Milky Way images?

General discussion about StarTools.

Milky Way images?

Postby 67ssdan » Fri Mar 30, 2018 1:14 pm

Hi, new user here. I had a quick question before I started messing with a demo and all that. Can StarTools be used to process Milky Way images, where there are foreground elements present? It seems like all the examples I was seeing were of deep space type images.

Thank you!

Dan
67ssdan
 
Posts: 10
Joined: Thu Mar 29, 2018 12:23 pm

Re: Milky Way images?

Postby admin » Tue Apr 03, 2018 7:18 am

Hi Dan,

You can indeed use StarTools for Milkyway images. However be mindful of the foreground objects when using Wipe (for light pollution removal) especially. Mask out the foreground objects first (you can use the lasso tool for that). If you have any trouble, feel free to upload your data to Google Drive, Dropbox, One Drive, etc. if you'd like to share it here.

It is possible StarTools may be overkill for your purpose. The GIMP could possibly be a more cost-effective (free) solution for your needs. Much depends on how deep you are going and/or whether there is any significant magnification. When you start stacking multiple images and get into faint nebulosity territory, StarTools will definitely have significant advantages.

However if you are more into landscape astrophotography (which usually entails very wide fields) a traditional photo-processing package may be more suited.

Let us know and wishing you clear skies!
Ivo Jager
StarTools creator and astronomy enthusiast
User avatar
admin
Site Admin
 
Posts: 1717
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne

Re: Milky Way images?

Postby 67ssdan » Tue Apr 03, 2018 10:53 pm

Thanks for the reply!

To be honest, I've seen a LOT of people that seem to get way more color out of their Milky Way shots then I seem to be able to, so I've sort of been on a mission to figure that out. I stumbled on *he who shall remain nameless* blog, and while he gets spectacular results, I find his process to be extremely buggy if not antiquated (though it still works for him). Anyway... digging for gold in a coal mine!

Thanks again,

Dan
67ssdan
 
Posts: 10
Joined: Thu Mar 29, 2018 12:23 pm

Re: Milky Way images?

Postby admin » Wed Apr 04, 2018 6:32 am

67ssdan wrote:Thanks for the reply!

To be honest, I've seen a LOT of people that seem to get way more color out of their Milky Way shots then I seem to be able to, so I've sort of been on a mission to figure that out.


Happy to help if I can! The MW is a very colorful place and "correct" coloring can tell you so much about an object (or region) - its past, future, composition, temperature, etc.

If you could upload some data somewhere I'd be happy to have a look. It could have something to do with pre-processing, or it might be your post-processing.

I stumbled on *he who shall remain nameless* blog

:lol:

, and while he gets spectacular results, I find his process to be extremely buggy if not antiquated (though it still works for him).


If it's the *he who shall remain nameless* that I'm thinking of, last thing I saw was that he was actually on to something for once (but of course, as he does, also claiming he just invented it for the first time).

Non-linear stretching skews color saturation. This then causes bright things to desaturate. Other processing that meddles with brightness (e.g. most processing) contributes to this as well.
There have been various techniques to undo this desaturation. They all invariably involve retaining the color ratios in the linear image (color ratios don't change no matter how bright something is!), and then "transplanting" these to the stretched image. The transplanting part is where methods differ.

It is called "Color constancy" in StarTools and it implements a great number of "transplanting" methods.

This nearly 20 year old article explains the problem/theory/solution;

http://www.allthesky.com/articles/colorpreserve.html

Anyway... digging for gold in a coal mine!


Compress coal hard enough, you might get diamonds... :)
Ivo Jager
StarTools creator and astronomy enthusiast
User avatar
admin
Site Admin
 
Posts: 1717
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne

Re: Milky Way images?

Postby 67ssdan » Wed Apr 04, 2018 9:54 pm

Ah nice, thanks for offering to help. Here's a couple of examples of files I've messed with, and the end result I've gotten, but feel there is more there.

File 1: https://drive.google.com/file/d/1DFuET2 ... sp=sharing
My result (sky): https://flic.kr/p/25iX1Kw

File 2: https://drive.google.com/file/d/1nQ-_ua ... sp=sharing
My result: https://flic.kr/p/XVGSCE

Both are created using Sequator to stack multiple images (F1 15 images from Canon 60D, ISO 4000, F2 3 images from Canon 6D, ISO 6400 I *think*). Happy to share anything else that would be useful.

Thanks again,
Dan
67ssdan
 
Posts: 10
Joined: Thu Mar 29, 2018 12:23 pm

Re: Milky Way images?

Postby admin » Thu Apr 05, 2018 12:03 am

Hmmm... looking at this stack it's almost like you stacked JPEGs?

The image looks stretched, filtered and stars look devoid of color already....

Could you give some more details about how you acquired this and how you stacked this?

Is there anything you did to the CR2 files before getting them off your camera and handing them to your stacker?

Stars look over-exposed, especially in the 2nd stack. The ISO setting is a bit too high - try (from memory) ~640 for a 60D.

ISO is not really a thing any more with digital SLRs - the sensor has only one sensitivity at its "native" ISO. Try finding that native ISO. To emulate other ISOs, DSLRs simply multiply the recorded signal by a factor (also multiplying the recorded noise, so you don't get extra "magic" noise benefits from higher ISOs). For any other ISO than the native ISO, DSLRs may clip the signal, throw bits away or even apply unwanted noise reduction (looking at you, Sony :evil:).
Ivo Jager
StarTools creator and astronomy enthusiast
User avatar
admin
Site Admin
 
Posts: 1717
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne

Re: Milky Way images?

