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

Newbie hopefully past the basics... C&C solicited...

  • Please log in to reply
21 replies to this topic

#1 sbharrat

sbharrat

    Apollo

  • -----
  • topic starter
  • Posts: 1,217
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 18 October 2021 - 08:08 PM

UPDATE: this will disappear off the first page of posts soon so updating with a plaintive plea. Really interested in some critical review. If you need a different format to properly review, please let me know what and I will upload and provide link. Thanks!

 

I am a relative newbie having gotten the full set of an ok mount (Sirius/HEQ5), a good apo (TS PL 80mm triplet), and a cooled camera (asi533) about 6 months ago. (Started in Dec 2020 and used an achro and a DSLR at the start.) I use NINA/Phd2 for capture control, and PixInsight for processing.

 

At this point, I think I have the fundamentals down. On capture I use either an L-eNhance for emission neb or a Neodymium for everything else (I am in Bortle 8 skies), typically 180s w/ L-eNhance, 60s with Neodymium, guided (RA & Dec 0.5 to 0.8" error), ditter every 3rd frame, use darks, flats, and dark-flats for calibration.

 

In processing, I typically do DCR, DBE, PCC, Deconv, MLT (NR), masked stretch, HTF, TGV (NR), Curves, and MLT (star erosion). Sometimes I will split out the stars using starnet and process the nebula separately and recombine at the end. I have gone through the Braken and Keller books and at least understand the techniques. I have also purchased the Adam Block Fundamentals and have watched many of them but there are a lot!

 

I am now trying to determine my biggest-bang-for-the-bug areas to focus on. Below are three recent images along with links to a higher res jpg. If you spend a couple of minutes poking at them and giving me your comments, I would greatly appreciate it. What are the top two or three issues that jump out at you? Thanks in advance...

 

elephant-ic1396-hoo-small-sep2021.jpg

https://astro.bharra...96-sep2021.html

 

(continued in next post)


Edited by sbharrat, 19 October 2021 - 08:26 AM.

  • mikewayne3, Ballyhoo, david.shadrach and 5 others like this

#2 sbharrat

sbharrat

    Apollo

  • -----
  • topic starter
  • Posts: 1,217
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 18 October 2021 - 08:12 PM

ngc6960-small-jul2021.jpg

 

https://astro.bharra...60-jul2021.html

 

m16-poc-small-jun2021.jpg

 

https://astro.bharra...oc-jun2021.html

 


  • Tim J Fowler likes this

#3 Ballyhoo

Ballyhoo

    Fly Me to the Moon

  • *****
  • Posts: 6,937
  • Joined: 22 Jul 2011
  • Loc: San Diego

Posted 18 October 2021 - 08:22 PM

I like them all. 


  • sbharrat likes this

#4 sbharrat

sbharrat

    Apollo

  • -----
  • topic starter
  • Posts: 1,217
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 20 October 2021 - 08:35 AM

@moderator. Given the absence of critique, I think maybe the Experienced Forum isn't the right place for this (though I could swear I saw similar requests in the past). Can you please move to the Beginner's Forum frown.gif



#5 raguramm

raguramm

    Vostok 1

  • *****
  • Posts: 146
  • Joined: 15 Jan 2019

Posted 20 October 2021 - 09:15 AM

For having done this for less than a year, these are amazing images. Well done. Your stars are round to the edge of the frame, you've got good detail, no gradients, pictures are calibrated well. 

 

With regards to your request, I'll give you 4 points of feedback. But please know that I really had to nitpick to find issues. You should be proud.

 

1) you're doing too much noise reduction. I see a soft, plasticky kind of look and a mottle in wavelet layers 4-5 at 1:1 scale that comes from going too hard with TGVDenoise. IMO, image looks most pleasing when there's some fine grain noise at 1:1 - the goal being noise *reduction* and not noise *elimination*. You can also try running a masked MultiscaleMedianTransform with denoise enabled on just wavelets 4-5 (again, very light - threshold of 2-5 and amounts set to <0.4) to address some of the inevitable mottling from TGV.

 

2) you're reducing stars a bit too much. Small stars are fine, but I can pick out artefacts around stars if reduced to excess, or if they're added back to a starless image that's been pushed too far without them (starnet is not a particularly clean program). 

 

3) You've got a bit of lateral color fringing in your stars, probably from atmospheric dispersion given it's an OSC. Try separating your channels and realigning them using StarAlignment's distortion correction function at the very beginning of your processing workflow.

 

4) The most important - if you're not totally happy with your image, get more data! More data will allow you to stretch harder, have smaller stars, do less NR, etc. I spent 4-8 hours per target during my first year. Now I'm disappointed if I get less than 40 hours on a target.

