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

M81/M82 with Startools log and link to raw stack

  • Please log in to reply
20 replies to this topic

#1 dx_ron

dx_ron

    Ranger 4

  • *****
  • topic starter
  • Posts: 341
  • Joined: 10 Sep 2020
  • Loc: SW Ohio

Posted 17 May 2021 - 11:21 AM

I've reprocessed this many times now, to the point where I'm never sure how truly different the results are. The real answer of course is "more data" - especially from my "bright suburban" yard - but by the time clear moonless returns this will not be a viable target. Longer focal length would help also, but when your only tool is a hammer...

 

120 x 45" + 270 x 60" = 6 hours total; with the gear in my signature

 

Preprocessed and stacked in Siril. 40 flats, 40 dark-flats, 40 darks for 45", 55 darks for 60", everything at -10C

 

I had a lot of trouble with the gradient, played around quite a while in the Wipe module before deciding it was as good as I was going to get. Binned to 50% after Wipe. I also messed around with color balance a lot. If anyone plays with the data I'd be really interested to see what you choose.

 

Attached File  StarTools.log.txt   6.6KB   8 downloads

 

 

Link to unstretched fit on dropbox Hopefully someone will tell me if that does not work

 

M81 4-12 4-14 4-16 5-01 2021 ST noNR 1600x1033
 

 

 

 


  • jeremiah2229 and Mike in Rancho like this

#2 copper280z

copper280z

    Vostok 1

  • -----
  • Posts: 186
  • Joined: 01 Mar 2019
  • Loc: Rochester, NY

Posted 17 May 2021 - 11:43 AM

Here's a take. I feel like I'm never totally happy with the color that I get out of startools, in spite of spending a lot of time playing with settings in the color module.

 

M81_4-12_4-14_4-16_5-01_ST.jpeg


Edited by copper280z, 17 May 2021 - 11:43 AM.

  • Thomas Marshall and pyrasanth like this

#3 pyrasanth

pyrasanth

    Skylab

  • *****
  • Posts: 4,213
  • Joined: 08 Jan 2016

Posted 17 May 2021 - 12:19 PM

Thats's not a bad image- colour is always a matter of taste- so its very much our own interpretations that matter.

 

I had a look at your image- I toned the red down slightly to give the blue a bit more of a showing and ran an enhancement on the image to see if I could bring out any more detail however since I'm using the forum Jpeg its not as good as it could be but there is more detail to bring out. I imaged this previously with my RASA V2 and the QHY168C and the ASI6200 to produce this image https://www.astrobin...c4mejv/?nc=user

 

I have attached your image with my take on the enhancements I did.

 

All in all you have a real nice image- so be proud of that achievement. The thing that is good is that you can always go back & revisit with new equipment & techniques latter.

 

M81-M82_REWORK.jpeg


  • RedLionNJ, Thomas Marshall, francov and 1 other like this

#4 bobzeq25

bobzeq25

    ISS

  • *****
  • Posts: 24,642
  • Joined: 27 Oct 2014

Posted 17 May 2021 - 12:50 PM

Here's my take, a quick hack job in PixInsight.  Click on it for a better view.

 

A significant problem is lack of dithering, that meant the black level had to be raised some, or the background was very noisy (fixed pattern noise).

 

I recommend Astro Pixel Processor for beginners.  Gets fairly close to PixInsight, with much less time and effort.

 

anothers M81_82.jpg


Edited by bobzeq25, 18 May 2021 - 12:28 PM.


#5 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,243
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 17 May 2021 - 12:59 PM

And mine with Startools and some tweaking in GIMP

 

Click the link below to the png file, much better then the jpg

 

https://www.cloudyni...911_4033516.png

 

More color in the stars, maybe a bit too much, yellowish core and grey blue spiral arms....

Also quite some detail in M81, nice data

Attached Thumbnails

  • M81M82.jpg

Edited by F.Meiresonne, 17 May 2021 - 01:07 PM.

