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

Why Do People Use DeepSkyStacker and Then PixInsight?

Astrophotography
  • Please log in to reply
118 replies to this topic

#1 Jim Waters

Jim Waters

    Fly Me to the Moon

  • *****
  • topic starter
  • Posts: 5,854
  • Joined: 21 Oct 2007
  • Loc: Phoenix, AZ USA

Posted 21 June 2021 - 04:06 PM

As the title says.  Why do people use DeepSkyStacker for calibration, aligning and stacking then switch to PixInsight?


Edited by Jim Waters, 21 June 2021 - 04:07 PM.

  • dswtan, elmiko, Greg F and 1 other like this

#2 Maritime

Maritime

    Apollo

  • *****
  • Posts: 1,014
  • Joined: 19 Aug 2020

Posted 21 June 2021 - 04:19 PM

Better yet, why pixinsight?  I’d rather take a beating in a dark alley….


  • happylimpet, Dynan and TelescopeGreg like this

#3 D_talley

D_talley

    Soyuz

  • *****
  • Posts: 3,997
  • Joined: 07 Jul 2005
  • Loc: Albuquerque, New Mexico

Posted 21 June 2021 - 04:25 PM

Deepskystacker is only used to stack photos.  You can't process the image with DSS.  Pixinsight will be the next step needed to process your image.  Pixinsight is easy to use after watching the tutorials and a little practice.  I love using the two programs to process my images. 

 

Yes Pixinsight can stack but DSS is easier to use for that. 


  • John Rogers likes this

#4 Dynan

Dynan

    Fly Me to the Moon

  • *****
  • Posts: 6,001
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 21 June 2021 - 04:32 PM

...and not APP then PI or ST. lol.gif


Edited by Dynan, 21 June 2021 - 04:42 PM.

  • leviathan likes this

#5 spacemunkee

spacemunkee

    Fly Me to the Moon

  • *****
  • Posts: 5,739
  • Joined: 22 Oct 2017
  • Loc: Cincinnati, Ohio

Posted 21 June 2021 - 04:37 PM

Learned how to integrate with PI not too long ago. Boy, that's a chore! Had to write up a three page cheat sheet to follow!



#6 bobzeq25

bobzeq25

    ISS

  • *****
  • Posts: 24,984
  • Joined: 27 Oct 2014

Posted 21 June 2021 - 04:48 PM

For no good reason.  Calibrating and stacking is one of the easier things to do in PI, if you're going to process you might as well do those operations in PI also.  Among other things it avoids possible errors going between two completely different programs.

 

PI is not for everyone, but using it _only_ after stacking in DSS makes little sense.


Edited by bobzeq25, 21 June 2021 - 04:51 PM.

  • mikefulb, dswtan, Jim Waters and 3 others like this

#7 Dynan

Dynan

    Fly Me to the Moon

  • *****
  • Posts: 6,001
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 21 June 2021 - 04:50 PM

Better yet, why pixinsight?  I’d rather take a beating in a dark alley….

funnypost.gif lol.gif lol.gif lol.gif 

 

Akin to "Why are you hitting your thumb with a hammer?"..."Because it feels so good when I STOP."

 

I learned PI from Keller's book to a decent degree, as recommended by Bobzeq, and I still prefer Star Tools on good data.


Edited by Dynan, 21 June 2021 - 04:52 PM.

  • Readerp likes this

#8 AbsolutelyNot

AbsolutelyNot

    Lift Off

  • -----
  • Posts: 2
  • Joined: 17 May 2021

Posted 21 June 2021 - 04:53 PM

From my (extremely limited) experience PI generates a large number of large files during the process of stacking, etc.  I use Blink in PI to delete unsuitable subs and then quickly stack them in DSS.  If I shoot on later nights I can simply add the new subs on a new tab.  Easy-peasy and no file bloat.



#9 TelescopeGreg

TelescopeGreg

    Gemini

  • -----
  • Posts: 3,220
  • Joined: 16 Jul 2018
  • Loc: Auburn, California, USA

Posted 21 June 2021 - 04:58 PM

.... for the same reason I don't use PI at all.  Ease.  I want the software to work for me, not the other way around.  I use either DSS or ASTAP for stacking, the StarTools for the rest.

 

PI reminds me of the PBS TV show, The Woodwright's Shop.  Want to build a table?  Grab that axe over there and follow me to the forest...



#10 kraegar

kraegar

    Apollo

  • *****
  • Posts: 1,480
  • Joined: 17 Apr 2012

Posted 21 June 2021 - 05:07 PM

PI's stacking is more powerful and can produce much better results, but a lot of people find the multiple steps slow and daunting.