 

Really though - nitpicky feedback, and your pictures are coming along great.


  • sbharrat likes this

#6 sbharrat

sbharrat

    Apollo

  • -----
  • topic starter
  • Posts: 1,217
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 20 October 2021 - 09:23 AM

For having done this for less than a year, these are amazing images. Well done. Your stars are round to the edge of the frame, you've got good detail, no gradients, pictures are calibrated well. 

 

With regards to your request, I'll give you 4 points of feedback. But please know that I really had to nitpick to find issues. You should be proud.

 

1) you're doing too much noise reduction. I see a soft, plasticky kind of look and a mottle in wavelet layers 4-5 at 1:1 scale that comes from going too hard with TGVDenoise. IMO, image looks most pleasing when there's some fine grain noise at 1:1 - the goal being noise *reduction* and not noise *elimination*. You can also try running a masked MultiscaleMedianTransform with denoise enabled on just wavelets 4-5 (again, very light - threshold of 2-5 and amounts set to <0.4) to address some of the inevitable mottling from TGV.

 

2) you're reducing stars a bit too much. Small stars are fine, but I can pick out artefacts around stars if reduced to excess, or if they're added back to a starless image that's been pushed too far without them (starnet is not a particularly clean program). 

 

3) You've got a bit of lateral color fringing in your stars, probably from atmospheric dispersion given it's an OSC. Try separating your channels and realigning them using StarAlignment's distortion correction function at the very beginning of your processing workflow.

 

4) The most important - if you're not totally happy with your image, get more data! More data will allow you to stretch harder, have smaller stars, do less NR, etc. I spent 4-8 hours per target during my first year. Now I'm disappointed if I get less than 40 hours on a target.

 

Really though - nitpicky feedback, and your pictures are coming along great.

Thank you very much for the feedback. Two of these (1,3) I would not have thought about at all. I will look into these in more detail. For #2, I myself am unhappy. Currently I use MorphTransform erosion at the minimum size of 3x3 but this is too much. I do have on my list to incorporate the scripted version of Adam Block's star reduction. 

 

Thanks again. 



#7 terry59

terry59

    Voyager 1

  • *****
  • Posts: 10,552
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 20 October 2021 - 12:31 PM

Some hopefully helpful inputs:

 

#1 has too much blue IMO. I'd dial that channel back some. There is a problem with your star colors that appears to be the result of the channels not well fitted

 

#2 needs a blast of SCNR

 

#3 also too blue.  I see what appears to be Starnet++ artifacts                               


Edited by terry59, 20 October 2021 - 12:34 PM.

  • sbharrat likes this

#8 raguramm

raguramm

    Vostok 1

  • *****
  • Posts: 146
  • Joined: 15 Jan 2019

Posted 20 October 2021 - 02:33 PM

Re: my 2nd point, check out this wonderful tutorial:
https://thecoldestni...star-reduction/

If done right and in moderation, it's the best star reduction method I've found.
  • sbharrat likes this

#9 TimN

TimN

    Soyuz

  • *****
  • Moderators
  • Posts: 3,808
  • Joined: 20 Apr 2008
  • Loc: Ontario, Canada

Posted 20 October 2021 - 04:20 PM

Moved to Beginning DSI



#10 ngatel

ngatel

    Mariner 2

  • -----
  • Posts: 298
  • Joined: 10 Dec 2006
  • Loc: Palm Springs, CA

Posted 20 October 2021 - 04:57 PM

I am in the same place as you. I have the acquisition part down, post-processing is a constant challenge. It takes time. I have a friend who has been imaging for decades and he is expert at it. He reminds me that it has taken him years and years of experience to get to where he is at. So he offers encouragement and tells me that I have come a long way in a year. He also tells me some of his best images took months or even years of work to get perfect. I'm impatient.

 

So I offer you my encouragement.

 

Sounds like raguramm provided some good technical feedback.

 

You have posted 3 images of targets that can be difficult to process, especially the Western Veil. Plus narrowband allows some great data, but then has its own difficulties in processing. It looks like you are only using an L-enhance filter. You might want to add some RGB and do a HaRGB or similar composite.

 

From my own experience, I tend to over-process, too much saturation, too much of a certain color, etc. Small incremental adjustments to be always better. Star reduction done right is difficult I am finding.

 

Add to this, you are using the most complicated, full-featured software available, which has a steep learning curve. Plus each target usually requires a little different strategy in processing. Time and experimentation will guide us in what strategy to employ with each.

