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

Would you process my Heart and Soul (OSC+Ha)?

  • Please log in to reply
29 replies to this topic

#1 the Elf

the Elf

    Soyuz

  • *****
  • topic starter
  • Posts: 3,635
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 12 November 2020 - 01:35 PM

Dears,

 

often I download other user's data and throw in my 2 cent. This time I'd like to ask you to do some processing for me. I have just finished my version but I will not post it yet because I don't want to point into and direction. I will post it later in this thread. There are two reasons why I'd like to see your processing:

a) I ran into a technical problem that either occurred for the first time or I simply have not seen it in my previous images (you will see in my version later)

b) I'm seeking inspiration from the aesthetic point of view. I feel like my images look too much the same every time. I'd like to improve.

 

The data:

 

54 x 3min OSC from an unmodded Canon EOS 800D/T7i at ISO 200, thru Baader semi APO filter, cropped and background removed, 195MB

http://elf-of-lothlo...SC_crop_ABE.fit

 

32 x 15min Ha from a mono modded (Bayer pattern removed) full spectrum EOS 600D/T3i at ISO 800 thru Baader 7nm Ha, 65MB, aligned to the OSC image

http://elf-of-lothlo...Ha_crop_ABE.fit

 

0 x nothing O-III. I tried but the signal is too weak for my uncooled camera. The image basically contained all the stars and a very tiny noisy blob without any shape or structure.

I was hoping for an HOO-image but it was not meant to be.

 

Scope is the askar 180mm focal length, 40mm aperture (semi-)APO on an EQ6-R, guided and dithered under Bortle 4(?) skies, some 5°C ambient temperature

 

Looking forward to your images. Thank you very much!

 

CS

the Elf


Edited by the Elf, 12 November 2020 - 01:36 PM.


#2 Tapio

Tapio

    Skylab

  • -----
  • Posts: 4,333
  • Joined: 24 Sep 2006
  • Loc: Tampere, Finland

Posted 12 November 2020 - 02:13 PM

I have been ill lately and just beginning to recover.

So it's been a while doing any processing.

Never imaged this target (for some reason I don't like this target), but considering circumstances I did a quick processing run.

And I'm not too happy of the result.

I'm sure you can do way better.

Attached Thumbnails

  • 01_OSC_Ha_crop_ABE.jpg


#3 Joe G

Joe G

    Vanguard

  • *****
  • Posts: 2,007
  • Joined: 10 Jan 2007
  • Loc: Southern California

Posted 12 November 2020 - 02:38 PM

When I click on your links I just get a bunch of computer gibberish.



#4 guidoforrier

guidoforrier

    Sputnik

  • -----
  • Posts: 33
  • Joined: 25 Sep 2017

Posted 12 November 2020 - 02:45 PM

can you provide the raw's from your camera ? Wetransfer ? Dropbox ?

I can try APP .

Guido



#5 Tapio

Tapio

    Skylab

  • -----
  • Posts: 4,333
  • Joined: 24 Sep 2006
  • Loc: Tampere, Finland

Posted 12 November 2020 - 02:49 PM

Joe, right click image and choose save as.

#6 Joe G

Joe G

    Vanguard

  • *****
  • Posts: 2,007
  • Joined: 10 Jan 2007
  • Loc: Southern California

Posted 12 November 2020 - 03:05 PM

Doh, thanks Tapio!



#7 the Elf

the Elf

    Soyuz

  • *****
  • topic starter
  • Posts: 3,635
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 12 November 2020 - 03:13 PM

Thank you for taking the time!

Guido, I am afraid I can't. Lights and Flats alone are 3.5 Gigs, if I include bias and flats that's another 10 Gigs and my upload bandwidth is too low to get this done in a reasonable time.

Tapio, apart from you flipped the data vertically your processing shows exactly the same problem as mine: strong horizontal color banding.

Any suggestions where the banding comes from? It it ISO 200? Should I return to ISO 400? Is there anything I can get rid of it?


Edited by the Elf, 12 November 2020 - 03:13 PM.


#8 OhmEye

OhmEye

    Viking 1

  • *****
  • Posts: 986
  • Joined: 15 Sep 2019
  • Loc: Western NY Southern Tier

Posted 12 November 2020 - 03:30 PM

My clumsy/quick/aggressive effort. I clipped a lot of saturation to hide the banding.

 

elf_heartsoul_500k.jpg

 

Larger size https://cdn.discorda...f_heartsoul.jpg


  • dswtan, LauraMS and Cfreerksen like this

#9 the Elf

the Elf

    Soyuz

  • *****
  • topic starter
  • Posts: 3,635
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 12 November 2020 - 03:34 PM

That is a cool interpretation! Looks like fire and smoke. This is the kind of inspiration I was looking for. Thank you for taking the effort.