Postby 67ssdan » Thu Apr 05, 2018 2:47 am

Interesting. I've been trying to make heads or tails of the whole ISO thing. Some of what I've read lines up with what I've seen personally in messing with images, and some of it seems to not. How do you find the native ISO? At 640 ISO, I wouldn't see any stars at all hardly, and definitely not the MW. Sounds like you're expecting that I'd be tracking the sky. Is that really what I should be doing for the best results?

In thinking about it, the second one may have been stacked JPGs. I'm trying to remember if I had discovered that Sequator stacked TIFs before or after I did those. The first one is definitely stacked TIFs though, and I don't think I did anything to them in post (I've messed with that in the past and found that it works better to not touch them prior... but I'm definitely open to suggestions. There are likely combinations I've overlooked). Would it help to see the pre-stacked images at all? Is there something that stacks better than Sequator? It may be doing some processing (stretching) of it's own... no idea really.

Thanks again for the help, this is awesome!

Dan
67ssdan
 
Posts: 10
Joined: Thu Mar 29, 2018 12:23 pm

Re: Milky Way images?

Postby admin » Thu Apr 05, 2018 4:27 am

67ssdan wrote:Interesting. I've been trying to make heads or tails of the whole ISO thing.

It's incredibly confusing with a lot of incorrect (film-based) information out there.

Your sensor has a single sensitivity, set in stone. It is called quantum efficiency (QE) and is expressed in a %. That percentage is the chance that an incoming photon is converted into an electron (the uncertainty of a photon really exciting an electron is the origin of the shot noise!). The excited electron then goes into your electron well (you have an electron well for every pixel). When the exposure is done, the electrons in every well are counted. This analog count is then converted to a digital number representing brightness levels.

To mimic ISO/sensitivity, DSLRs can do three things;

  • They can count multiple electrons for a single electron to fake higher sensitivity. This fills your electron wells quicker, overexposing your image quicker.
  • They can simply multiply the digital integer number by a factor > 1 to fake higher sensitivity. This creates "steps" in between brightness level numbers (e.g. 0, 1, 2 becomes 0, 2, 4 with values in between 0 and 2, and 2 and 4 simply missing).
  • They can simply multiply the digital integer number by a factor < 1 to fake lower sensitivity. This reduces precision as brightness levels need to be rounded to the nearest integer (e.g. 0, 1, 2, 3, 4 becomes 0, 0, 1, 1, 2, 2).


All methods also increase or decrease the shot noise component proportionally. There is no free lunch here. There is no "free" improved signal created. The higher (or lower) sensitivity is fake.

For astrophotography, keeping your DLSR from doing all of the above is the preferred way to get as much precision as possible into the digital domain. Your stacker and post-processing software can make the most of the recorded signal.
Keeping your DLSR from doing all of the above is accomplished by setting your DLSR to the native ISO - the setting where your DSLR is not forced to do any tricks to mimic any other sensitivity than its native sensor sensitivity.

Sounds like you're expecting that I'd be tracking the sky. Is that really what I should be doing for the best results?


Not at all! I hope the above clears up how this all works.

In thinking about it, the second one may have been stacked JPGs. I'm trying to remember if I had discovered that Sequator stacked TIFs before or after I did those. The first one is definitely stacked TIFs though, and I don't think I did anything to them in post (I've messed with that in the past and found that it works better to not touch them prior... but I'm definitely open to suggestions. There are likely combinations I've overlooked). Would it help to see the pre-stacked images at all? Is there something that stacks better than Sequator? It may be doing some processing (stretching) of it's own... no idea really.


I don't know of any camera that outputs TIFFs, so you must have done some sort of conversion somewhere. ;) Don't do any conversion! Use the straight CR2 files your camera produces.
Deep Sky Stacker (also free) is a popular way to stack images. It should work well with your 60D.

For specific settings for StarTools, have a look at this tutorial by a long-time StarTools user.

Hope this helps!
Ivo Jager
StarTools creator and astronomy enthusiast
User avatar
admin
Site Admin
 
Posts: 1717
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne

Re: Milky Way images?

Postby 67ssdan » Fri Apr 06, 2018 12:46 pm

Okay wow, this has taken a while to process. I'm not sure I still get it, but, I'll come back to that later. LOL For dummies though, how do I find this native ISO? I've rented the Canon 6D Mark II to take to Japan next week, and I'd love to get all this right if possible.

I don't know of any camera that outputs TIFFs, so you must have done some sort of conversion somewhere. ;) Don't do any conversion! Use the straight CR2 files your camera produces.


Oh! I didn't even try RAW files in there! Okay, so quick question. When I use RAW files, it spits out a FITS file (also does this with TIFF) that is just a sliver of the full image. When I flip back to importing TIFF files, it works as expected, without changing anything in the settings.

Example: https://drive.google.com/file/d/1eXSJHsg55-JW6GBWt_Hb6GNc_xCFdhQO/view?usp=sharing

Thoughts on what I'm doing wrong there?

Also, thanks a ton for all the info and the link. That link will be most helpful when I get over this current hurtle.

Dan
67ssdan
 
Posts: 10
Joined: Thu Mar 29, 2018 12:23 pm

Re: Milky Way images?

Postby 67ssdan » Fri Apr 06, 2018 5:17 pm

Never mind! I should have Googled before asking. Turned out the stable versions of DSS don't like some Canon RAW files, so you have to use the beta. Moving on now, I'm sure I'll have more questions in a bit.

Dan
67ssdan
 
Posts: 10
Joined: Thu Mar 29, 2018 12:23 pm

Next

Return to General Discussion

Who is online

Users browsing this forum: AstroOlly, Bing [Bot] and 2 guests