 

I think WBPP2.x alleviates a lot of that - I get great  stacking and rarely need to tweak anything with it (though I typically do a manual integration after so I can use SFS and/or the new NSG Script before integration)

 

PI, overall, adds a ton of flexibility - which is super powerful, but at the cost of not having a simple, clear path.

 

My guess is if someone made a form of WBPP2.x with the exact  same options, but took away all the extra fields for adjusting and stripped it to the bare minimum (so it looked like DSS more or less), and scripted it to clean up the interim files, a lot of people who use DSS would use that instead... and get better results.  It would be just like running with the defaults... but simpler to follow.


  • mikefulb, Jim Waters, 17.5Dob and 1 other like this

#11 terry59

terry59

    Voyager 1

  • *****
  • Posts: 10,457
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 21 June 2021 - 05:09 PM

I expect the answer is generally going to be that the PI calibration processes are intimidating, especially given that PI is routinely advised against around here for new people. Using the WBPP script I get very good results with both CCD and CMOS at default settings except for adding a pedestal to the CMOS data and selecting the correct file for dark optimization, for CCD it is the flats and for CMOS it is the lights

 

Anyone that decides to get "serious" about AP will learn how and what settings to use for calibrating data anyway

 

smile.gif


Edited by terry59, 21 June 2021 - 05:32 PM.

  • elmiko and Dynan like this

#12 Jim Waters

Jim Waters

    Fly Me to the Moon

  • *****
  • topic starter
  • Posts: 5,854
  • Joined: 21 Oct 2007
  • Loc: Phoenix, AZ USA

Posted 21 June 2021 - 05:22 PM

For no good reason.  Calibrating and stacking is one of the easier things to do in PI, if you're going to process you might as well do those operations in PI also.  Among other things it avoids possible errors going between two completely different programs.

 

PI is not for everyone, but using it _only_ after stacking in DSS makes little sense.

Exactly.  I learned the basics of PI in a weekend and using Batch Preprocessing (BPP) or the newer Weighted Batch Preprocessing (WBPP) makes that MUCH simpler.  People that have issues with PI need to spend a few hours viewing the video's on YouTube. 

 

When I tried DSS many years ago the results were always poor at best.  I don't understand why people use DSS for calibration, ...stacking and then jump to PI.


Edited by Jim Waters, 21 June 2021 - 05:25 PM.

  • dswtan, bobzeq25 and Bretw01 like this

#13 John Rogers

John Rogers

    Viking 1

  • *****
  • Posts: 830
  • Joined: 08 Feb 2008

Posted 21 June 2021 - 05:57 PM

PixInsight is one of the tools in my toolbox.  As with most software, they each have their own strengths and weaknesses.  As users, we discover what works for us.

 

My current set of image processing tools include AutoStakkert!, DeepSkyStacker, PIPP, Registax6, PixInsight and Photoshop 2021 (with Noel Carboni's Astronomy Tools Action Set)



#14 kathyastro

kathyastro

    Gemini

  • *****
  • Posts: 3,280
  • Joined: 23 Dec 2016
  • Loc: Nova Scotia

Posted 21 June 2021 - 06:30 PM

People often upgrade their workflow gradually to include PI.  Its reputation for having a steep learning curve is warranted.  For one thing, nothing in PI is called by its normal name.  Look for "stacking": you won't find it.

 

When I started using PI, I had been stacking with DSS for a couple of years and was comfortable with it.  Going to PI, there was no process or script called "stacking".  I had frames to stack, and the only stacking tool I could find was DSS.  Rather than waste time, I stacked with what I had, and started learning how to process in PI. 

 

Once I was confident that PI gave me good enough results to be worth the investment, I started researching how to stack with it.

 

I am sure that a lot of people follow this pattern of gradually migrating their workflow.


  • dswtan, Jim Waters, Woodbridge_Dave and 6 others like this

#15 WadeH237

WadeH237

    Fly Me to the Moon

  • *****
  • Posts: 7,350
  • Joined: 24 Feb 2007
  • Loc: Ellensburg, WA

Posted 21 June 2021 - 06:47 PM

As the title says.  Why do people use DeepSkyStacker for calibration, aligning and stacking then switch to PixInsight?

For my part, I don't have a strong incentive to use just one package to do all of my processing.  I am happy to use different tools for different tasks, if I prefer the workflow.

 

Before PixInsight was a viable option, my workflow was CCDStack for all linear work.  When I was ready to switch to non-linear processing, I would save the CCDStack result to TIF files and then switch to Photoshop.  When I wanted to switch to PixInsight, I didn't do so all at once.