 

So keep at it. I like what you have done with the three images — better than what I did with the same targets! Go back and revisit. That is what I am doing a lot of these days.


  • sbharrat likes this

#11 dswtan

dswtan

    Viking 1

  • *****
  • Posts: 941
  • Joined: 29 Oct 2006
  • Loc: Morgan Hill, CA

Posted 20 October 2021 - 06:26 PM

Nice work! The stand-out comment for me is your color choices. If deliberate, I respect your choice. If not intentional, the too-blue elephant and eagle needs some color work, and the veil is too green (for me). If these are not artistic choices you are making, then perhaps you are using an uncalibrated screen?


  • sbharrat likes this

#12 Mike in Rancho

Mike in Rancho

    Surveyor 1

  • -----
  • Posts: 1,624
  • Joined: 15 Oct 2020
  • Loc: Alta Loma, CA

Posted 20 October 2021 - 06:37 PM

Tried to graduate yourself over to expert before we sent you over there, eh?  lol.gif  Hey, even Jonny's still here.  And I'll probably never get over there.

 

Well, I can't help you with any PI techniques, but for matters of general appearance -- what were you going for?  And remember that different targets can have quite different processing needs.

 

I like the Elephant - dark, but rather spooky.  I don't think it works well with the bright border, but when opened with a darker border, it really has a great look to it.  For that target I'm ok with the softening and star reduction, but the stars have some color errors that should probably be looked at.

 

For my taste, the Veil could be a bit sharper, 52 Cyg brought under control, and here I think the anti-star look has made the image lose it's pop.  Seems dead.  Maybe just some size reduction instead of dimming, show the little stars more as bright pinpoints?

 

Same with the Eagle, though here I am totally fine with going smoothed-over at least for the outer nebulosity.  The colors aren't quite working for me on this target though, meaning they are somewhat distracting for one, and not enhancing any details for another.  Being a narrowband palette, the colors and mapping are entirely up to you.  Might be able to find a combo that gels a little better, and also helps reveal the differing structures more?

 

But again those would just be my personal goals, based on my own personal tastes and processing guidelines.  Key is to develop the looks you want, while learning all the complex ways to accomplish your vision in PI.


  • dswtan, jonnybravo0311 and sbharrat like this

#13 sbharrat

sbharrat

    Apollo

  • -----
  • topic starter
  • Posts: 1,217
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 20 October 2021 - 07:32 PM

Nice work! The stand-out comment for me is your color choices. If deliberate, I respect your choice. If not intentional, the too-blue elephant and eagle needs some color work, and the veil is too green (for me). If these are not artistic choices you are making, then perhaps you are using an uncalibrated screen?

I am working on a screen calibrated but only with the low end Spyder device. I have to check it on some other screens to figure out if I just unconsiously like too much blue or if the particular screen I am using is far off. 

 

Thanks for the feedback...


  • dswtan likes this

#14 sbharrat

sbharrat

    Apollo

  • -----
  • topic starter
  • Posts: 1,217
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 20 October 2021 - 07:55 PM

Inline...

 

Tried to graduate yourself over to expert before we sent you over there, eh?  lol.gif  Hey, even Jonny's still here.  And I'll probably never get over there.

 

SJB> Hey! I tried the soliciting image advice here several times already... Figured I was read for some more heavy-handed smacking. 

 

Well, I can't help you with any PI techniques, but for matters of general appearance -- what were you going for?  And remember that different targets can have quite different processing needs.

 

I like the Elephant - dark, but rather spooky.  I don't think it works well with the bright border, but when opened with a darker border, it really has a great look to it.  For that target I'm ok with the softening and star reduction, but the stars have some color errors that should probably be looked at.

 

SJB> Creating HOO from OSC/dualband is definitely something I can't do right. The star technique here was remove stars before splitting channels to make it HOO and then add back the stars. When you say star color errors, do you mean too many blue stars to be right or multiple colors within a star? 

 

 

For my taste, the Veil could be a bit sharper, 52 Cyg brought under control, and here I think the anti-star look has made the image lose it's pop.  Seems dead.  Maybe just some size reduction instead of dimming, show the little stars more as bright pinpoints?

 

SJB> I have just been reading on shapening. Nothing like that on this image (didn't know how to... only now starting to see it).

Problem with stars is coming across as a common refrain. I do want to go look at the star reduction technique Raguramm pointed to. If it is one thing that came out of this exercise, seems like everyone agrees that could use work. 

