The use of in camera processing appears to be low hanging fruit for all
the manufacturers. I think CPU's in the lenses with 2 way communication
was novel when Oly implemented that. It seems increasingly difficult to
extract the lens optics form the
"system." Have seen one review site looking at a DZ sans cpu effects
but not sure what that means anymore as it not usable w/o the E-x.
There must be some modest cost in IQ as the Dr. Focus first principle
of optics (there is no free lunch) can not be violated. There is some
cost too even in expensive T-S lenses to correct with a PC lens as the
highly thought of Canyon 90mm T-S loses a ton of resolution at full
shift. I am not even sure how to properly devise an experiment to sort
the question out whether the no holes barred lens design can trump IQ
contrasted with a more modest lens, designed with easily correctable in
lens/camera cpu aberrations.
Perhaps even the lens MTF won't be sacrosanct as deconvolution
algorithms can steal back some resolution---used in DXo software and
others. Too processor intensive to put in camera/lens for now but
perhaps not for too much longer.
The D3 even looks at the image to correct LCA. see below What next?
Mike (student of Dr. Focus)
from imaging-resource.com on the Nik D3:
"Lateral Chromatic Aberration Correction. While other cameras have had
lens distortion processing built-in, notably the Olympus E-1, none have
done the processing based on the distortion they see in the image like
the D3 and D300 do with their Lateral Chromatic Aberration correction.
The E-1 took its distortion-correction cues from whichever lens was
mounted, and applied a pre-set amount of correction; but no image
analysis actually took place. That's also the approach taken by most
software applications. But the new Nikons have the power, thanks to the
EXPEED processor, to actually analyze each image after capture and fix
the chromatic aberration before saving the JPEG file.
We'll have to see how well this works once we get a production sample
to play with, but the prospects are exciting, particularly on the D3,
whose full frame sensor places more demands on most lenses. We'll
report more on this exciting development, including sharing the results
of some tests made with DxO Analyzer to compare before/after
performance, once we have a chance to test a production model of the
D3."
Dr. Focus wrote:
"Does anyone here know what the "system performance" is? When we
correct
distortion with software we pay a price in resolution from having
expanded some pixel locations. Of course, we pay the same sort of price
in using a shift lens which probably does not have the same resolution
at far (shifted) edge as in the center. Correcting CA involves shifting
pixel positions based on color and angle from center. Just like
distortion correction, moving a pixel requires something else be
invented to fill the space previously occupied by the moved pixel.
What's the actual effect? Can the software do as well as a moderately
expensive lens? (I don't include no-holds-barred Leicas here)"
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|