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

Where do you cut off varying sky brightness?

  • Please log in to reply
8 replies to this topic

#1 Mike in Rancho

Mike in Rancho

    Viking 1

  • -----
  • topic starter
  • Posts: 646
  • Joined: 15 Oct 2020
  • Loc: Alta Loma, CA

Posted 04 May 2021 - 06:14 PM

I got a late start on M106 last night.  Thinking I would be clever, I decided to lead the target into the meridian flip, but Tom Brady I am not.  For whatever reason, probably bad settings, Synscan balked and doing a proper pre-flip for me, so eventually I just did a "manual GOTO," if there is such a thing, using RA and DEC coordinates from Stellarium.  Long story short I was close to meridian at the start, and just kept going to rack up time.

 

This northern part of the sky is pretty good for me, and I was doing 2 minutes at iso400 for 1/4 histogram.  I've been at or over 1/3 before, at iso200, nearer to the celestial equator where I can start edging toward more prominent light domes.

 

I finally cut things off when the brightness started approaching, but had not yet hit, 1/3.  Meridian sub and my last sub shown.

 

M106 BYN sub zenith.jpg

 

M106 BYN sub setting.jpg

 

I haven't tested anything in the stacker yet or run numbers for skyglow percentage to see how drastic the variation is.  I imagine I might have to run background calibration if I keep them all?

 

So, I was wondering, how much brightness variation do you let happen before you call it a night and pack up, or put another way if you kept on going, at what differential would you decide to discard all the subs after that?



#2 happylimpet

happylimpet

    Fly Me to the Moon

  • *****
  • Posts: 5,666
  • Joined: 29 Sep 2013
  • Loc: Southampton, UK

Posted 04 May 2021 - 06:38 PM

Use them all, unless theyre perhaps 5x the background of the best. I know this sounds counterintuitive, but it still helps the image. Try it with and without.

 

Ideally, you could weight them and then the brighter frames get a lower weighting, then its definitely worth keeping absolutely everything.


  • Juno18 and Mike in Rancho like this

#3 jonnybravo0311

jonnybravo0311

    Surveyor 1

  • *****
  • Posts: 1,551
  • Joined: 05 Nov 2020
  • Loc: NJ, US

Posted 04 May 2021 - 07:05 PM

Happy Birthday, Mike!

 

Keep them. There's no reason to toss those subs with a bit more sky glow.


  • limeyx and Mike in Rancho like this

#4 Mike in Rancho

Mike in Rancho

    Viking 1

  • -----
  • topic starter
  • Posts: 646
  • Joined: 15 Oct 2020
  • Loc: Alta Loma, CA

Posted 04 May 2021 - 07:41 PM

Use them all, unless theyre perhaps 5x the background of the best. I know this sounds counterintuitive, but it still helps the image. Try it with and without.

 

Ideally, you could weight them and then the brighter frames get a lower weighting, then its definitely worth keeping absolutely everything.

Thanks.  5x?  Wow.  So that's still adding useful desired signal and not blowing it out with LP?

 

I guess my blue peak only went up about 1/16 histogram when I look at things again, but on-screen it was starting to seem pretty bright and the galaxy shape was hazing over.  I've used far far worse, of course (I'm a beginner!), but I'd kind of like to make this a nice one and I haven't worked anything out yet for keep/reject on skyglow.

 

Happy Birthday, Mike!

 

Keep them. There's no reason to toss those subs with a bit more sky glow.

Thanks.  I may have been overreacting, though still wondering what reasonable "range from optimum" I should be thinking about.

 

And gracias, I guess it is my birthday...in UTC. lol.gif     The fourth with me be not.

 

My brother sent me the Bracken book.  Of course I knew it all already, just from being here on CN. tongue2.gif

 

Flipping through the PI sections made my eyes bug out.  I can't believe you guys are doing all that work!



#5 bobzeq25

bobzeq25

    ISS

  • *****
  • Posts: 24,159
  • Joined: 27 Oct 2014

Posted 04 May 2021 - 10:48 PM

Thanks.  5x?  Wow.  So that's still adding useful desired signal and not blowing it out with LP?

 

I guess my blue peak only went up about 1/16 histogram when I look at things again, but on-screen it was starting to seem pretty bright and the galaxy shape was hazing over.  I've used far far worse, of course (I'm a beginner!), but I'd kind of like to make this a nice one and I haven't worked anything out yet for keep/reject on skyglow.

 

Thanks.  I may have been overreacting, though still wondering what reasonable "range from optimum" I should be thinking about.

 

And gracias, I guess it is my birthday...in UTC. lol.gif     The fourth with me be not.