 

Same with the Eagle, though here I am totally fine with going smoothed-over at least for the outer nebulosity.  The colors aren't quite working for me on this target though, meaning they are somewhat distracting for one, and not enhancing any details for another.  Being a narrowband palette, the colors and mapping are entirely up to you.  Might be able to find a combo that gels a little better, and also helps reveal the differing structures more?

 

SJB> More trouble with color frown.gif Obviously I have to boost the OIII coming from L-eNhance green to get it anywhere near the Ha. Seems like I am way overdoing it. I don't know if I just have an unconscious liking to over-blue or if my Spyder calibrated monitor is just far off.

 

But again those would just be my personal goals, based on my own personal tastes and processing guidelines.  Key is to develop the looks you want, while learning all the complex ways to accomplish your vision in PI.


  • Mike in Rancho likes this

#15 ngatel

ngatel

    Mariner 2

  • -----
  • Posts: 298
  • Joined: 10 Dec 2006
  • Loc: Palm Springs, CA

Posted 20 October 2021 - 09:41 PM

I am working on a screen calibrated but only with the low end Spyder device. I have to check it on some other screens to figure out if I just unconsiously like too much blue or if the particular screen I am using is far off.


Everyone tells me my stars are greenish. I don’5 see that. My x-rite calibrator should be delivered tomorrow. I hope that is the problem and not my old eyes.

#16 mayhem13

mayhem13

    Viking 1

  • -----
  • Posts: 775
  • Joined: 10 Jan 2021
  • Loc: New Jersey

Posted 21 October 2021 - 06:03 AM

Fantastic images given your time in the sport!

 

My only critique would be to graduate to a mono camera now…..it appears to me that you’ve pretty much peaked with your gear. I’m just starting with the 533 a few weeks ago and I’m already jealous of the the mono images on Astrobin.



#17 sbharrat

sbharrat

    Apollo

  • -----
  • topic starter
  • Posts: 1,217
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 21 October 2021 - 07:59 AM

Fantastic images given your time in the sport!

 

My only critique would be to graduate to a mono camera now…..it appears to me that you’ve pretty much peaked with your gear. I’m just starting with the 533 a few weeks ago and I’m already jealous of the the mono images on Astrobin.

Thanks. I am sure at some point I will move to mono. But right now it seems it would be a "lazy" upgrade. I know I have a way to go in learning processing (star quality and coloring seems to be a common complaint) that I expect will help me in general. Besides, my $$/image was only now starting to come down to reasonable numbers... lol.gif



#18 ChiTownXring

ChiTownXring

    Viking 1

  • *****
  • Posts: 778
  • Joined: 12 Aug 2020
  • Loc: Sugar Grove, IL

Posted 21 October 2021 - 02:14 PM

Its almost like we are both watching the same tutorials for processing.. Who are you following if anyone?


  • sbharrat likes this

#19 Mike in Rancho

Mike in Rancho

    Surveyor 1

  • -----
  • Posts: 1,624
  • Joined: 15 Oct 2020
  • Loc: Alta Loma, CA

Posted 22 October 2021 - 12:51 AM

I was just goofing, Shaun!  lol.gif

 

HOO can indeed be a tough balance, as I am learning myself.  Tulip was a fail for me on that, possibly a lot to do with the pesky Hb in my L-eNhance.  The Pacman allowed for a lot more balancing practice, however, as there is plenty to be found -- if you just get the Ha out of the way.  I think I was able to do okay on it for my challenge entry, which was more wide field, but I struggled trying to do close-up crops.  The blue stuff just seems to come with more noise, and I'm not skilled enough yet to know how to handle it.

 

Did you have an entry on that?  If not, I would seek some duoband data on that as it would make great practice for you.  I can link mine up in the discussion thread if you'd want to download it.  Only 8 hours worth, but there's for sure Ha and OIII to be had.  This month has been a lot of L-eNhance learning for me.

 

Bicolor stars can be tricky.  Most are going to show a lot stronger in the blue-green than the red.  I have a few that are both, but the colors tend to be concentric.  Possibly something with my acquisition using the doublet.  Some of that artifact I can deal with using highlight repair in ST's color module.  I'm sure PI has 5 or 6 options for that?  Mostly what I was seeing in the Elephant though were a number of blue stars that had a small but bright red appendage on them.  I suppose they could be actual double stars, I didn't check.  Seemed maybe too many to be so, though.

