Looking at the specs of various digital cameras I realized that 4:3
aspect ratio 5MP picture of CP5000 actually has the same height
( 2000 pixels) as the 3:2 aspect 6MP frame of say D1x or DCS760.
Thus, since most of (my) pictures end up printed in 8 x 10 inch
format anyway, we are finally getting the same "essential"
resolution as the "slightly" more expensive pro cameras,
conveniently "pre-cropped" for printing... ;-)
Not too bad, eh?
--
Michal Urban
The critical "test" of this idea will come in the resolution you
see at 100% size on your computer monitor as you compare images
from the two kinds of cameras. The D1-class cameras have a
pronounced better-looking image structure at full size, so
naturally they look better in a print or magazine.
There will be a point at which a downsampled image from a 5000 will
look as good (in a practical sense) as a larger-chip SLR image
--per pixel. That might happen at 75% size or 65% size, I can't
tell, yet.
Once the --per pixel-- resolution of the larger camera has been
reached, the comparison will have sharper teeth.
When you downsample this way with a 995, for instance, images
resharpen to D(X) camera level within this range. By the time you
get to 50% of the original size, the image is obviously totally
pixel perfect as far as sharpness and detail are concerned.
I've done a test of taking a 3 megapixel image down to 1600 pixels
wide in Photoshop, thus compacting the image information into 61%
of the original pixel area, and then expanding it back up to 2048
pixels wide. Then compare the sharpness of the down/up image to the
original. Hmm. Not much different from before... You lose
about--what? 8%? Meaning that the best image may come from a full
frame shot, but a 92% image can come from a shot that takes 61% of
the storage space.
At some magic number for size reduction, virtually no difference
should show from a down/up -sample operation. That might be thought
of as the "true resolution" point. It's not really that simple, but
the full size image from the compact cameras isn't as clean an
image structure as images processed through the algorithms Nikon
packs into their big cameras. Are they sitting on image processing
tech that could "improve" the smaller cameras' images? It would
make marketing sense, wouldn't it? Why give away the superior
processing they need to cause their major professional devices to
look so good.
When you compare the 100% size image from a D-1 to a Canon D30, you
see that Canon hasn't achieved the "per pixel" picture power that
Nikon has. But watch out. Competition is a great motivator of
technology, and there are competitors out there who would LOVE to
have their cameras spit out pixel perfect images. Sony, for
instance, has no SLR line to defend. If they came up with D-1-like
image processing algorithms, they might turn their next camera into
a super cam and force Nikon to do a technology transfer into the
compact camera line. Wow, was that speculation or what?
I would think that the 5000 will be about the same amount of detail
--per pixel-- as the 990/995 -range cameras. Meaning that they will
be somewhere in the range of 70% of the image detail of a D-1x. But
that's a guess.
-iNova
