4.19
#...
·
 |
|---|
|
I would love feedback on my 2022 redo edit of the famous Dark Shark. I had a few minutes before my wife and kids returned this evening and I decided to redo edit of my data from last year instead of starting on new data. Just amazing what a year in this hobby can change. Image B and C , are versions of my image from last year. The mouse over provides a stark demonstration of how the processing and stars have changed in the two images. I have also included several comparison zoomed in images. With 2022 on the left and 2021 on the right. Still some noise that I left to avoid making it too fake looking to my taste. Also, there are some artifacts from Starnet original the were left from my base edit that Starnet2 has done much better with on other images. In general over the last year, I have had much better luck bring out dark nebula while decreasing the noise with helpful advice over the past 9 month. In addition, the biggest change for me lately has been the introduction of the GHS stretch. While I have had verifying luck with it. I have been doing this for most of my images lately with good results compared to star deemphasis and star reduction methods. Initially for GHS stars, I STF stretch the linear data. Then I remove the stars and complete sharpening, saturation, curves, etc. Then I go back to my linear image image and stretch it with GHS at 1.5-2.5 stretch factor and around 9-10 local stretch intensity. I mark an area near the nebulosity and set this as the SP point. I then take starnett2 and remove the limited stretched stars and add them to my edited starless image. The results have definitely been a step up for me in my option. Thanks for the feedback good, bad, or ugly!  Dark Shark Nebula 2022 Edit (GHS stretched stars and Editing Improvement) LDN 1235 |
0.00
#...
·
·
1
like
|
|---|
|
This is very good processing Brandon. Congrats! A few things imo: 1. Surrounding clouds are biased to purple. try invert image, SCNR to remove green and invert again. 2. Also reguarding to the surrounding, they are also over stretched a bit. you can see unnatual black spots which shouldn't be features. Try improve masking to avoid this. 3. I think you can go further with chrominance denoising. 4. There are ringing like artifacts around stars. Starnet only tries to remove stars and fill the area at the stretch level of the image. So if you stretch after removing the stars, you will see stretched artifacts. What you can do here is to do a two step star removal. First, stretch the image to where the stars look right in the final image, and keep the stars image. Then stretch the image further to where the features are right. Then do another SN, and process the SN image. Finally, combine the earlier and often darker stars image and the SN image. In this way, you can keep the artifacts at minimal level. |
|
4.19
#...
·
|
|---|
|
Thanks Dong! Your comments are very insightful 1) I often use the invert SCNR transformation for stars, but had not applied it to purple bias in a nebula. In fact, I just failed to notice the purple base line hue. 2) I think the dark spots may also be related to my linear denoise process. I just didn't notice them until the end. 3) What approach do you usually use on the chrominance once you have gotten to the non linear? I have been using TGV denoise and MMT in the linear stage as demonstrated by Tim Hutchison on TAIC. As needed, I will use MMT on the non linear image to noise preference. 4) I had not though about partial stretching the image and removing the stars, then completing another complete stretch. Thanks for sharing. |
|
0.00
#...
·
·
1
like
|
|---|
| TGV is fine with removing chrominance noise. I usually do this after stretching, and maybe one more time after stretching saturation. But if you have Photoshop, it's camera raw filter does a very good job removing chrominance noise. |
13.14
#...
·
|
|---|
| I'm not worthy or skilled enough to comment on this fine image! -john |