 

At first, I kept using CCDStack for calibration, registration and integration.  From there, I switched to PixInsight for linear processing.  This replaced CCDStack for some things, most notably deconvolution.  I was also quick to adopt DBE.  CCDStack has a gradient reduction feature, but I preferred GradientXTerminator in Photoshop.  DBE was a big improvement on that.  The last thing that I did in PixInsight was to stretch the data and save it as a TIF for Photoshop and the non-linear processing.

 

The next step was to start doing non-linear processing in PixInsight and move almost completely away from Photoshop, although I still use it once in a while when I think that it will solve a particular problem well.  For example, I have just started to experiment with Topaz Denoise AI to do strong noise reduction on the chrominance data (I don't use it for the luminance data).  I'm using Topaz as a filter plug-in in Photoshop.

 

It was much later that I finally adopted calibration, registration and integration in PixInsight.  The reason is that CCDStack is really good at this, and it has (at least in my opinion) a nicer workflow than PixInsight for this stuff.  It's much more interactive and does a nice job of separating data rejection from integration, and does so in a way that you can easily try different rejection settings without having to redo much work.

 

Even though CCDStack is a pretty old tool, I still consider it to be one of the best ways to do this part of processing.  I only switched to PixInsight to pick up a couple of features, like SubFrameSelector and some newer rejection algorithms, like Winsorized Sigma or, more recently, Generalized Extreme Studentized Deviate Test (how's that for a name?)  LocalNormalization and the new NormalizedScaleGradient script are also powerful tools that I've not seen elsewhere.

 

Since I have not used DSS, I can't speak to how it compares to PixInsight, but my experience with CCDStack at least suggests some reasons why people may use tools other than PixInsight for this part of processing.


  • Jim Waters and Morefield like this

#16 Juno18

Juno18

    Ranger 4

  • *****
  • Posts: 346
  • Joined: 07 Dec 2018
  • Loc: Long Beach, Mississippi, USA

Posted 21 June 2021 - 07:22 PM

PI's stacking is more powerful and can produce much better results, but a lot of people find the multiple steps slow and daunting.

 

I think WBPP2.x alleviates a lot of that - I get great  stacking and rarely need to tweak anything with it (though I typically do a manual integration after so I can use SFS and/or the new NSG Script before integration)

 

PI, overall, adds a ton of flexibility - which is super powerful, but at the cost of not having a simple, clear path.

 

My guess is if someone made a form of WBPP2.x with the exact  same options, but took away all the extra fields for adjusting and stripped it to the bare minimum (so it looked like DSS more or less), and scripted it to clean up the interim files, a lot of people who use DSS would use that instead... and get better results.  It would be just like running with the defaults... but simpler to follow.

My experience is very much the same. WBPP produces great results even with the defaults. If I compare simply a stretched DSS stack and the same data stacked by WBPP (weighed batch pre-processing script) in PI, there is a definitely more detain in the PI stack. 

 

I decided about 6 months ago to bite the bullet and get started in PI. Just wasn't getting the results that I wanted in other applications (never tried APP though).

I am amazed at what I find that PI is capable of. I am just scratching the surface and the array of processes and scripts are mind blowing! 

 

Its like everything else in AP, the tough road can pay off in big dividends!


  • dswtan, Jim Waters and mackiedlm like this

#17 ChrisWhite

ChrisWhite

    Fly Me to the Moon

  • *****
  • Posts: 6,851
  • Joined: 28 Feb 2015
  • Loc: Northeast USA

Posted 21 June 2021 - 08:09 PM

Beats me Jim.  When I started out I tried DSS and compared this to PI and the results were always better with PI and just as easy with BPP. 

 

Edit-  Actually, when I first started out I used ImagesPlus for a while.  That program had a learning curve!!!


Edited by ChrisWhite, 21 June 2021 - 08:10 PM.

  • Jim Waters and cobestluc like this

#18 SteveL42

SteveL42

    Vostok 1

  • *****
  • Posts: 108
  • Joined: 18 Sep 2020
  • Loc: Colorado

Posted 21 June 2021 - 08:25 PM

I went from DSS to Siril. So far, I have found it quite usable, and much faster than DSS.  I take the output from Siril and finish it up in Photoshop with Astronomy Tools and a little bit of Topaz Denoise.  

 

From what I have read about PI, it's incredibly powerful, but quite complicated as well.  Photoshop has incredible depth, but is pretty usable with minimal instruction (I have probably 15+ years using it).  It's also not a astro specific tool, so it's been useful for work as well as my hobbies.

