Deconvolution Vs No Deconvolution

[Rodd]

I thought I'd post this comparison to show that deconvolution can be useful in accentuating fine details.....when it works as designed.  therein lies the rub, for me.  This is probably the best result I have gotten using deconvolution in the linear state.  It happens to be my recent M101 image that I reprocessed using deconvolution.  It is amazing that I was able to post this because this was a complete reprocess from the beginning--and I typically can not achieve the same look in subsequent processing efforts.  This one is pretty close though-close enough to be able to ignore the differences and concentrate on the deconvolution.  To remind: TOA 130 with ASI 1600, about 28 hours of data HaLRGB.  I followed the Light Vortex tutorial except for one thing--my star selection process was simplified to just the brightness (between .3 and .9).  The tutorial uses about 4-5 selection criteria that weans down 100 stars to 15 or 20.  I find that hard to do because sometimes the number of appropriate stars is small--especially in narrowband images.  So I just use brightness and about 30 stars.  The biggest issue for me using deconvolution is targeting appropriate high signal areas.  Often, the process creates little bright spots that look like like little stars--but I think they are artifacts.  One must take care when creating a star mask for local star support--as well as a mask that protects the low signal areas.  I am sure it is easier in PS.  Anyway--here is the result.  It is subtle, yet unmistakable.  It works to remind me that in astrophotography very small changes can impart big results on images.

 

No deconvolution

 

Decomvolution in linear state

 

 

Edited May 2, 2020 by Rodd

[matt_baker]

Looks great! 

Was wondering if you could upload the masters, so I can play around with my processing?

[Kinch]

Yes....deconvolution shows a fine improvement   👍

[Rodd]

10 minutes ago, matt_baker said:

Looks great! 

Was wondering if you could upload the masters, so I can play around with my processing?

Thanks Matt.  If you go back a page or two I uploaded the masters on my original post   You’ll find it. It has like 6 pages of comments. I am not near my computer now.  But you’ll find them 

[Rodd]

3 minutes ago, Kinch said:

Yes....deconvolution shows a fine improvement   👍

Thanks Brendan.