 

My brother sent me the Bracken book.  Of course I knew it all already, just from being here on CN. tongue2.gif

 

Flipping through the PI sections made my eyes bug out.  I can't believe you guys are doing all that work!

What I do.  Bortle 7.  Close to a big city.

 

Don't image below 45 degrees elevation.

 

Don't try to image galaxies and nebulae when the Moon is anywhere near the target.

 

Brightness variations are not a problem.

 

Bracken's PI processing is kiddy stuff.  <smile>

 

Here's an adult.

 

https://www.amazon.c...h/dp/1138055360

 

And here's a serious pro.  The best noise reduction technique I've ever seen, and I've seen a lot.

 

https://jonrista.com...duction-part-1/

 

Illustrative _small_ quote from the above.

 

"Now here is the critical part about creating a good TGV mask. It needs to be low contrast. We want to have higher protection in brighter areas than in darker areas, however it is critical that darker areas are actually protected fairly heavily. This is where manual masking for TGV differs from normal use…we are applying strong protection, and therefor significant attenuation, of the TGV denoising effect to the darker areas. It’s increased protection is what preserves detail and fine grain, while still allowing significant NR. To modify the mask, you need to perform two steps. First, open CurvesTransformation and reset it. Adjust the black point Output level to somewhere between 0.15 and 0.25, for now start with 0.2. Adjust the white point Output level to somewhere between 0.5 and 0.65, for now start with 0.5. Apply CT to your TGV_mask image.

 

This will brighten the dark parts and darken the light parts, resulting in a lower contrast gray image. We’ve started with 0.2 and 0.5, however in general you will want to tweak these levels depending on the initial contrast of your data…higher contrast data could use more tuning, while lower contrast data could use less tuning."

 

The first time I used the techniques, I took maybe 4 hours.  I've got it down to an hour.  Most of the time.

 

It takes me 10 (simple black and white H alpha) to 50 (multiple filters, both broad and narrowband) hours to process an image.   In my best year, I did 17 (seventeen).

 

The thing about PI.  It does not magically process better.  It provides you a vast array of extremely adjustable tools so that YOU can process better.  It's truthfully said that, if there are two ways to do something, PI has five.  To get the most out of PI, you have to know which of the five to use when.  <smile>  And how to use them, what parameters to adjust how.

 

I generally use a combination of at least three stretching techniques.  Now you have to know in what order to do them.  It varies with the data.

 

Why I recommend Astro Pixel Processor to beginners.  They have more important things to do than to learn to use PI well.


Edited by bobzeq25, 04 May 2021 - 10:52 PM.

  • Juno18, limeyx and Mike in Rancho like this

#6 happylimpet

happylimpet

    Fly Me to the Moon

  • *****
  • Posts: 5,666
  • Joined: 29 Sep 2013
  • Loc: Southampton, UK

Posted 05 May 2021 - 05:56 AM

Thanks.  5x?  Wow.  So that's still adding useful desired signal and not blowing it out with LP?

 

 

Thats been my experience. 5x may not be an exact figure but you'd be surprised. I used to cull a lot of data around dusk/dawn with elevated backgrounds but the re-stack would always have a lower signal/noise. Now I cull a lot less.


  • Mike in Rancho likes this

#7 Mike in Rancho

Mike in Rancho

    Viking 1

  • -----
  • topic starter
  • Posts: 646
  • Joined: 15 Oct 2020
  • Loc: Alta Loma, CA

Posted 05 May 2021 - 05:32 PM

What I do.  Bortle 7.  Close to a big city.

 

Don't image below 45 degrees elevation.

 

Don't try to image galaxies and nebulae when the Moon is anywhere near the target.

 

Brightness variations are not a problem.

 

Bracken's PI processing is kiddy stuff.  <smile>

 

 

Thanks, Bob. waytogo.gif   Are you staying above 45 due to extinction, since you guys are saying brightness variation isn't a problem?  Or just...time and effectiveness?  Yes lately I'll point away from the moon, or, observe or image the moon itself!  Another thing have I tried is the L-eNhance (to strip the red channel), which didn't seem terribly awful under the moon.

 

I don't doubt the complexity and power of PI, what with everyone talking about video series and even seminars.  But even Bracken's dip into the shallow end (it's a Primer, after all) surprised me with the vast array of possibilities and customization.  I was aghast, and the one word that came to my mind was...tedium.  Plus much of it seemed like levels/curves all over again, exactly the kind of PS/Gimp deal I did not want to deal with anymore.  I imagine, like any software, a comfort zone is reached and it becomes second nature.  But it didn't look like what I consider to be fun.

