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

Tell me what you think is wrong here?

  • Please log in to reply
15 replies to this topic

#1 arrowspace90

arrowspace90

    Viking 1

  • -----
  • topic starter
  • Posts: 643
  • Joined: 15 May 2009
  • Loc: United States

Posted 16 May 2021 - 11:15 AM

I am not brand new at this hobby, I've been at it for a bit over a year now, plodding along in the Bortle 6 1/2 suburb.

 

But this camera is new to me, a ZWO ASI2600MM.  I use it on my RASA 8 and for this session had the 2 in, Optolong RGB filters.  1 night for each filter.

 

The individual stacks looked ok in mono, with the usual, significant gradients of LP.  But when I channel combined them in PixInsight, the blue and yellow seem too intense and out of control.

 

This is a cropped image of M63 that I tried DBE and PCC to get back to reality, but it failed.

 

I don't know what caused this problem, so I don't know what to try to do about it. Typically, my raw combined channels look more blandly white prior to PCC.

 

If your experience gives you some idea about what went wrong here or if this is somehow "fixable", please offer me a clue?

 

These were 2 minute subs, each filter getting about 150ish frames. Right now, it's a lot of work for nothing.

Attached Thumbnails

  • M63-DBE.jpg
  • M63-PCC.jpg


#2 kathyastro

kathyastro

    Gemini

  • *****
  • Posts: 3,184
  • Joined: 23 Dec 2016
  • Loc: Nova Scotia

Posted 16 May 2021 - 11:25 AM

You didn't register the masters.  You have a significant rotation between the blue and the other colours.  You need to pick one single frame (pick your best frame) as the registration master, and then star-align all the other frames of all colours to that one frame.

 

Or, if each master looks okay on its own, you just need to register the masters,not the individual frames.


Edited by kathyastro, 16 May 2021 - 11:26 AM.

  • dswtan, jdupton, james7ca and 1 other like this

#3 arrowspace90

arrowspace90

    Viking 1

  • -----
  • topic starter
  • Posts: 643
  • Joined: 15 May 2009
  • Loc: United States

Posted 16 May 2021 - 12:04 PM

You didn't register the masters.  You have a significant rotation between the blue and the other colours.  You need to pick one single frame (pick your best frame) as the registration master, and then star-align all the other frames of all colours to that one frame.

 

Or, if each master looks okay on its own, you just need to register the masters,not the individual frames.

Hmmm.  I sure thought I checked all the boxes in proceeding with WBPP.  But!  It is certainly worth a re-stack from scratch, to see if I can improve this bad outcome.  In fact, not knowing where I went wrong, I had already concluded that a complete re-stack was in order, to rule that out.

Thank you for help, always appreciated.

When I use APP, it automatically picks your best (highest scoring) frame and labels it.  Pix doesn't do that.  I did just arbitrarily pick a frame from each color, not noting which one Subframe Selector judged the best. 



#4 kathyastro

kathyastro

    Gemini

  • *****
  • Posts: 3,184
  • Joined: 23 Dec 2016
  • Loc: Nova Scotia

Posted 16 May 2021 - 02:31 PM

I did just arbitrarily pick a frame from each color, not noting which one Subframe Selector judged the best. 

That may be the problem.  You pick one single frame for all colours, not one from each colour.  All frames of all colours need to be registered to the same frame.  Otherwise each colour will register separately, and you'll get what you got.


  • dswtan, Jon Rista and bortle2 like this

#5 arrowspace90

arrowspace90

    Viking 1

  • -----
  • topic starter
  • Posts: 643
  • Joined: 15 May 2009
  • Loc: United States

Posted 17 May 2021 - 10:35 AM

What I believe may be my biggest error:  I am not practiced sufficiently in PI to automatically do the right thing.  As I noted, I used to use APP.

When I got to star alignment, I ran it 3 times, once for each filter, as I would do for Integration.  This time, aligning with all frames combined.  Then I will integrate for 1 at a time, and then channel combine the 3 integrated results.  I hope this is correct.  

Easy to get the procedure mixed up, many tutorials spend an hour on the theory of flats, darks, etc., instead of outlining simply the WBPP and Align and Integrate steps.



#6 kathyastro

kathyastro

    Gemini

  • *****
  • Posts: 3,184
  • Joined: 23 Dec 2016
  • Loc: Nova Scotia

Posted 17 May 2021 - 10:44 AM

If your work flow makes more sense to align each colour separately, go right ahead.  There is nothing wrong with that.  However, if you do that, you still need to align the colour masters with each other, so that, in the final combination, everything ends up aligned with everything else.



#7 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 5,732
  • Joined: 14 Jun 2009

Posted 17 May 2021 - 11:55 AM

well but then you end up doing two interpolations, once on all the subs and then again on two of the masters. while you probably wouldn't notice a difference, it's still suboptimal. there's no harm at all in specifying a registration reference from a single filter and then registering all 3 filters to that reference.

 

WBPP should just align everything to the reference frame that you choose, or the reference frame that it has chosen.

 

if you do it manually, you are calibrating 3 times (matching lights with flats, but with the same darks and bias, usually), then register once (all frames from all filters loaded against a single reference frame) and then integrate 3 times, once per filter, to obtain 3 masters. then you use channelcombination to create an RGB master. WBPP does these exact same steps, but it stops at the 3 mono masters and you still have to pick up with channelcombination.

