl. Forty six minutes proves the point even better! Thanks.

#21 Achernar

Achernar

    Voyager 1

  • *****
  • Posts: 11,714
  • Joined: 25 Feb 2006
  • Loc: Mobile, Alabama, USA

Posted 12 August 2020 - 03:44 PM

Jupiter and Saturn rotate very rapidly, and if you record more than 90 seconds or two minutes rotational smear starts to intrude into the picture.

 

Taras



#22 JayMo7

JayMo7

    Sputnik

  • -----
  • topic starter
  • Posts: 33
  • Joined: 05 Aug 2020
  • Loc: Montreal, Canada

Posted 12 August 2020 - 04:34 PM

Hi, I have never tried to derotate images of the Moon. It looks to be possible... but as I do not image the Moon, I can't tell you whether or not it is interesting to do so ! 

Except maybe for correcting the altaz field rotation. It looks simple to do it by image derotation. However it must be extremely complicated to use if the field is high res and does not contain either the limb or the terminator...

Thanks Christophe.

 

I just read your article "Planetary imaging with an altazimuth Dobsonian (2018)" where you convincingly demonstrate that WinJupos with AutoStakkert will efficiently handle both object de-rotation and field rotation correction. The process is as simple as selecting the De-rotation of a video stream function in WinJupos and then just check the box Compensation of field rotation in altitude over azimuth mounts... and you're done! The resulting video is then stacked with AutoStakkert. You also mentioned in the article that, depending on the planet, AS alone would do a good job of correcting the field rotation on short videos of < 5 min.

 

I still wonder if WinJupos would do a good job of correcting the field rotation on a Moon time-lapse video long enough to show field rotation caused by an alt-az mount, when the limb and/or the terminator are visible on the frames...


Edited by JayMo7, 12 August 2020 - 04:35 PM.


#23 JayMo7

JayMo7

    Sputnik

  • -----
  • topic starter
  • Posts: 33
  • Joined: 05 Aug 2020
  • Loc: Montreal, Canada

Posted 14 August 2020 - 02:48 PM

...I think you're referring to this Paul...actually the longest was nearly 46 minutes. wink.gif

 

https://momilika.net...2-Square.png      Image can be clicked on for full-sizing...

I've just spent half an our browsing through the site momilika.net... WOW!  that's something different from what I read the past few months. Looking at astrophotography from such an artistic angle is refreshing and inspiring. It will give me hours of interesting reading...

 

BTW, are you Darryl or Mo or Nicholas and what is the meaning of Kokatha man?



#24 Kokatha man

Kokatha man

    Voyager 1

  • *****
  • Posts: 14,492
  • Joined: 13 Sep 2009
  • Loc: "cooker-ta man" downunda...

Posted 14 August 2020 - 08:36 PM

 

BTW, are you Darryl or Mo or Nicholas and what is the meaning of Kokatha man?

lol.gif I'd probably answer to any polite call...I'm Darryl or Mo to most people. Nicholas is the surname of my partner Pattie - & I'm a proud Kokatha man, part of the Western Desert "First Nations" people...like some of those well-known groups in places like Quebec, such as the Algonquin & Cree etc. wink.gif

 

I had the pleasure of meeting with a group of people from the West Coast (B.C.) of your country many years ago when they came here... smile.gif


  • eros312 and KiwiRay like this

#25 astrovienna

astrovienna

    Gemini

  • *****
  • Posts: 3,016
  • Joined: 04 Dec 2006
  • Loc: The NoVa White Zone

Posted 16 August 2020 - 08:45 AM

I take a "blended" approach with Jupiter using monochrome camera and filters.

1 minute per filter...accounting for time between to switch filters and re-focus, about 15 seconds, it is 3 and half minutes to collect RGB data. It is not long enough "smear" the data, and if I want to go longer, say 90 second captures, WinJUPOS can de-rotate.

The limit to the length of the video is really the file size and abilities to process in post.

It's been a long time, Astrovienna showed it was possible to de-rotate long video without having to chop it up.

In addition, Darryl has shown it is possible to combine multiple stills from multiple videos, in a what I would dub, "a super de-rotation scheme." I think he covered a 22 minute spread into a single image.

Hello Tox and all!  Yes, it's been a while, but I'm back now.  The planets all went into the trees from my obs a few years ago, but a new neighbor took a chainsaw to all his trees, and now I'm getting back up to speed.  I'll have plenty of questions.  I've forgotten a lot!

 

Kevin




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics