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.


How much star reduction/star removal?

  • Please log in to reply
7 replies to this topic

#1 mackiedlm


    Ranger 4

  • -----
  • topic starter
  • Posts: 397
  • Joined: 16 Jul 2019
  • Loc: Galway, Ireland

Posted 08 March 2021 - 04:30 AM

I recently posted an image of IC 405 on the beginning imaging forum. This was an OSC image, without any filters and with significant star reduction carried out using MT in PI.  One of the comments I received was "but most of it is obscured by the stars". I found this intriguing because I had deliberately reduced the stars to the point where I felt they were no longer detracting from (obscuring?) the nebula but were still adding to the overall feel of the image.


This got me to thinking about what is an appropriate level of star reduction. And what about star removal?


I think that the bottom line is that it is totally subjective so I think it would be interesting to hear the views of CN on this question.


I'll start the discussion;


My feeling is that stars are a very important part of any AP image. Yes in areas where there is a heavy starfield they stars can detract from the view of the nebula which is the main subject of the image. And in that situation star reduction to the point where the stars are no longer a distraction is the right way to go. And I think that is a long established principle in AP.


However, I think that with narrow band imaging, and the new dual band and tri band filters used with OSC's, we are becoming conditioned to seeing images with many fewer, largely colourless stars, and and there is a resultant trend towards a level of star reduction which I feel is more than is necessary. Personally I feel that as long as the stars are not detracting from the ability to appreciate the main subject then they are a very important component of a well balanced AP image. If they are reduced to the point where they are barely noticeable I feel the image begins to look lifeless and unexciting. On narrow-band or OSC duo/tri band images I think that more reduction can be applicable because I think that the largely colourless stars need a bit more reduction for them to become less distracting.


Which brings me to star removal. Now, there are many absolutely stunning starless images being produced and I will look at those and appreciate them for what they are. But I almost always find myself feeling that I'd like to see it with some stars as it make the image complete and whole.


So, I'd really like to hear the views of the CN cognoscenti on what level of star reduction you think works best for different types of images and what are your views on completely starless images. Is there a trend, a "fashion", for reducing stars more now than say 5 years ago? If there is, then is that a good thing?


  • PirateMike and F-Astro like this

#2 Epox75



  • -----
  • Posts: 439
  • Joined: 19 Apr 2017
  • Loc: Netherlands

Posted 08 March 2021 - 04:46 AM

I personally like to see narrowband starless pictures, it makes a picture look more like a painting. But even if I like to see them I don't really like to produce them because there's too much modification involved and I feel like it's not real anymore. 


Concerning the star size reduction i've tried it different times in the past but now I don't use it anymore, the deconvolution in Pixinsight with a custom PSF does a good job already and I prefer to be very careful and use masks during the stretch of the image in order to keep a small star size and mantain the color. 

Edited by Epox75, 08 March 2021 - 04:49 AM.

  • PirateMike likes this

#3 terry59


    Voyager 1

  • *****
  • Posts: 10,286
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 08 March 2021 - 07:25 AM

Synthesizing your post, here is an approach...


(NB) When ready for linear stretching, do a mild stretch of each channel, remove the stars, blend, process then add in stars from one of your channels


(WB w/Ha) When ready for linear stretching, do a mild stretch of the Ha channel, remove the stars, process then blend in the starless Ha








Edit: I'm not trendy....too much of that comes and goes



Edited by terry59, 08 March 2021 - 07:27 AM.




  • *****
  • Posts: 9,224
  • Joined: 10 Feb 2014
  • Loc: Springwater, NY

Posted 08 March 2021 - 07:38 AM

You're right; they're wrong. Inexperienced amateurs (and also inexperienced professionals) have a tendency to way overprocess... over-sharp, over-bright, over-contrasty, over-colorized, over-filtered, over-doctored, over-sampled, over-exposed, over-thought, over-forced... It's like the kids getting into mommy's makeup kit for the first time. Or flailing away at dad's drum set and hearing their pounding as the most musical thing since sliced bread. Eventually the clown wears off and most seasoned imagers settle down into subtleties that are well-balanced and truly aesthetic. The longer you've been at this art form... the less and less processing you will favor. I believe what you are discovering is that most often less is more --- and that sometime is now for you! Welcome to the truly expert community! Leave the others behind in the dust with their finger paints. Send out the clowns. Carnegie Hall ain't for everyone.    Tom


~click on~ >>>

Attached Thumbnails

  • 47 carnegie hall 80.jpg

  • AhBok and David I like this

#5 kathyastro



  • *****
  • Posts: 2,983
  • Joined: 23 Dec 2016
  • Loc: Nova Scotia

Posted 08 March 2021 - 08:22 AM

It is definitely a subjective thing.  While starless images are interesting for studying an object in a documentary way, I don't like them esthetically.  On the other hand, I also don't like big, blobby stars that obscure the target.


I reduce my stars even in narrowband images, but I leave them in.  My normal workflow is to separate the stars and process them and the target separately.  But I always add them back in in some form at the end.