#10 the Elf

the Elf

    Soyuz

  • *****
  • topic starter
  • Posts: 3,635
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 12 November 2020 - 03:40 PM

OK, time to show my version:

 

CN.jpg

 

Looks awful in the browser because 10,000 stars in the image are single pixel in full res and just disappear when scaling down. Here is a link to full res and fullHD

http://elf-of-lothlo...__2020_full.jpg

http://elf-of-lothlo...__2020_1080.jpg

 

Critique and suggestions welcome!


Edited by the Elf, 12 November 2020 - 03:44 PM.

  • Scott Mitchell likes this

#11 Joe G

Joe G

    Vanguard

  • *****
  • Posts: 2,007
  • Joined: 10 Jan 2007
  • Loc: Southern California

Posted 12 November 2020 - 03:42 PM

Quick PS process.

 

 

Attached Thumbnails

  • Elf H and S CNs.jpg


#12 the Elf

the Elf

    Soyuz

  • *****
  • topic starter
  • Posts: 3,635
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 12 November 2020 - 03:52 PM

Thank you!

The color banding is in the data. I was thinking my processing enhanced it but it seems to be quite prominent. Got to track this down. Sigh.



#13 calypsob

calypsob

    Fly Me to the Moon

  • *****
  • Posts: 6,475
  • Joined: 20 Apr 2013
  • Loc: Virginia

Posted 12 November 2020 - 10:16 PM

The banding is in your color data.  I think I did an ok job attenuating it into the background with the canon banding script in PI in 2 iterations of varying intensity. 

I cannot find for the life of me an old post I made where someone explained very clearly how I can extract the red channel from OSC without debayering and then integrate it with my monochrome Ha.  It worked really well to clean up the red channel.    Its bee a long time since I last messed with Ha+OSC data but Im going to give it a shot. 



#14 calypsob

calypsob

    Fly Me to the Moon

  • *****
  • Posts: 6,475
  • Joined: 20 Apr 2013
  • Loc: Virginia

Posted 12 November 2020 - 11:03 PM

Ok I gave it a shot.  First I completely processed the Ha data in photoshop since this is obviously where all the signal exists. I did a star removal, shrink, sharpening and noise reduction and tonal adjustments with curves. I saved this file for later use in PI   

 

Then back to PI - I l ran canon banding script twice and linear fit the RGB data.  I used NBRGB to add the processed Ha to the linear RGB data with the Ha scale setting at 30.   This produced an ok image with a green background.

 

I took this back into photoshop and used highlight and shadow split toning to shift the green back into its appropriate white balance range in the shadows and highlights.  After that I treated it like a normal image using match color to add saturation to the stars and selective color and color balance to make subtle shifts in the low, mid, and high frequencies. 

 

That Ha data is great, I really need to shoot some and add it to a recent image of my own.  The color data is not bad but definitely needs a deeper integration to really overcome the noise floor.   It may be worth your while to integrate the osc and ha into a single image to use in an LLRGB workflow. 

 

I think this process worked out ok, though I obviously lost the Oiii that is preserved in the other images above, I have a bad habit of doing that. If you say the Oiii you shot was unusable, you may want to try integrating with the osc green and blue channels to see if the integrated signal is usable. 

 

 

Sorry its so small, CN has a 500kb limit and my flickr is not behaving tonight.  here is a link to the full size on google drive https://drive.google...iew?usp=sharing

 

heart-data-elf.jpg  


Edited by calypsob, 12 November 2020 - 11:11 PM.


#15 rekokich

rekokich

    Apollo

  • *****
  • Posts: 1,110
  • Joined: 14 Sep 2014
  • Loc: Virginia

Posted 13 November 2020 - 01:46 AM

I don't know the reason for the banding. Have you had similar problems with iso 400 or 800? What about different ambient temperatures? Different filters?

 

In my processing I made a reverse mask of all light features, and desaturated the sky to 0. There is still evidence of banding in the nebulae and in star colors.

 

h7.jpg


Edited by rekokich, 13 November 2020 - 01:51 AM.


#16 Tapio

Tapio

    Skylab

  • -----
  • Posts: 4,333
  • Joined: 24 Sep 2006
  • Loc: Tampere, Finland

Posted 13 November 2020 - 02:30 AM

Banding is not too obvious to me.



#17 the Elf

the Elf

    Soyuz

  • *****
  • topic starter
  • Posts: 3,635
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 13 November 2020 - 02:38 AM

Wes, thank you for taking the time. Nice one. Your remark about the red channel makes me think I split the OSC data to channels and check if the banding is in all channels or only red which is almost the one with the weakest signal. Going to find out this weekend.

 

rekokich, it was only recently that I started using ISO 200. As I own the camera for less than a year I have only few ISO 400 images for comparison. So far I have not seen any banding there but I will dig out the stacks and stretch them much more to check.


