I agree and also think the practical limits for most films is probably
somewhere around the 3200 dpi mark because the size of the grain
ultimately becomes the limiting factor. But most of us can't scan at
6400 dpi or higher to capture the 3200 that might be there. I say might
because I well believe that most photos don't contain more than 2400 or
thereabout based on the system limitations of photographer, lens,
camera, vibration, wind, subject motion, etc. etc.
Chuck Norcutt
On 7/20/2012 4:59 PM, Moose wrote:
> On 7/20/2012 7:33 AM, Chuck Norcutt wrote:
>> I wasn't talking about your specific results, only pointing out that
>> most photos end up with less than 50 lines/mm based on the
>> photographer's technique and environmental factors such as subject
>> motion, wind, vibration... not the film, lens or camera. Capturing that
>> much is actually difficult. 30-40 lines/mm would be more typical of hand
>> held work. If you check some of Modern Photos lab test resolution
>> results you will find the Zuiko 24/2.8 hitting 80 lines/mm at f/5.6 but
>> only 50 at f/2.8 and 63 at f/16. The Zuiko 135/2.8 ranges from 45 to a
>> max of 56 at f/8-11. The Nikkor 28/2.8 gets from 48-60. And that's in
>> the lab with a high resolution target and high resolution film on a very
>> heavy duty support.
>
> Notice I only referred to scanner dpi. I have no specific info about how lppi
> of b&W resolution charts relate to the dpi
> of the scanner. AG has had a lot to say about that that makes sense to me.
>
>> Also, you can't actually say that your own photography is producing more
>> than 2400 dpi (47 lines/mm).
>
> Again, you make a hard definition of how lppi and dpi are related. It may be
> mathematically correct, but to my mind is
> subject to variables such as AG has detailed. Additionally, it is dependent
> on contrast, which is reduced in the process
> of digital sampling, which means the results of scanning may differ from
> those of analog, visual measurement as the old
> tests were conducted. It's a reasonable surrogate for lens ability to resolve
> detail, but not an accurate measure of all
> aspects.
>
> Tests using color shots of complex subjects, say leaves in dappled sun/shade,
> for one example, might well show different
> relative results between lenses (and films/sensors/scanners). We wouldn't
> like them, of course, because they wouldn't
> produce exact seeming numbers for comparison. This is one reason those lenses
> with the best reputations are not always
> those with the best test results.
>
> So no, I can't say with reliability that my own results meet or exceed a
> specific number. I can say that I have handheld
> and tripod mounted images from nature taken with lenses for which I also have
> carefully taken, heavy tripod mounted
> images and for which there are lppi results available where the field results
> appear to resolve as well as the more
> careful images. I don't keep good notes, and it's been a while, but I recall
> concluding that my best film results were
> likely about 3000-3200 lppi. Could I be wrong? Sure.
>
>> You can only say that the film scanner outperforms the flat bed (which is
>> claimed to have higher resolution).
>
> And that's all that really matters - to the original questions raised in this
> thread.
>
>> The actual resolution of either device and the film you're scanning is
>> unknown to you without specific resolution based testing.
>
> I agree BUT I'm gonna contend again that that supposedly hard number just
> isn't. Perhaps more accurately, that it isn't
> a reliable overall measure of the ability of a component or chain of
> components to resolve detail in complex subjects -
> to the naked eye. I'll go further and assert that not only old fashioned lppi
> testing and newer MFT testing are less
> meaningful than they appear for the entire subject to display image chain for
> the simple reason that post processing is
> capable in the majority of images of increasing overall and edge contrast,
> and thus visual detail, without much, if any,
> other adverse effects.
>
> Visual 'sharpness' isn't a defined measurement. It's a combination of
> contrast and resolution - and human vision.
>
> I can say that scanning at 4000 dpi on the FS4000 resolves more detail in
> complex, natural subjects than scanning at
> 2700 dpi on my older scanner and more than 4800 dpi on the flatbed 9950F. I
> can also say that the 9950F resolves a bit
> more than the 2700 dpi film scanner, because I have Raw scans from that
> scanner of film I still have, and have made
> comparisons.
>
> I can also say that multiple pass scanning on the FS4000 resolves even a bit
> more detail, indicating there may be still
> more that higher sampling frequency might reveal. Per AG's comments on
> sampling frequency and resolution, that doesn't
> say anything specific about lppi resolution numbers.
>
> I seem to recall that Victor's conclusions from testing a lot of scanners was
> that film resolution probably topped out
> somewhere around 3200 dpi. But per AG's comments about audio sampling and
> waveform shapes, that doesn't necessarily mean
> that even much higher sampling frequencies might not give improved detail.
> Straight, sharp edged, B&W bars will never
> tell the whole story.
>
> More Detailed Moose
>
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|