On lens "corrections" -
While researching a lens I was interested in I stumbled across a comment that struck me. It went something like (yes, I'm paraphrasing in the following)...
"... you can use software to sharpen up this old lens, but you won't have a true understanding of how bad this lens really is..."
I wondered if the writer understood something fundamental to digital photography and current image processing. Software "corrects" for all manner of lens design and implementation "faults."
To see what I mean turn off software intervention. Specifically, turn off -
- Capture Sharpen - which ostensibly counters AA filter effects
- Lens Correction Profile (LCP) - which corrects for
- Chromatic Aberration
- Field Distortion
For in-camera jpgs this means locating and changing the settings there on the camera. For RAW this means locating these settings in the image processing software, where switches and controls could be very well hidden.
Now have a close look at an image and compare it against a software "corrected" image of the identical scene.
It can take a lot of image processing just to reach a decent starting point. What's good for the goose might be good for the gander, right?
Question: Why not apply software "corrections" to old manual focus lenses?
Sony NEX05T + Sigma 24mm f/3.5 DG DN
Illustrating the results of
all the software interventions applied
by default on file import into a RAW
image processing software
On Lens Diffraction limits ~
Who hasn't read lens reviews that tell us things get mushy when shooting at apertures below the limits of what a sensor can resolve? The phrase "diffraction limits" comes up shockingly often.
For full frame cameras 12 to 24mpixel that's f/16 and for 40 to 60mpixel that's f/11. The caution is to avoid those apertures if you want the sharpest rendition possible. The implication being that images shot at apertures below the resolution limits of a sensor are <insert favorite unsavory expletive>.
In light of software intervention capabilities, I wondered if this was strictly true.
Taking the question seriously, I used a beautiful old Nikon Nikkor 20mm f/3.5 Ai lens, focused two feet, adapted to a 42mpixel Sony A7RII (shooting RAW) and shot three photos starting at f/11, then f/16 and finally at f/22.
Looking at the images with Capture Sharpened turned off I could see a slight difference in the sense of sharpness between f/11 and f/22. Between f/11 and f/16 it was a little more difficult to tell a difference at full rez or 200 percent rez.
Then I turned Capture Sharpen on and... <drum roll, please>... I see zero sharpness difference between them. As in, it don't matter (bad English intended). Software did what it was designed to do: Make things sharp. OK. OK. There was, however, a clear difference in depth of field. But that's also the point of shooting at small apertures, right?
To check if this was strictly true I then took a fine little Pentax-M 28mm f/2.8 and reran the f/11, f/16, f/22 comparison. In this case the sharpness difference between f/11 and f/22 was more obvious, even when using Capture Sharpen. Because of the way the Nikkor performed, there's likely something in the Pentax-M design that adds a bit more softness at really small apertures. However...
... for grins, I took the Capture Sharpened f/22 Pentax-M image and applied a gentle UnSharp Mask (USM) and compared the result to a Capture Sharpened bitingly/critically/fabulously sharp f/11 image. The result is... <another roll on the drum, please> ... zero, zip, nutt'n, nada difference between them. Software intervention of the kind applied to digital lenses now applied to the wee-Pentax-M is able to make a f/22 image look as good as a Capture Sharpened f/11 shot.
Question: How many people avoid shooting at small apertures because they've been told something awful happens down there?
Recap question: With the kinds of beneficial image improving software tools available to us, why not use them, regardless of the lens?
No comments:
Post a Comment