  • DeanCN, dx_ron and Mike in Rancho like this

#6 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,243
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 17 May 2021 - 01:12 PM

Startools log

 

first thing i do is binning actually , then autodev-crop-wipe

Attached Files


Edited by F.Meiresonne, 17 May 2021 - 01:14 PM.

  • Mike in Rancho likes this

#7 dx_ron

dx_ron

    Ranger 4

  • *****
  • topic starter
  • Posts: 341
  • Joined: 10 Sep 2020
  • Loc: SW Ohio

Posted 17 May 2021 - 01:39 PM

Here's my take, a quick hack job in PixInsight.  Click on it for a better view.

 

There's something wrong with the data, it showed up as having quantization error, as if it had been converted to 8 bits at some point.

 

A significant problem is lack of dithering, that meant the black level had to be raised some, or the background was very noisy (fixed pattern noise).

 

I recommend Astro Pixel Processor for beginners.  Gets fairly close to PixInsight, with much less time and effort.

 

attachicon.gifanothers M81_82.jpg

Can you please explain or show how/where you see it as having quantization error? (you started with the .fit, not the .tiff, right?) The .fit file is straight out of Siril stacking - if there's some weirdness there then it would affect all of my projects.

 

The subs were dithered every 2 frames, for both the 45" and 60" exposures.



#8 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,243
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 17 May 2021 - 01:43 PM

I dither to avoid walking noise

 

I saw no walking noise in the .fit stack...so imho the dithering is ok..



#9 dx_ron

dx_ron

    Ranger 4

  • *****
  • topic starter
  • Posts: 341
  • Joined: 10 Sep 2020
  • Loc: SW Ohio

Posted 17 May 2021 - 01:56 PM

And mine with Startools and some tweaking in GIMP

 

Click the link below to the png file, much better then the jpg

 

https://www.cloudyni...911_4033516.png

 

More color in the stars, maybe a bit too much, yellowish core and grey blue spiral arms....

Also quite some detail in M81, nice data

I like the look of your version, Freddy. Will try to spend some time tonight going over your Startools log to see what I can learn. What did you use for the initial white sample? I tried using most of M81, didn't care for the result (I think that was one of my very pink versions). The version I posted started with a stars mask, then I manually unmasked the big blown-out stars as best I could, then sampled with that mask and tweaked from there. It started with too much green, so I toned down the green and then toned down just a bit the red.

 

I particularly like how you retained a wisp of Holmberg IX, which I had lost in my stretch.



#10 F.Meiresonne

F.Meiresonne

    Fly Me to the Moon

  • *****
  • Posts: 7,243
  • Joined: 22 Dec 2003
  • Loc: Eeklo,Belgium

Posted 17 May 2021 - 02:06 PM

Yes , looking at the log if you have time, cause i allready forgot a bit which steps i took.

 

I did brighten M81 in Superstructure by giving it a mask, made is stand out more, if i recall right...

 

In the HDR module i used reveal core, i used it twice i believe to try to get out more structure in the galaxie...

 

I also colored the stars with a separate mask because initially in the color module not that much color came out in the stars.

 

The trouble was then a bit the bloated stars,i have that in my processing too, tried via GIMP to reduce that a bit...


  • dx_ron likes this

#11 Mike in Rancho

Mike in Rancho

    Viking 1

  • -----
  • Posts: 914
  • Joined: 15 Oct 2020
  • Loc: Alta Loma, CA

Posted 17 May 2021 - 04:48 PM

I don't think your processing was bad at all, looked pretty good in the linked larger tiff.

 

Here's my go at it plus the log, mask coding deleted.  I didn't really like the colors even after trying to balance them out.  If I got the galaxy where I wanted, I'd lose the stars.  It's been a while since I've wanted/needed to color process stars and target separately.  Maybe it's just what was captured, I don't know, but it seemed heavy on the red/pink, even after handling the greens.  All in all pretty cool though.