 

rob



#8 Marcelofig

Marcelofig

    Viking 1

  • -----
  • Posts: 658
  • Joined: 21 Jan 2015

Posted 17 May 2021 - 03:54 PM

If your work flow makes more sense to align each colour separately, go right ahead.  There is nothing wrong with that.  However, if you do that, you still need to align the colour masters with each other, so that, in the final combination, everything ends up aligned with everything else.

 

That's what I do. It can take several weeks before I have all the channels complete for a given target, so as I get one ready, I process it, generate the master and wait to get the rest ready.

 

Then when all the masters are ready, I register them (StarAling in PI) and combine them for post processing.



#9 kathyastro

kathyastro

    Gemini

  • *****
  • Posts: 3,184
  • Joined: 23 Dec 2016
  • Loc: Nova Scotia

Posted 17 May 2021 - 04:41 PM

That's what I do. It can take several weeks before I have all the channels complete for a given target, so as I get one ready, I process it, generate the master and wait to get the rest ready.

 

Then when all the masters are ready, I register them (StarAling in PI) and combine them for post processing.

Well, I think that is the step you missed in the example image.

 

I took your image, separated the channels into R, G, and B, registered the B with the red, and then re-combined them.  Here is the result:

Image06.jpg


Edited by kathyastro, 17 May 2021 - 05:05 PM.

  • dswtan, arrowspace90 and thekubiaks like this

#10 arrowspace90

arrowspace90

    Viking 1

  • -----
  • topic starter
  • Posts: 643
  • Joined: 15 May 2009
  • Loc: United States

Posted 18 May 2021 - 09:59 AM

Well, I think that is the step you missed in the example image.

 

I took your image, separated the channels into R, G, and B, registered the B with the red, and then re-combined them.  Here is the result:

attachicon.gifImage06.jpg

Well, that does look better, thanks for your interest and trouble.  I am restacking these images now.  They are certainly c_cc_r on the FITS headers. I did pick out a reference frame.  

The blue frames are rotated more than I wanted, the camera must not have been straight up (I don't have the option of leaving the camera attached for a filter change), but of course I can crop out the rotated corners anyway, the rest should be adjustable.

If there is other stuff wrong, and there may be, I haven't figured out exactly what so far.  Hopefully sometime this year, PixInsight will start to click better with me, it's easily the most difficult software that this non IT person has ever tried to use.  I usually don't start having so much trouble while still in the stacking phase.



#11 arrowspace90

arrowspace90

    Viking 1

  • -----
  • topic starter
  • Posts: 643
  • Joined: 15 May 2009
  • Loc: United States

Posted 18 May 2021 - 10:57 AM

I restacked with this result.  Definitely better.  Now has the pinkish cast that I more frequently see and not the bizarre, bright blue and yellow that was a definite screw up.

I still find good color saturation difficult and challenging on PI, but watching yet another tutorial today to improve.

Sometimes, my stars, as can be seen, come out decidedly purple.  I have yet to grasp fully why this happens.

But for my initial bad stack, combing all RGB for aligning definitely helped, appears to have helped. 

Attached Thumbnails

  • M63-restack.JPG

  • dswtan and kathyastro like this

#12 SeymoreStars

SeymoreStars

    Gemini

  • *****
  • Posts: 3,320
  • Joined: 08 May 2014
  • Loc: Pennsyltucky

Posted 18 May 2021 - 12:46 PM

It would benefit everyone if we all learned to manually calibrate images before using WBPP.


  • JuergenB likes this

#13 arrowspace90

arrowspace90

    Viking 1

  • -----
  • topic starter
  • Posts: 643
  • Joined: 15 May 2009
  • Loc: United States

Posted 18 May 2021 - 03:53 PM

It would benefit everyone if we all learned to manually calibrate images before using WBPP.

Isn't calibration mostly a lot of math muscle?  I used APP and sometimes still do.  All through the stacking process it is doing heavy math.  There's a low probability I am gonna be able to absorb that.



#14 SeymoreStars

SeymoreStars

    Gemini

  • *****
  • Posts: 3,320
  • Joined: 08 May 2014
  • Loc: Pennsyltucky

Posted 18 May 2021 - 09:21 PM

Manual calibration in pixInsight means using processes Calibration, CosmeticCorrection, StarAlignment and ImageIntegration, instead of WBPP.



#15 Mert

Mert

    Cosmos

  • *****
  • Posts: 7,894
  • Joined: 31 Aug 2005
  • Loc: Spain, Pamplona

Posted 19 May 2021 - 05:49 AM

For purple stars, you might try to seperate the stars, invert, SNCR, invert back again and combine the stars again.

#16 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 5,732
  • Joined: 14 Jun 2009

Posted 19 May 2021 - 10:36 AM

Manual calibration in pixInsight means using processes Calibration, CosmeticCorrection, StarAlignment and ImageIntegration, instead of WBPP.

well, not really... manual calibration means the process "Calibration", period. it's the process of dark/bias subtraction and flat division.

 

i guess you can refer to the whole pipeline you've called out as "reduction" which is what professional astronomers say.

 

anyway it is useful to understand how to do it manually so that you can debug WBPP when something goes wrong. but WBPP is a huge, huge time saver (and it's only going to get better in the upcoming -8 release of PixInsight)

 

rob


  • JuergenB likes this


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