 

My take, if you're a professional astrophotographer, or a serious amateur, AND you have the time to invest, PI sounds like the way to go. But for 90% of the hobbyists out there, it doesn't seem needed to achieve excellent results with some of the easier to use packages out there.  Could you do better with PI in your tool belt?  Probably.  Is the return on investment worth the difference?  You'll have to answer that yourself. 

 

Steve



#19 Jim Waters

Jim Waters

    Fly Me to the Moon

  • *****
  • topic starter
  • Posts: 5,854
  • Joined: 21 Oct 2007
  • Loc: Phoenix, AZ USA

Posted 21 June 2021 - 08:56 PM

Beats me Jim.  When I started out I tried DSS and compared this to PI and the results were always better with PI and just as easy with BPP. 

 

Edit-  Actually, when I first started out I used ImagesPlus for a while.  That program had a learning curve!!!

I used ImagesPlus for 2 years before I switched to PI.  The final image quality was hands-down better with PI.  The PI learning curve was much easier.



#20 bobzeq25

bobzeq25

    ISS

  • *****
  • Posts: 24,984
  • Joined: 27 Oct 2014

Posted 21 June 2021 - 09:01 PM

My recommendation for beginners is Astro Pixel Processor.  Doing calibration, stacking, and processing in a single software is a big plus.

 

APP is easy to learn (and a built in workflow actually teaches you how to process), and capable of fine results.


  • lambermo likes this

#21 KTAZ

KTAZ

    Surveyor 1

  • *****
  • Posts: 1,930
  • Joined: 09 Apr 2020
  • Loc: Scottsdale, AZ

Posted 21 June 2021 - 09:07 PM

I just know that I hate coming up against a wall with any software; then having to change to something more powerful and “retrain” myself.

 

When I entered this hobby a year ago, I made the decision to go straight to PI and I’m glad I did. I have barely scratched the surface and it is a very powerful program.



#22 kraegar

kraegar

    Apollo

  • *****
  • Posts: 1,480
  • Joined: 17 Apr 2012

Posted 21 June 2021 - 09:12 PM

I must say, for anyone who has PI but finds stacking complicated, it's 100% worth the time to watch Adam Block's video series on Youtube about WBPP.  It's free, and walks you right through it so you actually understand it, but also he tells you what you need to do in order to "just stack" with pretty much solid defaults, too.

 

https://youtu.be/aZqPjDN8e40

 

Absolutely worth the time it takes to watch!


  • cucubits likes this

#23 whwang

whwang

    Soyuz

  • *****
  • Posts: 3,628
  • Joined: 20 Mar 2013

Posted 21 June 2021 - 09:15 PM

Maybe I am the odd one here. I once compared calibrated/stacked images from DSS and PI a few times, and I can hardly see the differences in quality. My pre-2015 images were stacked with DSS. They are still in my Astrobin gallery, and they speak for themselves.

I switched to PI calibration and stacking in 2015, not because of quality. It was because DSS was still a 32 bit program and it couldn’t handle the many large files from my camera. At the time of switching, it was quite painful. PI can handle huge files, but was very slow. For the same bunch of images, it took PI much longer than DSS. But I had no choices. Now PI is faster and acceptable in speed. So even though DSS became a 64 bit program, I stayed with PI.

And after stacking and perhaps a pass of ABE or DBE, the next thing I do is usually going to Photoshop and finish the rests there.

I would say, PI is a nice program. It’s very useful. But if one day PI is gone from my hard drive for good, I can still produce images of similar quality. On the other hand, if Photoshop is gone, I probably will quit this hobby because there will be no fun left.
  • Joe G and unimatrix0 like this

#24 Ron (Lubbock)

Ron (Lubbock)

    Apollo

  • -----
  • Posts: 1,096
  • Joined: 17 Aug 2012
  • Loc: Pittstown, NY

Posted 21 June 2021 - 09:27 PM

I still use DSS to stack images after nine years of AP.  People who say they get inferior results from DSS are probably not optimizing the various features like cosmetic correction.  One important thing I learned is to stack in DSS and save an unmodified FTS file, but absolutely do not stretch the image in DSS.  Other programs do a better job there.  The rest of my workflow is based around GIMP with a handful of accessory programs that do noise reduction or sharpening when needed.  I believe I downloaded PixInsight three times and gave it a fair shot, but I just didn't want to change my whole workflow and re-learn everything from scratch, with no guarantee of improvement.  Plus, if it ever switches to a subscription model and I'm addicted to it, then.... I'll be less than happy?


  • Jim Waters likes this

#25 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 24,929
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 21 June 2021 - 09:34 PM

Better yet, why pixinsight?  I’d rather take a beating in a dark alley….

Why so much hate for PixInsight?




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





Also tagged with one or more of these keywords: Astrophotography



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics