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 eclipse imagers: Anyone else perplexed by processing?

  • Please log in to reply
55 replies to this topic

#1 dan_hm

dan_hm

    Messenger

  • *****
  • topic starter
  • Posts: 426
  • Joined: 31 Dec 2012
  • Loc: New Jersey

Posted 26 August 2017 - 11:09 AM

Now that I've had a chance to sit down with my images of totality, I'm pretty puzzled as to how to get a decent HDR image.  I followed Jerry Lodriguss's guide pretty closely and bracketed at 1/125 and 1/4 to get sequences of 1/500, 1/125, 1/30, 1/15, 1/4, and 1 second exposures all at ISO 100 and f 2.8.  In retrospect I should've done shorter exposures overall, but the data doesn't seem so bad that I can't get a decent final product.  I downloaded Alan Dyer's "How to Photograph The Solar Eclipse" and followed his workflow but I'm still pretty lost.  I guess my main questions are as follows:

 

1) Should I stack each set of exposures prior to postprocessing?  If so, what program is appropriate for this?

2) I have a problem with my longer exposures where the brightness of the corona creates posterization around the center.  The whole center of the image is highly overexposed.  Would flats correct this?  If not, how would you go about fixing this?  I only really do deep sky imaging in narrowband so this isn't really an issue I've faced before.

3) Since I used a fixed tripod and not a tracking mount, my frames aren't aligned.  When I manually align the long exposures with the short ones, everything except Regulus seems to align. Any idea why this might be?



#2 Jerry Lodriguss

Jerry Lodriguss

    Vendor

  • *****
  • Vendors
  • Posts: 7252
  • Joined: 19 Jul 2008
  • Loc: Voorhees, NJ

Posted 26 August 2017 - 12:53 PM

Now that I've had a chance to sit down with my images of totality, I'm pretty puzzled as to how to get a decent HDR image.  I followed Jerry Lodriguss's guide pretty closely and bracketed at 1/125 and 1/4 to get sequences of 1/500, 1/125, 1/30, 1/15, 1/4, and 1 second exposures all at ISO 100 and f 2.8.  In retrospect I should've done shorter exposures overall, but the data doesn't seem so bad that I can't get a decent final product.  I downloaded Alan Dyer's "How to Photograph The Solar Eclipse" and followed his workflow but I'm still pretty lost.  I guess my main questions are as follows:

 

1) Should I stack each set of exposures prior to postprocessing?  If so, what program is appropriate for this?

2) I have a problem with my longer exposures where the brightness of the corona creates posterization around the center.  The whole center of the image is highly overexposed.  Would flats correct this?  If not, how would you go about fixing this?  I only really do deep sky imaging in narrowband so this isn't really an issue I've faced before.

3) Since I used a fixed tripod and not a tracking mount, my frames aren't aligned.  When I manually align the long exposures with the short ones, everything except Regulus seems to align. Any idea why this might be?

1. You should stack exposures with the same shutter speed. Any astro stacking program.

 

2. That's normal.

 

3. You need to align them manually to a base image. If you aligned on the Moon, it is moving, therefore Regulus is going to move. You really want to align on Regulus.

 

Here's how you do HDR:

 

http://www.astropix....git/e_comp.html

 

Jerry


  • ismosi likes this

#3 dan_hm

dan_hm

    Messenger

  • *****
  • topic starter
  • Posts: 426
  • Joined: 31 Dec 2012
  • Loc: New Jersey

Posted 26 August 2017 - 10:19 PM

Thanks, Jerry.  I finally figured out what I was doing wrong with manual align (had to move the reference point in Free Transform).  I stacked each set in PixInsight but got weird color in my integrations for 1/30 and 1/125.  Maybe I'll try Autostakkert.  I went for a quick HDR with luminosity masks and High Pass filters.  The results are...interesting:

 

get.jpg

 

 



#4 Calypte

Calypte

    Surveyor 1

  • -----
  • Posts: 1514
  • Joined: 20 Mar 2007
  • Loc: Anza, California

Posted 26 August 2017 - 11:00 PM

As soon as I got home I created jpegs of three frames to show to friends & family, to show something from all of the fuss about the eclipse.  But creating something better has left me defeated so far.  I bought the Photomatix tool today, which helped with the HDR combine, and I applied two rounds of high-pass filtering with Photoshop.  But troubles ensue from there.  I get nice extensions on the corona, but I also get a bright, glowing halo around the sun-moon.  I've tried removing the halo with GradientXTerminator, with limited success.  The problem is that, depending on how aggressive I am with this tool, I get dark, faint, concentric rings around the eclipse.  They show up easily when the photo is viewed on a cell phone.  I have no flats or darks -- wouldn't know how to make them with a DSLR (they're routine with my CCD setup), but I don't think that's the problem here.  The eclipse is a bit off-center in all of my shots, but the faint rings don't conform to the axis of the optical system.  Rather, they are centered on the eclipse.  Whatever -- the eclipse is behind me, and darks & flats didn't happen.  So I'm still stuck with curing the glowing halo problem.



#5 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 5172
  • Joined: 14 Jun 2009

Posted 27 August 2017 - 04:28 PM

i see the halo in my images as well - at first i thought it was some kind of artifact, but it must be real. i think you need to resort to multiscale processing techniques to tame this halo (meaning, create an image with only the largest scales present and then somehow use it to "flatten" the image such that a histogram transformation works right to isolate the corona itself)... or just operate on the small-scale image to enhance details.

 

the posterization seems to be a real issue. i do wonder if parts of the corona that fall into the nonlinear portion of the sensor's response are still participating in the HDR merge and as such have the wrong color/brightness relative to the properly exposed portions. for my part i had to use the masks output by HDRComposition to fix up the brightness in each posterization level a little bit.

 

get.jpg

 

rob


Edited by pfile, 27 August 2017 - 04:28 PM.


#6 Jerry Lodriguss

Jerry Lodriguss

    Vendor

  • *****
  • Vendors
  • Posts: 7252
  • Joined: 19 Jul 2008
  • Loc: Voorhees, NJ

Posted 28 August 2017 - 06:02 AM

This was done completely in Photoshop.

 

First one is with Russel Brown's Smart Object mean combine and then a radially graded filter:

 

Attached File  HDR_Corona_RB_PS.jpg   73.41KB   14 downloads

 

The next is a combination of RB's technique and the Shiota-Pellet technique:

 

Attached File  Lodriguss_2017_Total_Solar_Eclipse_HDR_Corona_800px.jpg   111.96KB   19 downloads

 

Jerry


  • EricCCD, rainycityastro, dan_hm and 5 others like this

#7 AUricle

AUricle

    Messenger

  • *****
  • Posts: 474
  • Joined: 05 Jan 2017
  • Loc: Frankfort, IL

Posted 28 August 2017 - 01:26 PM

The next is a combination of RB's technique and the Shiota-Pellet technique:

 

attachicon.gifLodriguss_2017_Total_Solar_Eclipse_HDR_Corona_800px.jpg

 

Jerry

Jerry,

 

NFW!!.......gorgeousbow.gif bow.gif bow.gif 



#8 pfile

pfile

    Fly Me to the Moon

  • -----
  • Posts: 5172
  • Joined: 14 Jun 2009

Posted 28 August 2017 - 09:58 PM

really nice, superb detail in the corona! i will have to make another go at it with these radial techniques.

 

rob



#9 rainycityastro

rainycityastro

    Apollo

  • *****
  • Posts: 1312
  • Joined: 29 Mar 2010
  • Loc: Seattle, WA

Posted 29 August 2017 - 12:55 PM

Hi Jerry, Amazing image and processing!

Could you provide some information on Russel Brown Smart object mean combine?



#10 Jerry Lodriguss

Jerry Lodriguss

    Vendor

  • *****
  • Vendors
  • Posts: 7252
  • Joined: 19 Jul 2008
  • Loc: Voorhees, NJ

Posted 29 August 2017 - 01:42 PM

Hi Jerry, Amazing image and processing!

