At 8/15/2021 11:53 AM, Master Sharpen Moose wrote:
>On 7/31/2021 7:16 AM, Wayne Shumaker wrote:
>>At 7/30/2021 07:38 PM, Mike wrote:
>
>>>WayneS writes:
>>>
>>><<Nice. Looks contemplative.
>>>
>>><<Is it Audubon Society material or <<did you use AI sharpen on it :-)?
>>>Thanks for looking. Good call. MR. Hawk was in shadow and a bit surprised
>>>by the noise at base ISO.
>>>Denoise AI first did the heavy lifting â??sharpeningâ?� and a wee tou
>>>touch up with Topaz sharpen. The lens is a tad soft at long end near max
>>>aperture and Iâ??m sure atmospheric ef effects did not help.
>>>Mike
>>Since you allow zooming in, and my experience with the AI tools, I can tell
>>by closer look at the lichen on the fence where it attempts to sharpen but
>>can't quite do it, then toward the top of the fence it gets a bit fuzzy but
>>distinct edge with the background. Zoomed out it is harder to tell. I think
>>the tools don't do quite as well with a combination of some motion blur and
>>some focus blur.
>
>Speaking not to this example, but to your "don't do quite as well"
>generalization, It is sometimes possible to "stack" Topaz AI applications,
>with some almost magical results.
>
>This shot with E-M5 and the old Oly 75-300 Mk I, is the kind that frustrates
>me. Unusually close subject, great light, etc., but not all that sharp. TAIs
>fix it!
><http://www.moosemystic.net/Gallery/tech/Process/Topaz%20AI/Sharpen/RedTail.htm>
>
>As focal distance increases, atmospheric effects increase, and how much the
>AIs can help lessens. Or perhaps I should say they still work wonders, but the
>end result is still not really sharp and artifact free.
>
>There are so many factors that affect "sharpness" of long lens shots that it's
>important for a good result to closely evaluate them. It would be nice to have
>simple rules about what tool and settings to use, but it's not so.
>
>Using my current extreme example, Oly 100-400 with 1.4x teleconverter:
>
>First, a baseline. Everyone is likely to say, long zoom, long end,
>teleconverter, wide open - has to be bad. Weeellll, not so much.
><http://www.moosemystic.net/Gallery/tech/Lenses/Oly%20100-400/O100-400Yellow.htm>
> Topaz Denoise AI "Standard" didn't do much, and "Clear" seemed too
>artificial, esp. in the vertical hairs. "Standard" plus Sharpen AI "Motion"
>seem to me to have done a good job.
>
>FL of about 20 ft., late-ish on an overcast afternoon, relatively clear air,
>no sun heating up the ground, and the pixel peeper's 100% sample isn't bad at
>all. After some Topaz AI massage, it becomes really clear that
>sharpness/detail at any point in the image depends more on the very shallow
>DoF than inherent lens sharpness.
>
>Sooo. . . I have a lens combo that's pretty darn sharp. I need to know that
>to evaluate shots taken at greater distances.
>
>Earlier, a Red Tail Hawk was circling above us. The Oly EXIF FD is very
>suspect, as it shows almost the same distance for images where the bird is
>clearly at quite different distances. In any case, over 400 ft.
><http://www.moosemystic.net/Gallery/tech/Lenses/Oly%20100-400/O100-400Hawk.htm>
>
>It seems that the tools can make a big improvement, but not overcome
>atmospheric effects to render anything close to the flower photo.
>
>Yet other factors arise, to complicate testing/comparison, beyond Ctein's
>caveats*:
>
>1. A middle distance shot of a turtle sitting still on a log, on the same
>trip, shows lots of motion blur. Much more than the other examples. What's
>that about?
>
>2. I was about to add examples of processing that shot - when an update to
>Topaz Sharpen AI arrived. Different, and much better results from Motion Mode!
>
>I've dragged this reply on long enough. The turtle will have to wait.
>
>The Good, the Bad and the Moose
>>AI-Nit-Picking WayneS
>
>* Commenting on his brief testing of my PL 100-400: "Hard to say how good the
>lens is by any objective measure. It's insanely hard to critically test an
>optic like that. Stuff that's far enough away that depth of field isn't a
>problem, there are atmospheric ripples and distortions to deal with at the 400
>mm lens. Stuff that's close enough that that's not a problem, I have to
>compare multiple frames made with the same aperture where I shift to the
>camera to move the point of focus to different parts of the field of view.
>It's a pain and not terribly precise"
>--
Thank you for the analysis. I would love to see you process Mike's image and
reveal some magic sauce. My nit-pixel peeping comments was coming from my
observation of things I see in my own photos when I attempt to do AI
sharpening. Not to put down Mike's image. I will have to go back and try
combining on some of my photos.
I tend to have less patiences to do denoise then sharpen. The combinations
become quite a lot. But you give inspiration. I wonder if doing the denoise
first helps the AI sharpen later do a better job?
AI challenged - WayneS
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|