  • dswtan, Jon Rista and mackiedlm like this

#6 Jon Rista

Jon Rista


  • *****
  • Posts: 24,803
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 08 March 2021 - 01:52 PM

Personally, I like the aesthetics of keeping stars in the image. I even like it as stars approach whitish when they get very bright...it conveys the intensity. The simple fact of the matter is, stars exist, out there in space, they are a part of it all, and without them...well, something is definitely missing. tongue2.gif


I've seen many images where the intensity of stars is decimated in order to attain "realistic color" ... the star colors usually end up overly saturated, while their differing intensities is lost. So you have this "flat" looking starfield, very richly colored, but you cannot really tell which stars are actually bright vs. those that are fainter. 


With starless images, the star removal process usually has a fairly devastating effect on the nebular details. If you do really wide fields, especially from mosaics, remove the stars, and downsample the image, that destruction of structure can be hidden to a degree, and you can grasp the large scale structure of the nebulas. Still, I have a real hard time destroying data like that. It is an intriguing exercise, and when it is done really well, being able to see the structure of nebula without the stars interfering with that view is even more intriguing. 


In the long run, though, I prefer to keep the stars as much "as they are" as possible. I even approach aligning my data in a very specific way (linear alignment, rather than a linear fit or any other kind of redistributive alignment), so as to portray my stars as naturally as I possibly can. Sometimes, that means some stars end up being pushed to the point of clipping or even a bit beyond...but, in relative terms, said stars ARE bright, often much brighter than most of the other stars in the frame, and I like to see that. They may be white, but at such intensities, our eyes wouldn't pick up much color anyway.


So, I don't do star removal, and I don't do much star reduction. If I do reduction, I try to only do it enough to enhance the relative differences in intensities, so that they can actually be detected by a viewer, and not try to remove or otherwise decimate the appearance of the stars in the image.


For wide fields...where starfields can often become dominating due to the fact that so much light is concentrated into so few pixels. This is where small pixels, even tiny pixels, really SHINE! With big pixels (i.e. 9 micron, 7 micron, even 5-6 micron), at short focal lengths, you concentrate a LOT of signal into just a few pixels, maybe just one pixel for each star. Significant numbers of stars then saturate, bleed, and grow well beyond their natural sizes. This is part of the reason you often see such extreme star reduction in widefield images.


My recommendation is to use a sensor with the smallest pixels you can get your hands on when doing wide field, and to use larger apertures. Large apertures allow higher resolution (small stars), and small pixels allow those stars to be better sampled. This can help avoid the "dominating starfield" problem. 

Edited by Jon Rista, 08 March 2021 - 01:55 PM.

  • dswtan and mackiedlm like this

#7 endless-sky



  • -----
  • Posts: 1,010
  • Joined: 24 May 2020
  • Loc: Padova, Italy

Posted 08 March 2021 - 02:11 PM

To my personal taste, I don't particularly like starless images. Sure, they are nice to look at, once in a while, to appreciate the nebular regions to their fullest, but immediately after I think to myself: "I wish I could see this with stars, too - I am sure it would look even better!"


I do some star reduction in my images, as I am shooting with a DSLR and a mild light pollution filter: I don't get the added benefit of narrowband filters reducing the stars already. When I'll have my monochrome + narrowband filters, my star reduction will likely decrease in strength/iterations.


I feel that too many/big stars detract from the real "star" of the image, while no stars at all feels too unnatural. I prefer a healthy balance of the two.

  • dswtan and mackiedlm like this

#8 Jon Rista

Jon Rista


  • *****
  • Posts: 24,803
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 08 March 2021 - 02:31 PM

For all the complaints they often get...small pixels are, IMO, one of the best things we have ever had for stars. Star size is a combination of factors. Seeing often being one of the biggest, but defocus as well (and sometimes worse than seeing), tracking errors, and of course bloat due to starlight falling on surrounding pixels. When pixels are bigger, that bloat can get very significant very quickly. With smaller pixels, and more specifically with better sampling, as starlight is thrown around by seeing and falls on neighboring pixels, the bloat is far less. 


For those who love stars, but don't want them to dominate the field, embrace small pixels! :p Also, automate your focusing, maybe employ temp compensation for your focusing. Focus is a HUGE factor in achieving the smallest star size possible. Finally, apertures 80mm and smaller will increase diffraction, which will also bloat stars...larger apertures allow smaller stars. There is a balance point, though...much larger than 8-9" and due to what seeing actually does, stars can actually bloat more with really huge apertures. For optimal results, an aperture 8-9" will deliver the tightest stars. For refractors, I have found that at least 100mm will start to minimize diffraction enough. By 150mm a refractor will have minimal diffraction...I've had stars with FWHMs as small as 1.6-1.7" with a 150mm refractor aperture, and 0.82" pixels. Usually with normal seeing, my FWHMs were around 2.1". With a 106mm aperture, I have stars usually around 2.4", with minimums around 1.9". 

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