Interesting article, confirming — to my considerable surprise — the conclusion at which I'd doubtingly arrived some while ago after finding that I was getting better onscreen results from my S95 than from what I think were reasonably decent scans of negs from my old Nikon SLR.
It would appear that what you though were decent scans may not have actually captured all the detail that was available on the film. Where did you obtain the scans from?
I'm unclear about one point, though. Can anyone explain in simpler (or at least other) terms why higher resolution doesn't necessarily equate directly to higher enlargability?
My observation is that the resolution on the film is only one step in the chain. One still needs:
1. Excellent Optics to focus the image onto the film
2. Correct exposure and focus with no camera or subject motion blur.
To make a fair comparison between two systems, items 1 and 2 must be exactly the same for both. For example, the test above only changed sensors while keeping the other variables constant.
3. Correct Processing - development temperature and agititation for example.
4a. Excellent optics to enlarge the image if printing directly or
4b. Excellent scanner to digitize it. The scanner in turn needs:
i. Much higher pixel count than the scanned medium due to sampling errors.
ii. Excellent Optics to read the image
iii. Just the right amount of exposure and post processing to bring the image out.
Unless
every step along the way is as described above, the final result will not be anywhere close to what the resolution figures alone would imply. That is because the performance is not based on what the weakest link can deliver as many might believe. It is based on multiplying together the performance factor (modulation transfer function) for each and every item in the chain.
But even under the best circumstances, for example, every item in a 4 stage chain performed at 95% of its potential, if you take .95x.95x.95x.95, the result is .81 or 81%. Each stage adds a reduction in quality. If any one of the items performed at say 75%, the results would be only 64% performance. See how quickly the performance dropped?
So this means, an all digital system that removes the scanning process will get a performance advantage with all other factors being equal.
If you'd like to learn more from a trusted source, check out:
Zeiss- How to read MTF charts Part1 and Part 2
http://www.google.com/#hl=en&sugexp=ldymls&xhr=t&q=Zeiss-%20How%20to%20read%20MTF%20charts&cp=29&pq=zeiss-%20how%20to%20read%20mtf%20charts%20part1%20and%20part%202&pf=p&sclient=psy&source=hp&aq=f&aqi=&aql=&oq=Zeiss-+How+to+read+MTF+charts&pbx=1&bav=on.2,or.r_gc.r_pw.&fp=a5452f260f79a214&biw=960&bih=753&bs=1
--
Robert
