On 2/6/2012 10:35 AM, Ken Norton wrote:
> If a lens happens to be technically better than another, but the rest of the
> system interferes then is it actually better?
Interesting turn that thread was taking, while I was doing some lens/system
stuff myself, although the direction is
different. :-)
I've been learning about the S100 and the files it produces as I process shots.
Although I bought the camera in part for
better IQ at higher ISOs, one thing that has impressed me is the IQ at base
ISO, 80. This is an entirely new sensor,
CMOS, as opposed to the CCD sensors in the prior S and G series cameras.
Here's an image with a little of everything, wide DR from direct sun and deep
shadows and lots of fine detail of various
types center and edge.
<http://www.moosemystic.net/Gallery/tech/S100_ISO80/IMG_0179.htm>
1. Full frame, RAW, ACR conversion.
2. 100% samples of parts of the image.
3. Neat Image applied. It's quite difficult to see the difference, but it's
there.
4. Shadow/Highlight Tool applied only to lower right, revealing only subtly
noisy detail from very deep shadows.
5. Focus Magic, out of focus correction, radius 2 pixels.
6. LCE & Curves.
7. FF, after.
Although the G11/12/S90/95 sensor is certainly not particularly noisy at 80,
there is definitely noise there in the
shadows of RAW conversions. This noise sets a floor for dynamic range. One
can't go too far with exposing to the right
if one wants to be able to hold detail in the shadows. As they are brought up
in post, they bring noise with them.
Also, processing like deconvolution, sharpening, LCE, etc. can make noise in
brighter areas that was initially not
visible start to become a factor. This puts a practical limit on what can be
done to shadows and low midtones.
The S100 lowers the shadow noise and changes it. It's possible in Neat Image to
really knock down even the subtle noise
that's there. It makes no practical difference until doing other processing;
most folks probably can't even see the
difference at 100%. But when deconvolution is applied, shadows are lifted and
LCE applied, no odd, grotty looking
noise-ish stuff shows up. I'd say it adds at least a stop of practical DR,
maybe more.
The other thing about S100 files is that they seem to naturally lend themselves
to deconvolution without revealing all
sorts of artifacts - they just get sharper! The result of these factors
together is an amazingly better IQ. You'd almost
think the lens and sensor were designed for such processing.
After post, these files look more like those from a small sensor DSLR of a
generation or two ago than a compact camera
the size of a pack of cards!
So how does one evaluate the S100's lens? It's quite adequate for web images
and prints up to 8x10, although even at
8x10, I'll bet post processed versions would look better. But this is a tiny
camera from which one could make quite
large prints.
What if shadow NR and overall deconvolution customized for the lens, focal
length, aperture and focal distance were
built in?
Next - I just peeked at an image with huge DR and some twigs against sun and
sky in Canon's Raw converter. Toggling the
lens correction adjustments, I see that Canon is doing some very interesting
things that I don't think I can duplicate
in anything else. Looks like it can send the image directly to PS, as ACR does.
Hmmmm
De' Convolutin' Moose
--
What if the Hokey Pokey *IS* what it's all about?
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|