Could you provide some information on Russel Brown Smart object mean combine?

Hi Ram,

 

Thanks!

 

Just google Russel Brown Eclipse Processing on YouTube. He has two videos.

 

Jerry


  • rainycityastro and ilovecomets like this

#11 SKYGZR

SKYGZR

    Vanguard

  • *****
  • Posts: 2016
  • Joined: 13 Aug 2009
  • Loc: Planet Earth, Milky Way Galaxy, Speeding towards the Virgo Supercluster

Posted 30 August 2017 - 04:05 PM

I've kind of hit the proverbial brick wall...Using PS6 & Bridge, began to follow Russel's procedure. First, tried to highlight all .cr2 raw files in Bridge & process in Adobe Camera Raw, yet got a warning that "a valid product needed to be launched first to use raw"...(or something to that affect). So, since each would individually open in Raw(7), I adjusted all the sliders, etc for each of the individual frames, saved 'em all to 16b, just like he suggests.

 

Highlighted them all & exported as layers to PS6..all good..Created/Converted to Sharp Object thumbnail...all good..here's the wall...upon going to the next step, "stack mode" is grayed out (ie;not active), so unable to select it to proceed any further...

 

So at this point seems I'm smacked into this wall & can't proceed any further with this method till it's figured out why this is happening.

 

Did go back into bridge & did a couple HDR's, yet would sure like to get past this wall to be able to continue with the process...

 

Any Hint's @ what to double check, look for, as to why?



#12 APshooter

APshooter

    Soyuz

  • *****
  • Posts: 3707
  • Joined: 14 Oct 2013
  • Loc: Camby, In.

Posted 31 August 2017 - 07:27 PM

Try 8 bit? I'm not at home in front of my computer tho test this. Stacking worked for me.

#13 SKYGZR

SKYGZR

    Vanguard

  • *****
  • Posts: 2016
  • Joined: 13 Aug 2009
  • Loc: Planet Earth, Milky Way Galaxy, Speeding towards the Virgo Supercluster

Posted 31 August 2017 - 11:16 PM

Try 8 bit? I'm not at home in front of my computer tho test this. Stacking worked for me.

No Go..8-bit still hits the wall..(had to convert to 8-.tif)..perhaps will try 16b .tif..yet seems this is gonna be a lesson in futility...

 

Might have to D/L the LR/CC trial & see if that works....?



#14 RutileQ

RutileQ

    Explorer 1

  • -----
  • Posts: 54
  • Joined: 27 Jun 2017

Posted 01 September 2017 - 12:19 AM

Not perplexed so much, but wishing i could automate more.  I was shooting at 8800 feet in the Sawtooths, you'd think I'd have minimal atmospheric distortion, but we had smoke roll in over the night.  People down in Stanley keep telling me they had clear skies and great conditions, and I know that even with a fair amount of smoke, zenith views can be really clear, but I'm trying to align my chromosphere and prominences exposures to stack them for a higher S/N, and the amount of dilation and contraction of the lunar disk, and the features around it can rival that of the thick air at 300 feet where I live.  



#15 Mad MikeE

Mad MikeE

    Vostok 1

  • *****
  • Posts: 129
  • Joined: 27 May 2006

Posted 01 September 2017 - 01:39 PM

I'm making progress but have more questions now smile.gif

First - latest attempt w/ using Pellett and photoshop (This was shot on the 135mm)

(Also inspiration to the frustrated and perplexed - if I can do this (total newbie) so can you smile.gif

 

Attached File  AlladdedandOriginalwithcolorcorrection1shrunk.jpg   58.73KB   2 downloads

I seem to have a lot of horizontal streaks (thinking enhanced sensor noise) - Any idea what to do to lower that (even if it means going and taking darks or flats (if so which?))?

 

And then there is the ever-present gray-zone...