 

I was also going to ask about your stacking and options therein, then saw you mentioned Siril which I know nothing about.  However, do you know if Siril "molested" your data prior to Startools?  Ivo probably has some tips on that.

 

I also don't know if this is related to what Bob is talking about, but your final stacked fits file is far smaller than I would expect for your camera resolution.  Like, maybe 1/2 to 1/3 the normal size.  My D5300 is something like 24.x mp, and I get stacks around 300mb, either tiff or fits 32 bit.

 

Attached File  StarTools dx ron m81.txt   8.43KB   2 downloads

 

gallery_345094_15786_138818.jpg


  • DeanCN likes this

#12 dx_ron

dx_ron

    Ranger 4

  • *****
  • topic starter
  • Posts: 341
  • Joined: 10 Sep 2020
  • Loc: SW Ohio

Posted 17 May 2021 - 07:48 PM

[EDIT for those reading this thread later - if you are using a 12-bit native camera (183C here), histograms like the one here are no cause for alarm]

 

I went back to Siril and opened a single sub .fits file. Doing nothing at all to it other than an 'auto' histogram transformation gives the same thing:

histo.jpg

 

So it's not anything Siril did during pre-processing or stacking. 

 

I don't see anything unexpected in the INDI driver settings:

Image_info.jpg

Image_settings.jpg

 

What actually should be expected? The QHY183C is a 12-bit camera (4096 levels). It upscales to 16-bit by multiplying by 16, so only 1/16th of all of the 65,536 possible 16-bit values should be populated? Or does that misrepresent how the upscaling works?


Edited by dx_ron, 18 May 2021 - 12:41 PM.


#13 DeanCN

DeanCN

    Mariner 2

  • -----
  • Posts: 209
  • Joined: 13 Sep 2011
  • Loc: Dundas, ON

Posted 17 May 2021 - 09:57 PM

really like your image: a lot can be done with the .fit file .  Here is what I get with StarTools.

I couldn't completely smooth out the gradient with wipe so I ended up pushing the background a bit dark

and then brought it back up in gimp with minor background exposure adjust.  Filter module with a lasso masked

separately on each galaxy helped boost some more color.

 

gallery_202569_16227_246916.png

 


  • Mike in Rancho likes this

#14 Mike in Rancho

Mike in Rancho

    Viking 1

  • -----
  • Posts: 914
  • Joined: 15 Oct 2020
  • Loc: Alta Loma, CA

Posted 18 May 2021 - 04:27 AM

really like your image: a lot can be done with the .fit file .  Here is what I get with StarTools.

I couldn't completely smooth out the gradient with wipe so I ended up pushing the background a bit dark

and then brought it back up in gimp with minor background exposure adjust.  Filter module with a lasso masked

separately on each galaxy helped boost some more color.

 

 

That's really sharp.  I think you've got the cleanest of that little blue-red double at about 10 o'clock from M81, though Ron's is pretty good too.

 

And I always forget that Filter can do more than fringe kill and reject.



#15 DeanCN

DeanCN

    Mariner 2

  • -----
  • Posts: 209
  • Joined: 13 Sep 2011
  • Loc: Dundas, ON

Posted 18 May 2021 - 07:32 AM

it was late and I was a bit lazy ... here is the log for anyone interested.

There are 2 sections where I masked off M81 and picked a yellowish pixel to boost

the blue spirals/plane (saturate visual H-beta/O-III filter) and then M82 to pick a bluish

pixel in the edge-on galaxy plane to try and boost the red tendrils/knots (saturate visual H-alpha).

In Color module I always create a star mask, then click Sample to initially set the bias levels,

then I delete that mask and create a full mask for the remaining adjustments.