 

 

Thats been my experience. 5x may not be an exact figure but you'd be surprised. I used to cull a lot of data around dusk/dawn with elevated backgrounds but the re-stack would always have a lower signal/noise. Now I cull a lot less.

Interesting.  I always figured I'd be dragging my data into the gutter by including such subs.  But, at least as long as it is weighted or background matched for proper combining or rejection algorithms, the end result is still improved?

 

Now, I'm always reading here how difficult faint galaxies and so forth are in LP, or, as noted above, moonlight glow, as well as how much extra integration is needed.  Is that the key then - a brighter sub away from zenith is still helpful to the stack overall, just not as relatively helpful as getting an extra sub at zenith would be?
 


Edited by Mike in Rancho, 05 May 2021 - 05:33 PM.


#8 bobzeq25

bobzeq25

    ISS

  • *****
  • Posts: 24,159
  • Joined: 27 Oct 2014

Posted 05 May 2021 - 06:06 PM

Thanks, Bob. waytogo.gif   Are you staying above 45 due to extinction, since you guys are saying brightness variation isn't a problem?  Or just...time and effectiveness?  Yes lately I'll point away from the moon, or, observe or image the moon itself!  Another thing have I tried is the L-eNhance (to strip the red channel), which didn't seem terribly awful under the moon.

 

I don't doubt the complexity and power of PI, what with everyone talking about video series and even seminars.  But even Bracken's dip into the shallow end (it's a Primer, after all) surprised me with the vast array of possibilities and customization.  I was aghast, and the one word that came to my mind was...tedium.  Plus much of it seemed like levels/curves all over again, exactly the kind of PS/Gimp deal I did not want to deal with anymore.  I imagine, like any software, a comfort zone is reached and it becomes second nature.  But it didn't look like what I consider to be fun.

 

 

Interesting.  I always figured I'd be dragging my data into the gutter by including such subs.  But, at least as long as it is weighted or background matched for proper combining or rejection algorithms, the end result is still improved?

 

Now, I'm always reading here how difficult faint galaxies and so forth are in LP, or, as noted above, moonlight glow, as well as how much extra integration is needed.  Is that the key then - a brighter sub away from zenith is still helpful to the stack overall, just not as relatively helpful as getting an extra sub at zenith would be?
 

My procedures have a number of advantages in light polluted skies.  One of them is that brightness variations simply cease to be an issue.  I've never gotten anywhere near 2X.  I'm often below 1.1X.

 

Faint anything is tough in moonlight, it's like moving downtown.  <smile>  I do star clusters in moonlight, little else.  Not worth the time.  Ha is possible, but not O(III).

 

One man's tedium is another man's challenge.  <smile>  But I do object to the thought that PI is the ultimate solution for everyone.  For people who don't like standard processing, there's StarTools.  For people who just don't like processing, period, there's EAA.

 

The concepts of levels and curves is integral to stretching in general.  Why I recommend Bracken, for his clear explanations.  PI does offer infinite adjustability, and some interesting variations, like MaskedStretch, and arcsinhStretch.  StarTools has a useful Autodev, like much of StarTools the computer is by your side, helping you.  PI?  You're pretty much on your own.
 


  • Mike in Rancho likes this

#9 Mike in Rancho

Mike in Rancho

    Viking 1

  • -----
  • topic starter
  • Posts: 646
  • Joined: 15 Oct 2020
  • Loc: Alta Loma, CA

Posted 05 May 2021 - 07:32 PM

My procedures have a number of advantages in light polluted skies.  One of them is that brightness variations simply cease to be an issue.  I've never gotten anywhere near 2X.  I'm often below 1.1X.

 

Faint anything is tough in moonlight, it's like moving downtown.  <smile>  I do star clusters in moonlight, little else.  Not worth the time.  Ha is possible, but not O(III).

 

One man's tedium is another man's challenge.  <smile>  But I do object to the thought that PI is the ultimate solution for everyone.  For people who don't like standard processing, there's StarTools.  For people who just don't like processing, period, there's EAA.

 

The concepts of levels and curves is integral to stretching in general.  Why I recommend Bracken, for his clear explanations.  PI does offer infinite adjustability, and some interesting variations, like MaskedStretch, and arcsinhStretch.  StarTools has a useful Autodev, like much of StarTools the computer is by your side, helping you.  PI?  You're pretty much on your own.
 

Ok, I think I get it now.  You don't worry about brightness variations because you avoid them outright.  Fair enough.

 

It's funny, I've always been a PC person, going back to DOS.  So I understand and typically prefer greater control even if it seems arcane to most.  But with images I find I'd much rather have a higher level interface to deal with, only diving down deep when required or wanted.  It's good that there are a multitude of processing platforms to choose from.


  • limeyx 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