Just noticed the ring around the moon, that one I will try some tinkering with before I ask more.

 

I also have 3 sets of images for this camera during totality, any thought on combining same exposure times?

Please and Thanks in advance


Edited by Mad MikeE, 01 September 2017 - 01:45 PM.

  • dan_hm and ismosi like this

#16 dghundt

dghundt

    Vostok 1

  • -----
  • Posts: 194
  • Joined: 01 Mar 2017

Posted 01 September 2017 - 06:32 PM

My canon had faint horizontal sensor lines when I initially pushed it too hard with too much clarity and raising shadows. You have to watch each step and stop at some point.
For the ring around the moon, you could at the end of creating your composite mask in a slightly enlarged moon on top.

Edited by dghundt, 01 September 2017 - 06:34 PM.


#17 dan_hm

dan_hm

    Messenger

  • *****
  • topic starter
  • Posts: 426
  • Joined: 31 Dec 2012
  • Loc: New Jersey

Posted 02 September 2017 - 10:07 AM

Here's what I finally settled on:

 

get.jpg

 

Remaining problems are the halo and getting the moon right.  I just can't quite figure out a workaround for these, but I was pretty happy with how the corona and background turned out, and that I was able to get 3 stars.


  • charotarguy likes this

#18 Bill G.

Bill G.

    Viking 1

  • -----
  • Posts: 600
  • Joined: 04 Apr 2004
  • Loc: Newton, NC USA

Posted 02 September 2017 - 04:43 PM

I've gotten some pretty good images by processing a single image but have had no luck with stacking. First, all the software I've been using seems to be confused by the inverted image and does not auto stack the images correctly. So, figured I'd stack manually. Well the framing I got with ED80 and D7000 seemed like it would be great....but then corona was larger than I expected and Regulus was outside the frame. So not exactly sure how to align the images without a common alignment point. I was using bracketing and got good under, normal, and over shots. But tracking was off so I will have to stack manually. 

 

Any suggestions on what program would be a good one to manually align shots with no well defined point of reference??

 

Thanks,

Bill G.



#19 dan_hm

dan_hm

    Messenger

  • *****
  • topic starter
  • Posts: 426
  • Joined: 31 Dec 2012
  • Loc: New Jersey

Posted 02 September 2017 - 05:38 PM

I've gotten some pretty good images by processing a single image but have had no luck with stacking. First, all the software I've been using seems to be confused by the inverted image and does not auto stack the images correctly. So, figured I'd stack manually. Well the framing I got with ED80 and D7000 seemed like it would be great....but then corona was larger than I expected and Regulus was outside the frame. So not exactly sure how to align the images without a common alignment point. I was using bracketing and got good under, normal, and over shots. But tracking was off so I will have to stack manually. 

 

Any suggestions on what program would be a good one to manually align shots with no well defined point of reference??

 

Thanks,

Bill G.

I opened my images as layers in Photoshop and used the "Difference" blending mode to align them.  When I used Regulus as a reference, the moon was somewhat out of alignment but not so bad that it would significantly impact quality.



#20 dghundt

dghundt

    Vostok 1

  • -----
  • Posts: 194
  • Joined: 01 Mar 2017

Posted 02 September 2017 - 07:42 PM

I found using the Fitswork eclipse tutorial worked well for the halo and radial contrast.  I used only step 3, and saved the LS-mask as a tiff, put it on top of my image layers in Photoshop using a hard light blend and partial opacity.

Attached Files


  • Mad MikeE and REC like this

#21 Mad MikeE

Mad MikeE

    Vostok 1

  • *****
  • Posts: 129
  • Joined: 27 May 2006

Posted 02 September 2017 - 07:54 PM

I found using the Fitswork eclipse tutorial worked well for the halo and radial contrast.  I used only step 3, and saved the LS-mask as a tiff, put it on top of my image layers in Photoshop using a hard light blend and partial opacity.

That came out REALLY well!

What is this Fitswork eclipse tutorial that you speak of -> directions please???