File loaded [/Downloads/unstretched_stack_M81_4-12_4-14_4-16_5-01_ST.fit].
Image size is 5544 x 3684
--- 
Type of Data: Linear and was Bayered, but not whitebalanced
--- Auto Develop
Parameter [Ignore Fine Detail <] set to [Off]
Parameter [Outside RoI Influence] set to [15 %]
Parameter [RoI X1] set to [0 pixels]
Parameter [RoI Y1] set to [0 pixels]
Parameter [RoI X2] set to [5544 pixels (-0)]
Parameter [RoI Y2] set to [3684 pixels (-0)]
Parameter [Detector Gamma] set to [1.00]
Parameter [Shadow Linearity] set to [50 %]
--- Rotate
Parameter [Angle] set to [1.0]
--- Bin
Parameter [Scale] set to [scale 50.00% / +2.00 bits / +1.41x SNR improvement]
Image size is 2803 x 1890
--- Crop
Parameter [X1] set to [206 pixels]
Parameter [Y1] set to [111 pixels]
Parameter [X2] set to [2733 pixels (-70)]
Parameter [Y2] set to [1807 pixels (-83)]
Image size is 2527 x 1696
Mask used (BASE64 PNG encoded)
--- Wipe
Parameter [Synthetic Dark/Bias] set to [Off]
Parameter [Gradient Edge Behavior] set to [Absorb 50%]
Parameter [Synthetic Flats] set to [Off]
Parameter [Sampling Precision] set to [256 x 256 pixels]
Parameter [Dark Anomaly Filter] set to [2 pixels]
Parameter [Gradient Falloff] set to [0 %]
Parameter [Synth. Bias Edge Area] set to [100 %]
Parameter [Gradient Aggressiveness] set to [88 %]
Mask used (BASE64 PNG encoded)
Redoing stretch of linear data
--- Auto Develop
Parameter [Ignore Fine Detail <] set to [2.4 pixels]
Parameter [Outside RoI Influence] set to [9 %]
Parameter [RoI X1] set to [1047 pixels]
Parameter [RoI Y1] set to [740 pixels]
Parameter [RoI X2] set to [1403 pixels (-1124)]
Parameter [RoI Y2] set to [985 pixels (-711)]
Parameter [Detector Gamma] set to [1.00]
Parameter [Shadow Linearity] set to [34 %]
--- Contrast
Parameter [Expose Dark Areas] set to [Yes]
Parameter [Brightness Retention] set to [Global Mode Align]
Parameter [Precision] set to [256 x 256 pixels]
Parameter [Shadow Detail Size] set to [10 pixels]
Parameter [Locality] set to [50 %]
Parameter [Shadow Dyn Range Alloc] set to [91 %]
--- HDR
Parameter [Small Detail Precision] set to [Max]
Parameter [Channels] set to [Brightness Only]
Parameter [Algorithm] set to [Reveal DSO Core]
Parameter [Dark/Bright Response] set to [Full]
Parameter [Detail Size Range] set to [596 pixels]
Parameter [Strength] set to [1.4]
--- Wavelet Sharpen
Parameter [Structure Size] set to [Large]
Mask used (BASE64 PNG encoded)
--- SNR-aware Wavelet Sharpening
Parameter [Scale 1] set to [100 %]
Parameter [Scale 2] set to [100 %]
Parameter [Scale 3] set to [100 %]
Parameter [Scale 4] set to [100 %]
Parameter [Scale 5] set to [100 %]
Parameter [Mask Fuzz] set to [4 pixels]
Parameter [Amount] set to [300 %]
Parameter [High SNR Size Bias] set to [85 %]
Parameter [Low SNR Size Bias] set to [0 %]
Parameter [Dark/Light Enhance] set to [50% / 50%]
Mask used (BASE64 PNG encoded)
--- Deconvolution
Parameter [Image Type] set to [Deep Space]
Parameter [Secondary PSF] set to [Dynamic Star Sample Small x Primary]
Parameter [Primary PSF] set to [Moffat Beta=4.765 (Trujillo)]
Parameter [Primary Radius] set to [1.8 pixels]
