> Date: Sat, 25 Jan 2003 18:44:33 -0600
> From: "Bill Pearce" <bspearce@xxxxxxxxxxxxx>
> I still maintain that comparing a print from a digital camera with a print
> from a film scan (inkjet or on photo paper), is not valid. Make a GOOD WET
> PRINT from that 6x7 chrome, and compare it to a print from the digital.
> That's a valid comparison, as the print from the scan has an additional step
> inserted.
Agreed. While I wait for digital to be both really good and affordable,
here's a bit of rough math, which may explain some perplexing results I've
seen in Reichmann's article and in Photo Techniques.
Recall that as an approximation the resolution of a combination of two steps
is related to that of the steps as
1/Rt = 1/R1 + 1/R2
Suppose you have a 70 lp/mm lens and 120 lp/mm film (Velvia at intermediate
contrast). The result comes out at 45 lp/mm, which is very roughly 7.6 Mpx.
However, if you scan the slide at 3200 dpi or about 64 lp/mm, you bring the
result down to 26 lp/mm or very roughly 2.5 Mpx. This may explain how 4 or 5
Mpx cameras can appear to outperform 35mm film. Even an 8000 dpi scan, about
160 lp/mm, brings the result down to 35 lp/mm or very roughly 4.2 Mpx.
Of course with the dollars and CPU being now thrown at the problem, digital
sharpening/interpolation algorithms in high-end cameras probably now
outperform anything from amateur outfits like the Soviet and US militaries.
:-) But there's only so much software can do to reconstruct information not
in the data.
Andrew
(hoping for a Kodak 14Mpx 24x36 back for his OMs :-) )
< 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 >
|