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

Trialing PixInsight - Basic Stacking Issue

  • Please log in to reply
9 replies to this topic

#1 philinbris

philinbris

    Vostok 1

  • -----
  • topic starter
  • Posts: 161
  • Joined: 10 Jan 2020
  • Loc: Brisbane, Australia

Posted 04 July 2020 - 06:39 PM

Hi All,

I have started learning PixInsight (PI) and have watched a fair few YouTubes and read many Google searches on the subject as well as investing in the "Inside PixInsigght" book. All up I have spent about 8 hrs worth of learning.

I have done a few stacking (Integration) runs and I have found that compared to Registax and AutoStakkert, I am not having much luck.

I have 12 basic subs of M104 that I had absolutely no issues stacking in Registax and AutoStakkert. For a first time go I was very happy. And I achieved that in about 30 minutes - Hour tops.

However, using PI, I have spent a good two hours trying to stack / integrate and unfortunately I have not yet achieved anything close to my first go with the freebie apps.

I have worked out the work flow to be basically:

1. Select the sub frames

2. Calibrate subs

3. Star alignment

4. Integrate

 

Here is a crop of the freebie Registax output:

Reg_Stack.jpg

 

Here is a crop of the Pi Integration Process:

PI_Stack.jpg

 

And here is another run of the same subs using PI FFT Reg script:

PI_FFTReg.jpg

 

 

I get very similar outcomes follwing the 4 steps above as I do if I try to integrate the subs without aligning / calibrating.

Now given the fact that many swear by PI, I am sure PI can do excellent stacking - where the heck am I going wrong with this expensive software? I see potential having everything in one application, I just would have expected a simple stack to be reasonably easy.

I am trying not to get too discouraged, but I am wondering if 230 Euros (about 370 bucks Australian) is a good investment for me given easy success with the freebie alternatives and post processing in Gimp.

Cheers



#2 rgsalinger

rgsalinger

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 6,559
  • Joined: 19 Feb 2007
  • Loc: Carlsbad Ca

Posted 04 July 2020 - 09:08 PM

It looks to me as if the registration step is not succeeding. Make sure that you have selected Register/Match Images as the working mode. Some of the other integration modes are for special cases.  That's all I can think of. Given the amount of drift, you're not going to succeed unless registration is successful. Run it and then blink the images, if there's still a shift in the output images, that's where you are going wrong. 

Rgrds-Ross



#3 futuneral

futuneral

    Apollo

  • *****
  • Posts: 1,268
  • Joined: 27 Dec 2014
  • Loc: Phoenix, AZ

Posted 04 July 2020 - 10:08 PM

one idea - post here what it says in the console after you do Star Alignment



#4 philinbris

philinbris

    Vostok 1

  • -----
  • topic starter
  • Posts: 161
  • Joined: 10 Jan 2020
  • Loc: Brisbane, Australia

Posted 04 July 2020 - 10:52 PM

Thanks for the ideas.

I did load all the registered images and superimposed them and there were two out of the eleven that had the stars way off - the other nine were fine.

The original images all drift from left to right, so for some reason PI calibrated all but 2 of them.

I will re-run with your suggestions.

 

As a side request, did both of you originally use AutoStakkert / Registax?

IS it a case for you both that you would never look back from PI, or do you still use the others sometimes?

Cheers



#5 futuneral

futuneral

    Apollo

  • *****
  • Posts: 1,268
  • Joined: 27 Dec 2014
  • Loc: Phoenix, AZ

Posted 04 July 2020 - 11:03 PM

I used Registax and DSS. Was years ago, but PI worked great for me always. Sometimes you just need to fiddle with it. It likes attention to details and provides the tools to make a lot of adjustments. The price you pay - the default settings may not work right away.

 

After you align your images, they will be saved to your destination folder with a suffix "_r" appended to them. Makes sure you ONLY use those files for integration



#6 philinbris

philinbris

    Vostok 1

  • -----
  • topic starter
  • Posts: 161
  • Joined: 10 Jan 2020
  • Loc: Brisbane, Australia

Posted 04 July 2020 - 11:27 PM

So in PI, when you have selected the reference image, you then add the lights.

When you add the lights, should you exclude the reference image?

When I looked at the YouTube tutorial, I am sure it also added the reference with the lights.

Cheers



#7 Alex McConahay

Alex McConahay

    Cosmos

  • *****
  • Posts: 9,229
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 04 July 2020 - 11:43 PM

I have a pdf tutorial on getting one through the first experience with PixInsight. The link is in my signature. It ain't much, and you can move beyond it pretty quick. But it does get you through the first few encounters. There is one for OSC, and another for LRGB. There were shows on The Astro Imaging Channel to go along with them.

 

Alex

 

http://alexastro.com...iclesindex.html


Edited by Alex McConahay, 04 July 2020 - 11:44 PM.

  • philinbris likes this

#8 philinbris

philinbris

    Vostok 1

  • -----
  • topic starter
  • Posts: 161
  • Joined: 10 Jan 2020
  • Loc: Brisbane, Australia

Posted 05 July 2020 - 03:36 AM

Hi Alex,

Thanks for the tutorial. I have gone through it and your steps are pretty much what I am doing. The only real difference is how you separate the Lum, Blue etc into separate BINs. I just load my RGBs into on BIN.

Oh and you taught me how to quickly add a reference image. So that answers my question above on if it should be excluded from the lights.

I still cannot get a clean stack.

And now I have another question about stacking in PI. It seems this process relies on stars, what about planets where you need top align on features like the red spot on Jupiter?

Thanks for the help

Cheers


Edited by philinbris, 05 July 2020 - 03:55 AM.


#9 Chucke

Chucke

    Viking 1

  • *****
  • Posts: 690
  • Joined: 12 Mar 2010
  • Loc: AZ

Posted 05 July 2020 - 02:55 PM

The reference image can be included in the lights if it uses the same filter.



#10 philinbris

philinbris

    Vostok 1

  • -----
  • topic starter
  • Posts: 161
  • Joined: 10 Jan 2020
  • Loc: Brisbane, Australia

Posted 07 July 2020 - 06:04 AM

Well I did it.

I registered my M104 in PI, but I am not so sure its the "best practice" way of doing it.

What I did is a standard image cal process followed by a FFTRegistration script of the calibrated images.

This is what I got:

PI Stacked.jpg

 

I cannot get ImageIntegration or StarAlignment to work, I just end up with my original problem as per this (using same calibrated images as above):

PI_Integ.jpg

 

So question time.

Is doing a cal followed by FFTReg OK?

What does FFTReg work, but ImageIntegration does not?

 

Cheers




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