Parameter [Iterations] set to [11]
Parameter [Error Diffusion] set to [16 %]
Parameter [Mask Fuzz] set to [8.0 pixels]
Parameter [Deringing] set to [50 %]
Mask used (BASE64 PNG encoded)
Mask used (BASE64 PNG encoded)
--- Color
Parameter [Bias Slider Mode] set to [Sliders Reduce Color Bias]
Parameter [Style] set to [Scientific (Color Constancy)]
Parameter [LRGB Method Emulation] set to [Straight CIELab Luminance Retention]
Parameter [Matrix] set to [Identity (OFF)]
Parameter [Dark Saturation] set to [2.6]
Parameter [Bright Saturation] set to [Full]
Parameter [Saturation Amount] set to [283 %]
Parameter [Blue Bias Reduce] set to [1.00]
Parameter [Green Bias Reduce] set to [1.70]
Parameter [Red Bias Reduce] set to [1.22]
Parameter [Mask Fuzz] set to [1.0 pixels]
Parameter [Cap Green] set to [100 %]
Parameter [Highlight Repair] set to [Off]
--- Super Structure
Parameter [Detail Preservation] set to [Linear Brightness Mask Darken]
Parameter [Compositing Algorithm] set to [Multiply, Gamma Correct]
Parameter [Brightness, Color] set to [Process Both]
Parameter [Brightness Retention] set to [Local Median]
Parameter [Mask Fuzz] set to [1.0 pixels]
Parameter [Airy Disk Radius] set to [50 %]
Parameter [Gamma] set to [0.50]
Parameter [Detail Preservation Radius] set to [20.0 pixels]
Parameter [Saturation] set to [50 %]
Parameter [Strength] set to [100 %]
Mask used (BASE64 PNG encoded)
--- Filter
Parameter [Filter Mode] set to [Saturate Visual H-beta/O-III]
Parameter [Sampling Method] set to [3x3 Average]
Parameter [Filter Width] set to [5]
Parameter [Mask Fuzz] set to [1.0 pixels]
Mask used (BASE64 PNG encoded)
--- Unified De-Noise
Parameter [Grain Size] set to [3.0 pixels]
Parameter [Walking Noise Size] set to [1.0 pixels]
Parameter [Walking Noise Angle] set to [0]
Mask used (BASE64 PNG encoded)
--- Shrink
Parameter [Mode] set to [Tighten]
Parameter [Halo Extend] set to [1 pixels]
Parameter [Iterations] set to [10]
Parameter [Regularization] set to [0.85]
Parameter [Color Taming] set to [2 pixels]
Parameter [De-ringing] set to [2.0 pixels]
Parameter [Un-glow Strength] set to [35 %]
Parameter [Un-glow Kernel] set to [30.0 pixels]
--- Unified De-Noise
Parameter [Grain Size] set to [3.0 pixels]
Parameter [Walking Noise Size] set to [1.0 pixels]
Parameter [Walking Noise Angle] set to [0]
--- Unified De-Noise
Parameter [Scale 1] set to [95 %]
Parameter [Scale 2] set to [95 %]
Parameter [Scale 3] set to [95 %]
Parameter [Scale 4] set to [95 %]
Parameter [Scale 5] set to [50 %]
Parameter [Grain Equalization] set to [0 %]
Parameter [Scale Correlation] set to [50 %]
Parameter [Color Detail Loss] set to [28 %]
Parameter [Brightness Detail Loss] set to [30 %]
Parameter [Grain Dispersion] set to [3.0 pixels]
Mask used (BASE64 PNG encoded)
--- Filter
Parameter [Filter Mode] set to [Saturate Visual H-alpha]
Parameter [Sampling Method] set to [3x3 Average]
Parameter [Filter Width] set to [5]
Parameter [Mask Fuzz] set to [1.0 pixels]
Mask used (BASE64 PNG encoded)
File saved [/Downloads/unstretched_stack_M81_4-12_4-14_4-16_5-01_ST.png].

