Fascinating to see the discussion about the artifacts with Denoise. I’ve not used it on a Milky Way image, but loved what it did for the image I just posted. I’ll have to keep an eye out for the squiggles, but hope Adobe can fix that because Denoise is so powerful for these high iso shots!
I agree with the above posts Keith, the Milky Way to the left would be optimal but it’s not a deal breaker. As to the foreground, I also agree there is a bit more variation than expected, and the redo is better. Did you happen to take a real long exposure of the foreground when shooting the Milky Way? I sometimes do that; even if I don’t use it I find it can help guide me on how to process the blue-hour photo.
Thanks for your comment @John_Williams. Yes, I did shoot another composition after this one with the Milky Way further to the left and a long exposure foreground. I haven’t processed that image yet, but when I do I’ll post it here.
I think it’s getting even better! I like the blue in the sky. There is now less of the eye jumping from the MW to the aspen grove – lovely as it was! Wonderful captures and processing!!
Just a quick footnote here re the NR discussion above. I unearthed a set of 15 MW shots I had done in quick succession to test SLS – which I never got around to doing. But I’m immersed in PixInsight just now, working on the comet (or trying to – it’s not simple) so I ran the set through the batch preprocessing procedure that gives a single registered (aligned) image and the noise was beautifully removed without any artifacts. The stars are still what they are, with a camera lens, but this is clearly the way to go for NR. I suspect SLS would give a similar result. Probably fewer frames would be enough. Good material for a Discussion. I won’t take time till “comet season” is over but anyone else is welcome to jump in and start one!
@Diane_Miller, I think what you have verified is that when processing astrophotography images, methods that involve mathematical averaging of pixels from multiple frames is superior to Lightroom’s AI algorithm in its Denoise function, which operates on a single image. It’s far easier to identify noise when you have data from 15 images than it is from a single image. Lightroom’s Denoise is great for normal photos, but for now I will continue to use programs like PixInsight, SLS, Sequator, etc. for astrophotos. Thanks for your testing, and good luck with the comet. I was finally able to image it a couple of nights ago, but the full moon was preventing me from getting good detail in the tail. I’m looking forward to seeing your results.
Thanks, Keith – I’d love to see what you got! Bottom line is – yeah – averaging multiple images is better than “AI”. I’m sitting here sweating comet processing in PI – I’ve done it on 2 previous ones and have halfway decent data from the 15th (with the anti-tail visible in the blinked fiies), but yes, no detail in the tail any night so far. Lots of moonlight and low in the increasingly-thick air. I’ve hit an error with local normalization of the comet-only frames… It thinks the LN files aren’t associated with the comet files, but they are right there in the same directory. Only difference is now I’m using FBBP instead of WBPP. Could that be the difference? Heading for Cloudy Nights…
Are you trying to combine comet-aligned images with star-aligned images? I found the process so complex in the past that I had to write it all down so I would have a procedure to use on future processing. I can send you my procedure if you would like to see it.
Yes, that seems to be the way to do comets. I have a long cookbook document mostly from Adam Block’s videos. I think I found the current problem – I was trying to use FBPP but apparently it doesn’t write all the files and data needed for comets. I’m running WBPP now. I’d love to see your process, though. If it’s in a format that won’t work in a Message, I’ll send you my email.
I think I have permanent brain damage from trying to put all the pieces together from video “tutorials” done by people who assume the viewer already knows 99% of it. (If you want to know how to do something, don’t ask someone who already knows how.)
Boy, I’m sorry I missed this excellent conversation! I almost always stack for noise reduction in my Milky Way images. For the last several of the season in '24 I experimented with blending in a longer exposure shot at the same time as the sky because as others have said above, the light in a blue hour foreground is still directional, vs the more averaged starlight. My normal workflow is to run the stack in SLS, then export to Lightroom. I’ll create a mask for the sky and for the foreground and process them seperately. I then send it to Photoshop for star reduction in the Ministars action. This reduces the brightness of the stars and makes some of the more faint ones go away. Too many stars can create the illusion of noise in the sky. Occasionally, I will run it through Topaz sharpen if I have close-up foreground elements. I usually mask out the sky for this, but I have found that you get far fewer artifacts when doing additional noise reduction in Lightroom or Topaz in a stacked sky. I love that image though! The aspens are amazing and deserve a little increase in their luminance. I dont mind the light on the mountains, but I would want to see it more evened out across the frame, even if some where more in the shadows as you saw it. Nice work!