Back to forums

[RCC] First attempt at M42 with a star tracker

iOptron SkyGuider Pro Request for feedback M42 Messier Objects Emission Nebula
4 replies•303 views
Maxence Ouafik avatar
Maxence Ouafik avatar
Hello ! 

I just finished my first complete workflow using PixInsight with M42 as a target and I hoped to get some constructive feedback on my picture to help me improve my skills in the future  
I just started this hobby and this picture is from my very first subs using a star tracker (the IOptron Skyguider Pro). My setup is still very basic, with a Sony A7II and a cheap(-ish) telephoto zoom lens. 

Acquisition details 
  • 50x60" lights ; 300mm ISO 800 f/6.3
  • 62 bias
  • 25 darks
  • 34 flats


Preprocessing
I followed this very nice tutorial from Light Vortex for the calibration and the stacking of my subs. I used local normalization and a x2 Drizzle. Unfortunately, during the Drizzle part, my l.norm files could not be parsed by PixInsight. I was wondering if anyone had the same issue and if you found how to solve it

Linear processing
For all the processing, I used the book "Inside PixInsight".  
  1. DynamicCrop
  2. DynamicBackgroundExtraction
  3. Deconvolution : this part was trickier than expected. I found very few stars with sufficient amplitude for the PSF and had to use very low values in the deringing. Still after a lot of tickering, I have a small issue with the bright star next to orion with show a small bit of dark halo. I welcome any advice to improve on this part.
  4. PhotometricColorCalibration
  5. Noise reduction through MLT


Non-linear processing
  • HistogramTransformation for the initial stretch followed by MaskedStretch to boost the contrast while keeping background at bay
  • HDRMT to restore the core details with a blend of 60% of the original and 40% of the clone
  • LocalHistogramEqualization in two pass, one for the small structures with a Kernel radius of 50, contrast of 1.5 and amount of .35 and one for the large structure with the same settings except for the Kernel radius, set at 100
  • MLT and UnsharpMask to increase sharpness with a protection mask consisting of luminance - star-mask
  • ColorSaturation to increase the colors in the tint of M42 then curves to boost the saturation as a whole, with a luminance mask to protect the background
  • DarkStructureEnhancer to reveal more of the dark nebulosity
  • ExponentialTransformation : just a small touch to reveal more faint details in the shadows


And you can see the final result here! 

Apart from the obvious "does it look good ?" I mainly interested to know if it's not overcooked, if you've spotted any step in my workflow I could improve on (either with another process or different settings) and, finally, if you had an answer for my l.norm and deconvolution issues.

Thanks a lot for your time and advice ! 
Maxence
Helpful Respectful Engaging
Funkonaut avatar
Funkonaut avatar
Looking good!  You're right about Deconvolution; it's a tough process to get good results.  I skip it completely as it tends to make the stars look artificial.  As far as suggestions:
  1. First and foremost, if you haven't already, follow along with this video: https://www.youtube.com/watch?v=Nd3gTUMO_J4.  I still to this day use most of the steps in that walkthrough to process nebula.  It's a great workflow.  And since you're working on the same target you will learn a lot.
  2. Try using CurvesTransformation rather than ColorSaturation.  In the tool, click on the far right option "Saturation", turn on preview, and for the first adjustment grab the line at the first major point (.25,.25) and drag it up and to the left to ~(.125,.375).  Apply and reset. Then take (.75,.75) and drag it up to (.75,.875).  Apply and close.  I've found that just those two saturation adjustments are sufficient.  Less is more.
  3. If your image is noisy, try using the EZ Denoise Script.  Its defaults produce great results.  Scroll down to "Installing the EZ Processing Suite" here: https://remoteastrophotography.com/2020/12/easy-and-effective-noise-reduction-in-pixinsight

Cheers!
Well Written Helpful Insightful Engaging Supportive
andrea tasselli avatar
andrea tasselli avatar
I'd say it's fine first go at M42. Well done. A couple of things (to get started): 

1. Too heavy handed at noise reduction. Noise is fine when subtle as it adds a veneer of reality. Trying to remove it its entirety results if an artificial look about things. Obviously you might have a fairly noise-free image which would have to struggle with noise management in first place but that is for another day.

2. Over-processed look, especially in how the stars look, i.e. flat roundish discs of uniform intensity. I would avoid most of what you have done and control the non-linear contrast by acting on the curve transfer only (use CurveTransformation process).
Helpful
Maxence Ouafik avatar
Maxence Ouafik avatar
Looking good!  You're right about Deconvolution; it's a tough process to get good results.  I skip it completely as it tends to make the stars look artificial.  As far as suggestions:
  1. First and foremost, if you haven't already, follow along with this video: https://www.youtube.com/watch?v=Nd3gTUMO_J4.  I still to this day use most of the steps in that walkthrough to process nebula.  It's a great workflow.  And since you're working on the same target you will learn a lot.
  2. Try using CurvesTransformation rather than ColorSaturation.  In the tool, click on the far right option "Saturation", turn on preview, and for the first adjustment grab the line at the first major point (.25,.25) and drag it up and to the left to ~(.125,.375).  Apply and reset. Then take (.75,.75) and drag it up to (.75,.875).  Apply and close.  I've found that just those two saturation adjustments are sufficient.  Less is more.
  3. If your image is noisy, try using the EZ Denoise Script.  Its defaults produce great results.  Scroll down to "Installing the EZ Processing Suite" here: https://remoteastrophotography.com/2020/12/easy-and-effective-noise-reduction-in-pixinsight

Cheers!

@Funkonaut 

Thanks for the feedback ! 
I watched the video you shared and bookmarked it. Thanks for it  

Why would you favor CurvesTransformation over ColorSaturation ? I tend to use CurvesTransformation for the initial boost in saturation then ColorSaturation when I want to make some specific colors pop  a little more  

I tried to install the EZ Processing Suite but had an issue. I need to troubleshoot it before I can try it but thanks for the suggestion anyway  

andrea tasselli:
I'd say it's fine first go at M42. Well done. A couple of things (to get started): 

1. Too heavy handed at noise reduction. Noise is fine when subtle as it adds a veneer of reality. Trying to remove it its entirety results if an artificial look about things. Obviously you might have a fairly noise-free image which would have to struggle with noise management in first place but that is for another day.

2. Over-processed look, especially in how the stars look, i.e. flat roundish discs of uniform intensity. I would avoid most of what you have done and control the non-linear contrast by acting on the curve transfer only (use CurveTransformation process).

@andrea tasselli 
Thank you for your feedback !  
My image was very noisy at first but I reckon that I was too heavy-handed. 
Over-processing is an issue I have and try to work on. Especially for my first images, I had the mindset that if I didn't use every tool in the toolbox, I would be wasting part of my data. I think I need more processing experience to find the proper balance in my workflow
Funkonaut avatar
Funkonaut avatar
I haven't used ColorSaturation much, but after playing with it just now it is more heavy-handed than CurvesTransformation.  My initial color enhancement comes from LRGBCombination.  After I've applied my STF to HistogramTransformation which makes the image non-linear, I run LRGBCombination from the extracted Luminance with a Saturation of 0.290.  That brings out a good amount of color from the image.  Then I boost it ever so slightly with CurvesTransformation.  Those steps are in the YouTube video I referenced above.  Again, less is more.  :-)
Well Written Helpful Insightful Concise