Edited by DeanCN, 18 May 2021 - 07:35 AM.

  • Mike in Rancho likes this

#16 Ivo Jager

Ivo Jager

    Vendor ( Star Tools )

  • *****
  • Vendors
  • Posts: 468
  • Joined: 19 Mar 2011
  • Loc: Melbourne, Australia

Posted 18 May 2021 - 07:52 AM

Depending on the outlier rejection algorithm chose, your stack may not increase in bit-depth from the original bit-depth the individual frames were in.

 

Some outlier rejection algorithms simply pick a value for a pixel from all the sub frames that they think are closest, but don't add/average sub-frames together, therefore never needing to encode more bits than were originally used.

 

For example median stacking (the simplest out of all outlier rejection algorithms) just picks the middle number when sorting all possible values from small to large (if there is no perfect middle one, it averages the two middle values instead, possibly requiring one extra bit to encode the result).

 

I had a quick look at the stack. Some quick notes;

  • Any sign of the gradient should be removed instantly if you use one of the "Uncalibrated" presets. This is, however a very busy area with lots of IFN, and Wipe will be very careful by default.
  • It appears there are some color aberrations in the highlights, with some stars not perfectly aligning across channels. This will affect their appearance, but will also affect the Color module's default settings (as it relies on highlight calibration - if the colors in the highlights are off, the module will give defaults that are off too). If Color Constancy is impossible/undesirable due to the aberrations, you can always consider using the Artistic options for the Style parameter. This conveniently hides the issues in the highlights by mimicking how older software desaturates highlights during processing.
  • Always use MaxRGB mode as a sanity check; in this image with the Color module's default color balance, you will likely find green dominance where none should exist. Balance out the green until it is gone. You can even just click on a green dominant area to make it disappear and re-balance itself.
  • As can be readily seen in most renditions, M81's core should render somewhat yellow and its outer rim more blue/purple. M82 should render a nice white with orange and red (if sufficiently Ha sensitive) matter streams at both its poles.The great juxtaposition in coloring of the pair, is actually really helpful when color balancing, as you have so many features to go by.
  • Star cores seem very well (too well) resolved, yet have massive halos with a "flat" ring. This is somewhat suspect - as if some high dynamic range feature was used. This obviously would make the data non-linear, while also possibly being one of the reasons why channel alignment appears to have been troublesome (causing the color aberrations in the highlights).
  • In AutoDev, set an RoI over M81, as it is the brighter of the pair. The AutoDev documentation actually specifically uses M81/M82 as an example for a scenario with multiple regions of Interest.
  • Use the Super Structure module's default Dim Small preset, to push back any noise on an "empty" background, while retaining all other structures, including Holmberg IX, and perhaps even some more faint IFN (really requires more integration time though).
  • @bobzeq25 is definitely correct that some fixed pattern noise is visible ("zipper" artifacts). Sufficient dithering should have helped this, but I have seen this in other datasets stacked in Siril as well from other instruments (where something like DSS did not produce this issue). Perhaps try a different outlier rejection algorithm or settings? If you cannot get rid of this, the upcoming StarTools 1.8 will be able to mitigate this somewhat with a new filter in the Wipe module, though it is always more ideal to address this during acquisition.

Hope this helps!


  • bobzeq25 and Mike in Rancho like this

#17 dx_ron

dx_ron

    Ranger 4

  • *****
  • topic starter
  • Posts: 341
  • Joined: 10 Sep 2020
  • Loc: SW Ohio

Posted 18 May 2021 - 12:20 PM

I posted my screenshots above on the Siril forum. Cyril believes it is not alarming from a highly stretched 16-bit image.

I am trying to test if there is an advantage to using 32-bit fits files (I can see how that might make a small difference in average stacking) - I haven't yet figured out exactly where I can best perform the conversion(s). [edit - I think this happens only at the time of stacking each master frame, and my version, at least, was defaulted to 16-bit unsigned; working on restacking one of the nights with the 32-bit setting to see what the stretch ends up looking like]

 

Now, if there is some fundamental change I can make to handling the bit-depth, will that result in a significant improvement in the final image? I hope so, because that would be awesome, but I rather doubt it. I think instead that, if the current processing were substantially limited by bit-depth, the current final versions would be pretty crappy.

