On 11/3/2014 7:17 AM, Chuck Norcutt wrote:
Moose has the correct answer.
Oh my!
The scanner will produce an image size commensurate with the stated scanning resolution. Taking Moose's example of
scanning 35mm film at 4000 dpi you'll get an image of about 5600 x 3700 pixels. However, the big mystery is whether
each of those pixels actually resolves some detail different from its neighbors. Resolution in film cannot be more
than the size of its individual film grains.
Clumps of grain, as I understand it from reading.
Consider a case where the scanner is creating supposedly high resolution pixels but where three of those pixels are
all seeing a different part of the same film grain. No matter the scanning resolution the maximum possible real
resolution may be considerably less. Resolution for a film camera will depend on the film itself along with all the
other considerations that tend to limit it.
There are other interesting variables for MF film that digital doesn't have. The major one that comes immediately to
mind is film flatness. 6x7 cm film has two opportunities to lose detail to this, Depth of Focus in taking and Depth of
Field when scanning. Unless one uses cut film, a vacuum back and/or glass scanner carrier (which brings its own
problems), ALL MF film is not flat to a greater degree than 35 mm and to a MUCH greater degree than a digital sensor.
This is a significant problem that's hard to pin down, as it will vary from shot to shot, depending on film tension,
temperature, humidity, etc. on the shooting end and the film characteristics change after processing, so the places
affected may be different in the scanner holder. When making prints to be viewed at reasonable sizes relative to viewing
distance, it's not such a big deal most of the time. When scanning, and with the ability to pixel peep at 100%, it
becomes significant.
Likewise, the image dimensions from a digital sensor can't be more than the number of pixels on the sensor. In the
case of the Canon 5D with 3912 pixels in the vertical dimension there can't be more pixels than that in the vertical
dimension of the image. The maximum theoretical resolution in lines would be 3912 / 2 = 1956 since it takes a minimum
of 2 pixels of different color or brightness to start forming a line.
But that's the max theoretical from the sensor. If the lens is not resolving that level of detail or there is motion
blur then the maximum will not be reached.
Running the numbers is fine, and useful, but not always the whole story. When I upgraded from the APS-C 300D to a 5D, I
calculated that the theoretical resolution of the 300D is greater than that of the matching size central portion of the
larger FF sensor of the 5D.
However, as I suspected from review images, with actual test shots on a tripod of static subjects, using the same lens,
there was very slightly more visible detail in the lower pixel count shots from the 5D.
As scientists find regularly, the theoretical calculations may or may not match the experimental results. Sometimes the
theory has to go.
On 11/3/2014 8:29 AM, Chuck Norcutt wrote:
I agree with your first statement about 6x7 cameras vs digital. But your second statement (that it would take
stitching 100+ digital images to equal a 6x7) is grossly overstated. The resolution of a 6x7 is limited to the
resolution of the film available. Apart from Kodak Technical B&W the resolution of color and B&W film is limited to
200 line pairs/mm (for very high contrast targets). If we equate that resolving power to a digital sensor an
equivalent digital sensor would need 400 pixels/mm or 400 x 60 = 24,000 on the short size and 400 x 70 = 28,000 on the
long side for 672,000,000 or 672 MP.
Mamiya makes an 80 MP digital back for medium format cameras with image size 7,752 x 10,320 pixels. Assuming a very
conservative 30% overlap for stitching this could be accomplished with a 4x4 pattern and would actually make a
slightly larger image than the 6x7 film image since the aspect ratio of the digital image is different at 4:3.
Certainly not as easy as taking a single 6x7 film image but 100+ images is way out of bounds. Also, the dynamic range
of the medium format digital (12.5 stops) might give the film some stiff competition. Also, choosing such high
resolution film would limit the film's ISO to 25... a very severe penalty in actual usage. Choosing higher ISO films
(say 400) would cut the film resolution by a factor of 2 and the equivalent MP count by a factor of 4. Then it would
be 168 MP vs 80 MP with the digital still having a dramatic advantage in available ISO range.
It's my guess that in a real world test, a sensor like the Phase One IQ280 and/or the Mamiya Chuck mentions will resolve
about as much visible detail as any practical film in the Mamiya 67. When you get through shooting, processing,
scanning, Raw conversion, etc., and get down to peeking at the images, the fine detail of leaves, cloth, eyelashes, and
so on, the large difference in simple theory numbers will boil down to little difference.
In real world photography, at least outside a tightly controlled studio, the far better ISO range/performance and useful
dynamic range will whup film for actually getting the shot - and with high IQ.
Just one Moose's opinion, but based in a lot of experience with 35 mm film and digital sensors. Hasn't someone done this
testing? I don't care enough to search myself. As far as I am concerned, contemporary µ4/3 sensor systems outperform the
35 mm film I used to use in every respect.
Moose D'Opinion
--
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/
|