  • calypsob likes this

#18 rekokich

rekokich

    Apollo

  • *****
  • Posts: 1,110
  • Joined: 14 Sep 2014
  • Loc: Virginia

Posted 13 November 2020 - 12:34 PM

Elf and Tapio,

 

I oversaturated the original image to accentuate banding.

 

I do not think this is amp-glow but, make sure Live View is OFF before starting exposures. Canon cameras have a known issue with amp-glow on long exposures if Live View is ON.

 

I think this is a type of sensor defect in all three channels. Consider contacting Canon service for their opinion regarding the issue.

 

Meanwhile, make long test exposures at iso 200, 400, 800, with Long Exposure Compensation OFF and ON. Also, make similar darks.

 

I would be very interested to see the results, especially the darks, but in JPG format please.

 

Good luck !

 

banding.jpg


Edited by rekokich, 13 November 2020 - 01:00 PM.


#19 rekokich

rekokich

    Apollo

  • *****
  • Posts: 1,110
  • Joined: 14 Sep 2014
  • Loc: Virginia

Posted 13 November 2020 - 12:58 PM

click on image if it does not blink

 

animation.gif



#20 George Simon

George Simon

    Viking 1

  • *****
  • Posts: 706
  • Joined: 01 Jul 2008
  • Loc: Chesnee, SC

Posted 13 November 2020 - 01:22 PM

Elf,

 

Here's my rendition of your data, using StarTools and GIMP. I processed the OSC and Ha data separately, and then blended, using the Ha data as the luminance layer. I think proceeding in this way eliminated the banding that was present in the OSC data, or at least I don't see it in the final product.

 

01_Ha_crop_ABE_blend.jpeg


  • calypsob likes this

#21 calypsob

calypsob

    Fly Me to the Moon

  • *****
  • Posts: 6,475
  • Joined: 20 Apr 2013
  • Loc: Virginia

Posted 13 November 2020 - 01:36 PM

Wes, thank you for taking the time. Nice one. Your remark about the red channel makes me think I split the OSC data to channels and check if the banding is in all channels or only red which is almost the one with the weakest signal. Going to find out this weekend.

 

rekokich, it was only recently that I started using ISO 200. As I own the camera for less than a year I have only few ISO 400 images for comparison. So far I have not seen any banding there but I will dig out the stacks and stretch them much more to check.

You are very welcome.  

 

Regarding the red channel, I cannot remember now what the measurement suggested but I would assume that the red plus Ha would certainly combine to have a higher SNR than the G and B channels. 

 

 

What I suggest you investigate further this weekend is taking your pre calibrated OSC data and batch splitting RGB. Once you have each channel separated, register and integrate the red non debayered osc channels with the monochrome Ha data. Then integrate it

 

Do the same with your poor quality OIII data and either the green or blue channel data.  When you are done you can recombine the data to RGB using NBRGB as I described before.  Just remember that before you do any NBRGB you should first run the canon banding removal script twice, I did it at 100% and then at 60%, that seemed to help the color data alot. 



#22 the Elf

the Elf

    Soyuz

  • *****
  • topic starter
  • Posts: 3,635
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 13 November 2020 - 05:22 PM

yeah, but if possible I'd like to avoid banding in the first place using the optimal settings rather than removing it in processing. Alas it is raining now and the forecast is bad for the next few days, so not a chance for any tests at all. Sigh.



#23 Tiab

Tiab

    Lift Off

  • -----
  • Posts: 7
  • Joined: 13 Sep 2020

Posted 14 November 2020 - 08:07 AM

I think this is a known issue. Pix insight has a workflow and tool specifically for Cannon debanding. Google it for a ton of results.

#24 Pete_xl

Pete_xl

    Apollo

  • -----
  • Posts: 1,326
  • Joined: 29 Jan 2015
  • Loc: Germany

Posted 14 November 2020 - 08:40 AM

Hi Elf,

 

also would like ti give it a try this weekend...but have no luck downloading the images. Right click on the links does not give me an opportunity to save an image. Can you provide another link here or via PM?

 

Cheers

Pete



#25 calypsob

calypsob

    Fly Me to the Moon

  • *****
  • Posts: 6,475
  • Joined: 20 Apr 2013
  • Loc: Virginia

Posted 14 November 2020 - 08:57 AM

yeah, but if possible I'd like to avoid banding in the first place using the optimal settings rather than removing it in processing. Alas it is raining now and the forecast is bad for the next few days, so not a chance for any tests at all. Sigh.

Did you turn off your dslr screen while imaging? I assume you did. I have had mine come on for whatever reason and it messed up my data.  Also on a side note, the low iso 200 might be causing banding, you would need to do some extensive testing.  Luckily you could just shoot a series of dark frames inside at different iso's and blink through each set. 

 

I do not have the canon 800d, but in my experience the dynamic range you gain from shooting below iso 800 does not warrant the read noise that goes through the roof at those lower iso's.  Iso 800 - 1600 has always seemed to do really well for color and narrow band, clipping the cores of only the brightest stars. 




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