I've thought all along that a big part of the problem is the combo of
automated printing and scanning a 4x6 print. Now I have some evidence to
offer
Check out
<http://home.attbi.com/0.000000E+00dreammoose/wsb/html/view.cgi-photo.html--SiteID-322698.html>
.
Side by side are 2 crops of the same shot. The original was shot 5/1/02
with an OM-4(T) and Tamron SP 60-300mm lens, set at 300mm/f5.6 and
either 1/60 or 1/125 with the lens resting on the car window (engine
off!!) to steady it.
The left one is a 600dpi flatbed scan of a 4x6in. print from Kodak
Royal processing. The right one is a 2720dpi scan with a Can*n FS2710.
This gives roughly the same image resolution without any resampling.
The area shown is about 300f the width and 260f the height of the
full frame, or about 80f the area.
As I expected, the film scanner image is much sharper. I've said before
that print emulsions are just not designed to hold high resolutions, but
I was a little surprised at how big the difference is.
What I was especially looking for was the effect on contrast and color
differentation, and I really found it. I selected a shot with rich and
complex variations in color and contrast. As you can see, virtually all
the subtle color variations in the film scan are lost in the print scan.
It's not just sharpness, a great deal of color info is gone. Most of the
subtle graduation in contrast is lost too.
What does this mean for web images? Enough of the detail in the film
scan is lost in downsampling and JPEG conversion, so that there may not
be much apparent difference in sharpness between equal sized small web
images from the 2 sources, although I suspect that at TOPE size and
above there will be an apparent difference. On the other hand, the image
derived from the film scan should retain noticeably better color and
contrast detail/graduation. Although I can't say for sure what the
separate contributions of the printing process and the flatbed scanning
process are, it is clear on the print that a lot of color/contrast
detail is already missing.
Maybe the next step is to make a lower resolution direct scan of the
whole print and do a careful multi step downsize sharpening on the film
scan to TOPE sizes for another comparison. But not tonight.
Moose
PS to Andew Beals: This is an extreme example of what a really long lens
(1,000-1,100mm) would find on Yosemite's walls, a mysterious face in the
rock. Looks like maybe a glacier broke his nose?
bsandyman@xxxxxxx wrote:
I have been taking pictures of this bright yellow flower
now for months. I have been disatisfied so far partly
because the thing that draws my attention to these things
is the very slightly orange portion of the petal. I have
so far failed to capture this aspect of the flower. Here
are some pictures to show what I mean.
http://www.photo.net/photodb/folder.tcl?folder_id=236230
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|