From what I had found online, in pre-14.4 LRC, you could use the mask function to limit what would get sharpened AFTER denoise ran. With 14.4, it looks like the image loses a touch of sharpness after masking but before DeNoise. In a way, this makes sense (I think) as the sharpening that is auto defined is being impacted by the masking. But what I have not been able to find for 14.4 is if DeNoise is still sharpening after denoising.
Does anybody know if sharpening is still an automatic post-denoise event?
What auto defined sharpening?
The sharpening sliders are controlled by the user. They function the same as they always have, whether you run Denoise or not.
There are default settings for all the sharpen sliders and this predefined (probably a better word) sharpening affects the whole image. Masking allows the user to say what to sharpen and what not to sharpen. From what I had read in the past, sharpening automatically occurs post-Denoising, and uses the sharpen settings to do so.
It is probably just a matter of not having noticed this before 14.4, but as mentioned, I noticed that some images look a touch less sharp after masking but before running DeNoise.
So, I am looking for confirmation (or refuting) of this observation and confirmation that DeNoise 14.4 still does post-denoising sharpening.
Denoise removes noise and paints in fake details to replace what it thinks was lost to noise. Being able to see more detail is not the same as it being sharpened.
With synced settings, a Denoised DNG created by an older Lightroom is identical to the Denoised result created by the new version.
RT--, there are two points of comparison in the DeNoise process; at least the way I was doing it based on how older versions of DeNoise worked. Based on what I had read and a video or two I had seen, the recommended way to do it was from the Details panel (AOT right click, DeNoise). With the older versions of DeNoise, reducing the slider resulted in more detail retained, with the appearance of greater sharpness most evident on things like fine feather detail on passerines. Of course, a lower setting meant more noise. The trick to minimize this was to use masking so that noise in low detail areas wasn't being sharpened.
When I do this now, I see a loss of apparent overall sharpness after masking but a return of the sharpness (give or take a bit) after DeNoise. I'm thinking this was also the case before but I just didn't notice it until now as, between the new computer (especially the beefy GPU) and the change in how DeNoise works, I am scrutinizing it closer.
I have noticed something else though. How heavily I plan on cropping an image plays a role in how aggressive I am with the slider (as does the ISO setting of the photo of course). There are times in the past were I would have changed the setting to 20 or 30 before running DeNoise. Now it runs automatically with 50. I then move the slider to 20 or 30 and compare the results. Unlike in the past, I am not really seeing a loss of detail or sharpness in 50 vs 20 or 30 (although occasionally 50 looks a touch plasticky to me).
In summary, I am not sure any more that doing pre-denoise masking "buys" anything nor does it feel as necessary to try to get the slider setting down from 50.
I am curious what other users' observations and perceptions are.
As mentioned, I find it hard to believe that a more aggressive denoising (say, 50) doesn't have any detrimental effect compared to less aggressive (20 or 30) and yet that seems to be the case. The only exception I have noticed is that on the rare occasion an image processed at 50 can look a touch plasticky. I have never seen that problem with 40 though and, unlike in the past, I am beginning to think there is no advantage to reducing it to 20 or 30 but there is the disadvantage of leaving some noise. Unless my image is 10000 or 12800 (where I would likely want 50), I am inclined to process it at 40 and not worry that a lower number might work.
Related (perhaps) to that point, I have noticed that the AI Remove is spectacular in modifying or removing content BUT if one exams the boundaries of the selection area at 100%, one can often find a slight "texture difference" boundary line (the selected/modified area being smoother (no texture or noise)). This is usually easy to fix with the heal tool and perhaps a bit of play with the opacity setting on heal. It does make me wonder though if it is affected in any way by the degree of noise left in the image. Perhaps DeNoise replaces with noiseless content and so any noise left in the surrounding images presents as a boundary or transition point between "clean" and slightly grainy or textured? BTW, I zoom into 100% and check this on all AI Remove actions because if I ever make a print of decent size, it will show.
Yes, 40-50 is my default range, with 50 winning out in the VAST majority of cases. I'm adding in grain after the fact quite often, so "plasticy" isn't really a concern. If all I want to do is remove color noise, a value of 1 often does the trick.
Thanks for the reply RoamingScott. Have you noticed the boundary/texture difference I mentioned? If you haven't, it may be that 50 is removing all noise and the noiseless (?) AI replace/remove blends in very nicely; unlike the fine lines/boundary of texture difference I am seeing.