 

Thanks in advance!



#22 dghundt

dghundt

    Vostok 1

  • -----
  • Posts: 194
  • Joined: 01 Mar 2017

Posted 02 September 2017 - 08:06 PM

http://www.gva-hambu...fitswork_uk.htm

Fitswork is free. You can Google for the two downloads you need to install.
The eclipse tutorial was harder to find, so I linked it above.
I used Alan Dyer's method from his ebook to process everything and added the Fitswork part at the very end.

I processed my raw images in Lightroom for clarity and shadows mostly, and then exported just three images to Photoshop.
Stacked and processed in Photoshop per Alan Dyer's tutorial.
At that point it was nice, but the radial contrast and halo was still a bit strong.
Then I went to Fitswork with a tiff of my Photoshop composite. Used only step three to create a LS-mask. In step three I additionally used the Fitswork option to soften the effect some.
In Fitswork I did not combine the LS-mask with any images. I simply saved/exported ONLY the LS-mask as a tiff and put it into my Photoshop document as a new layer above the image layers, below the adjustment layers. Changed blend mode to hard light and reduced opacity to taste.
I imagine one could opt to not soften the mask in Fitswork and use Photoshop to increase the mask contrast, but I wanted something not too strong.
Hope that helps.

Edited by dghundt, 02 September 2017 - 08:38 PM.


#23 ismosi

ismosi

    Viking 1

  • -----
  • Posts: 739
  • Joined: 19 Apr 2014
  • Loc: New London, Pennsylvania, USA

Posted 02 September 2017 - 08:16 PM

I'm perplexed but this image is my stopping point for tonight. I tried the Russell Preston Brown method with 11 exposures at 1-stop intervals, ranging from 1/500 to 2 seconds. I have longer exposures but they all seem washed out, I suspect because there was some high haze and/or thin clouds (I was in Lexington, SC). The imaging sequence was automated by Solar Eclipse Maestro, and, I used my WO 80 'Red Anniversary' scope and modded Nikon D610.

 

Being a total n00b eclipse photographer (in fact this was my first total eclipse!) I'm sorta happy with this, but, being me, I'll no doubt try a few more iterations and see how that goes.

 

It's also up on Astrobin as a not-JPEG ...

http://astrob.in/310718/0/

 

 

Attached File  HDR-Corona-Attempt-2.jpg   170.14KB   2 downloads


  • charotarguy likes this

#24 dghundt

dghundt

    Vostok 1

  • -----
  • Posts: 194
  • Joined: 01 Mar 2017

Posted 02 September 2017 - 08:24 PM

The problem with Photoshop's radial blur is the amount of image lost if the sun is not perfectly centered.
With the LS-mask in Fitswork, a much smaller amount of frame edge is lost. I simply repaired this in Photoshop which meant I didn't have to crop my final composite.

#25 Mad MikeE

Mad MikeE

    Vostok 1

  • *****
  • Posts: 129
  • Joined: 27 May 2006

Posted 02 September 2017 - 09:51 PM

The problem with Photoshop's radial blur is the amount of image lost if the sun is not perfectly centered.
With the LS-mask in Fitswork, a much smaller amount of frame edge is lost. I simply repaired this in Photoshop which meant I didn't have to crop my final composite.

Of course this is posted RIGHT after I pay for a year's subscription to PhotoShop (119$) fingertap.gif

FitsWork looks REALLY interesting smile.gif

Got it and will load / watch tutorials over the next couple of days.

 

Interesting that everyone that uses the radial filters has an arc going through Regulus (I have learned about the healing brush in PhotoShop to deal w/ that smile.gif)

And for those that are using the blurs w/ PS, centering the image is easy (After watching Russell...) 
Set the rulers to percent, throw up a cross hair (drag a line from top and left and watch the rulers).
Select all aligned image layers and move them as a unit so they are centered (move cursor to the 4 sides and check to get them pretty equal around the 50 percent center).

Then use your blur of choice and have at it!
Hope that helps...




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