 

Anyway, this thread is proving to be very instructive. Still thinking about how best to figure out how to approach Ivo's comment about the stars. There is not any intentional 'HDR' being applied during the pre-processing insofar as I can tell. I can try DSS to see if there is any difference, but will need some time to learn that first.

 

I think there are several of us who use the combination of Siril --> StarTools, so these questions are worth exploring.


Edited by dx_ron, 18 May 2021 - 12:31 PM.

  • bobzeq25 likes this

#18 bobzeq25

bobzeq25

    ISS

  • *****
  • Posts: 24,642
  • Joined: 27 Oct 2014

Posted 18 May 2021 - 12:28 PM

OK, I made some kind of mistake.  My 12 bit stack looks exactly the same.  Deleting post #11, and the "quantization" stuff.


Edited by bobzeq25, 18 May 2021 - 12:30 PM.

  • dx_ron likes this

#19 dx_ron

dx_ron

    Ranger 4

  • *****
  • topic starter
  • Posts: 341
  • Joined: 10 Sep 2020
  • Loc: SW Ohio

Posted 18 May 2021 - 12:37 PM

I think our brains naturally smooth out the histograms, and the reality can be alarming :)  I know I was alarmed when you posted that.


  • bobzeq25 likes this

#20 Ivo Jager

Ivo Jager

    Vendor ( Star Tools )

  • *****
  • Vendors
  • Posts: 468
  • Joined: 19 Mar 2011
  • Loc: Melbourne, Australia

Posted 18 May 2021 - 09:22 PM

I posted my screenshots above on the Siril forum. Cyril believes it is not alarming from a highly stretched 16-bit image.

I am trying to test if there is an advantage to using 32-bit fits files (I can see how that might make a small difference in average stacking) - I haven't yet figured out exactly where I can best perform the conversion(s). [edit - I think this happens only at the time of stacking each master frame, and my version, at least, was defaulted to 16-bit unsigned; working on restacking one of the nights with the 32-bit setting to see what the stretch ends up looking like]

 

Now, if there is some fundamental change I can make to handling the bit-depth, will that result in a significant improvement in the final image? I hope so, because that would be awesome, but I rather doubt it. I think instead that, if the current processing were substantially limited by bit-depth, the current final versions would be pretty crappy.

 

Anyway, this thread is proving to be very instructive. Still thinking about how best to figure out how to approach Ivo's comment about the stars. There is not any intentional 'HDR' being applied during the pre-processing insofar as I can tell. I can try DSS to see if there is any difference, but will need some time to learn that first.

 

I think there are several of us who use the combination of Siril --> StarTools, so these questions are worth exploring.

Indeed, nothing too alarming there, though it is good to know this is happening. Particularly if you're going for exceedingly faint stuff this may come into play and you may want to use some stacking method that adds/averages frames, rather than keeping bit-depth as-is.

 

I just re-read your inital post;

 

120 x 45" + 270 x 60"

"Well there's your problem." You are stacking different exposure times, which is essentially doing a defacto HDR combine (you are making the stacker figure out how to combine images shot with two different dynamic ranges).

 

Another thing that can cause this sort of issue that looks somewhat similar, is shooting over multiple nights however with atmospheric conditions deteriorating (or improving) drastically.

 

Hope this helps!


  • dx_ron likes this

#21 dx_ron

dx_ron

    Ranger 4

  • *****
  • topic starter
  • Posts: 341
  • Joined: 10 Sep 2020
  • Loc: SW Ohio

Posted 18 May 2021 - 10:20 PM

Thanks!

 

I reset Siril to produce 32-bit results from average stacking, instead of 16-bit. That did 'smooth' the autostretched histogram quite a bit. How much difference that makes in the final result we shall see. I am going back and re-pre-processing each session that way.

 

Now I have another experiment to add - one combined stack with and one without the 45-second exposures.




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