 

I don't do any star destroying myself, but it seems to have found its way into many workflows now as a normal part of things.  Some techniques can dim them too much for my taste.  If I can't otherwise back that off, I'll take an interim-saved image and poke them back through using a star mask in Layers, with a very slight fuzz for blending.  That said, what to do with stars seems to very much be a personal taste thing, and also may vary depending on the target and what you want to emphasize and show to your viewers.  I tend to like stars.  My inner galactic protector, I guess. tongue2.gif


  • sbharrat likes this

#20 sbharrat

sbharrat

    Apollo

  • -----
  • topic starter
  • Posts: 1,217
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 22 October 2021 - 10:12 AM

Its almost like we are both watching the same tutorials for processing.. Who are you following if anyone?

When I first did a HOO from OSC, I followed this tutorial

https://www.youtube....V6ObLVRvNk&t=9s

 

However, in general I have been just using the Adam Block Fundamentals to learn various processing and then trying to map them onto a relatively simple workflow (straight OSC): DCR, DBE, PCC, DCONV, MLT(NR), Masked Stretch, Starnet, HT, TGVDN, Curves, +stars, MorphTrans (star reduction). I want to work on better star reduction and also adding some sharpening in there.

 

The HOO attempts is really new for me and I am not happy with that yet so still trying to experiement. For that, I split the channels after the masked stretch+starnet, (try to equalize), recombine. 



#21 sbharrat

sbharrat

    Apollo

  • -----
  • topic starter
  • Posts: 1,217
  • Joined: 28 Nov 2020
  • Loc: NJ, USA

Posted 22 October 2021 - 06:54 PM

Inline...

 

I was just goofing, Shaun!  lol.gif

 

HOO can indeed be a tough balance, as I am learning myself.  Tulip was a fail for me on that, possibly a lot to do with the pesky Hb in my L-eNhance.  The Pacman allowed for a lot more balancing practice, however, as there is plenty to be found -- if you just get the Ha out of the way.  I think I was able to do okay on it for my challenge entry, which was more wide field, but I struggled trying to do close-up crops.  The blue stuff just seems to come with more noise, and I'm not skilled enough yet to know how to handle it.

 

Did you have an entry on that?  If not, I would seek some duoband data on that as it would make great practice for you.  I can link mine up in the discussion thread if you'd want to download it.  Only 8 hours worth, but there's for sure Ha and OIII to be had.  This month has been a lot of L-eNhance learning for me.

 

SJB> I only have 2h on that. Hope that I will be able to add at least another 4-6. From my Bortle 8, anything less that is so noisy that I generally don't bother creating a final image.

 

Bicolor stars can be tricky.  Most are going to show a lot stronger in the blue-green than the red.  I have a few that are both, but the colors tend to be concentric.  Possibly something with my acquisition using the doublet.  Some of that artifact I can deal with using highlight repair in ST's color module.  I'm sure PI has 5 or 6 options for that?  Mostly what I was seeing in the Elephant though were a number of blue stars that had a small but bright red appendage on them.  I suppose they could be actual double stars, I didn't check.  Seemed maybe too many to be so, though.

 

SJB> Hmm. I didn't notice this. I have to go look at this more closely, especially the original versus the jpg. 

 

I don't do any star destroying myself, but it seems to have found its way into many workflows now as a normal part of things.  Some techniques can dim them too much for my taste.  If I can't otherwise back that off, I'll take an interim-saved image and poke them back through using a star mask in Layers, with a very slight fuzz for blending.  That said, what to do with stars seems to very much be a personal taste thing, and also may vary depending on the target and what you want to emphasize and show to your viewers.  I tend to like stars.  My inner galactic protector, I guess. tongue2.gif

 

SJB> I do like to *slightly* deemphasize the stars. What I know how to do is make everything dimmer. But what I really want to do is to just remove some of the smaller ones. Something that I need to read through some of the suggestions and try.

 


  • Mike in Rancho likes this

#22 belliott4488

belliott4488

    Ranger 4

  • *****
  • Posts: 388
  • Joined: 05 Oct 2020
  • Loc: MD, US

Posted 22 October 2021 - 08:42 PM

Wow - you're farther along than I am (I'm not yet past the basics), but I'll be ecstatic if I can ever produce images like this!

 

I especially liked the Elephant's Trunk; it's got a very unique feel among all the images I've seen. I expected that some people might object to the blue cast to the background, but I came around to really liking it. For one thing, it looks pretty natural for any of us who have anything brighter than Bortle 2 skies, but more than that, I really like the way it sets off the reddish dust lanes in the foreground.

 

I also find the overall darkness of the image very appealing. I think most of us automatically try to brighten our targets to bring out the detail and enhance the contrast and drama, but you still have plenty of detail, and the dark appearance just seems appropriate for a target that is, after all, an object in the night sky.

 

This is one of my favorite images I've seen - congratulations!


  • Mike in Rancho and sbharrat like 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