filmscanner vs hi-res flatbed

[Kennedy McEwen]

I have seen your other post and I do understand it. But in the past I have seen
a lot of tests which proved that the "real" resolution of flatbed scanners is
only about half the resolution compared to the spec. Are these tests wrong, or
what is the reason for this result?

The tests are correct, but their (or your) use of the term "real
resolution" is incorrect. The difference is what MTF the tests use as
their limit of resolution. As explained in the other post, the flatbed
approach uses a design which guarantees that the MTF is near zero at the
resolution limit of the sample density. So, if the test calls the limit
of useful resolution and MTF of 10% then it will yield a measure of
resolution which is less than half the sampling density of the flatbed,
but equal to (and in some cases more than) that for the film scanner. If
the test calls the limit at 20% then the result for the film scanner
will be even lower still. However, the "real" resolution limit is the
point at which the MTF falls to zero, and that is extremely difficult
(but not impossible) to measure.

Another issue is that many of the tests insist that any image sharpening
is turned off before measurement on the grounds that this "evens the
playing field". However it does quite the opposite, biasing the
measurement in favour of the dedicated film scanner. If you followed
the other post you will also understand that the area of each cell in
the staggered array is approximately 4x that of the equivalent cell in
an unstaggered array. This means that the staggered array has around
twice as much intrinsic signal to noise (all other things being equal).
As mentioned in the other post, the MTF in the useful band is reduced
because of the desire to avoid aliasing and this should be compensated
by post scan sharpening, trading that improved noise floor for
increasing the effective MTF in the useful band.

The playing field is level when the intrinsic advantage of the larger
pixel has been compensated by the increased noise of the sharpening
filter with a staggered array, not when the filter is switched off.

A final issue is that the image that is usually used as a reference is
that produced by a linear, undersampled scanner - one which has a
significant MTF at and above the limit imposed by the sampling
resolution. Whilst such a scanner will certainly resolve the test
image, the excess contrast and consequential aliasing set an unrealistic
comparison for a critically sampled image - one where the MTF reaches
zero exactly at the limit imposed by the sampling resolution.

BTW, because of practicalities such as fill factors being less than
100%, the zero point of the MTF is generally quite a way above half of
the sampling resolution, but I tried to simplify the general principle.

 

[Peter D]

No, you will find some consumer flatbeds run up to 4800ppi.

Quite. Thank you for refining my words.

Stacking the CCDs at half pixel, as you put it, does not mean that the
resolution is limited to the original unstacked device - quite the
opposite.

Quite. The point I was making. Thank you so much for elucidiating.

It is quite simple really.

Really. It doesn't look or read like it. But thanks for the lesson anyway.

 

[Peter D]

There are flat beds available for several thousand dollars that will
rival in quality the scan obtainable for an under $1000 hobbyest
scanner.... but they are true exceptions and wouldn't be considered in
this instance. But in general, a good flat bed scanner sells for half
of what a good 35mm film scanner sells for, and doesn't begin to reach
the scan quality. As for a "cheap/low quality" film scanner.... well
it's always possible to find some products that are simply worthlsss,
so yes, you can find a cheap film scanner that will perform poorer
than a good flat bed. But why would anyone wnat to do that?

I'm not suggesitng anyone would want to do it, but that the absolute "No" is
innacurate.

 

[Peter D]

If you look around the web you can see many tests and reports of the the
apparent resolution of flatbed scanners being about one half the stated
optical resolution.

And most such 'reports' are designed to skew the results in favour of the
'reporters' position. Namely that film scanners are always better than
flatbed scanners. Someone has already presented a more accurate review in
this thread.

My film scanner (Minolta ScanDual IV) has a fixed sensor and it moves the
film to scan in a single pass.
Image sensor: 3-line primary-color CCD with 5340 pixels/line.
Optical input Resolution: 3200 dpi

But so what? Using less than the whole scanning bar is irrelevant unless you
can demonstrate that using less than the whole scanning bar is necessary.
IOW, if it's irrelevant, it's irrelevant.

 

[Charlie Hoffpauir]

I'm not suggesitng anyone would want to do it, but that the absolute "No" is
innacurate.

You did not quote the rest of my post. The answer was qualified in the
later statements.

I think that your finding fault with my post is a bit rude. Why ask
for replies if you want to try to pick apart the resopnses that you
don't happen to agree with?

Thats the same as someone saying "no" to the question of "Is it
raining?".... and then the questioner says well, maybe not here, but
is is certainly raining somewhere! In that case, yes, my "no" answer
is innacurate.

Remember, the question was "is the quality comparible to that of
dedicated film scanners?" Is "yes" a more accurate answer? I happen to
own one of each (Epson flat bed model 3170 Photo and Nikon Coolscan IV
(neither one very new), and from my experience, the quality of the
Epson scan is NOT comparible to the quality of the Nikon scan on 35 mm
film.

If you don't like that answer, fine. go ahead and buy the flat bed,
and you'll probably be quite happy with it.
Charlie Hoffpauir
http://freepages.genealogy.rootsweb.com/~charlieh/

 

[CSM1]

And most such 'reports' are designed to skew the results in favour of the
'reporters' position. Namely that film scanners are always better than
flatbed scanners. Someone has already presented a more accurate review in
this thread.


But so what? Using less than the whole scanning bar is irrelevant unless you
can demonstrate that using less than the whole scanning bar is necessary.
IOW, if it's irrelevant, it's irrelevant.

Peter, you asked for opinions, you got opinions.

If you do not like the answers, or the answers are not what you want to hear
or read, then that is fine.

The only opinion I express are my own, and I sometimes post links to sites
that have more information. Nobody is forcing you the accept or even use the
statements made on this public forum on Scanners.

If you want to buy and use a flatbed scanner to scan 35 mm Film, that is up
to you. It is your money and your film.

The fact is a dedicated film scanner does a better job of scanning film,
especially 35 mm film.

The newer Canon and Epson scanners claim up to 4800 dpi optical, but they
still do not compare to a Nikon Super Coolscan 5000 ED 4000 DPI scanner.

If you want to see real pictures by people like yourself and get comparisons
of scanners, goto:
http://www.photosig.com/go/photos then on the left side, click on Browse by
scanner.

 

[Peter D]

You did not quote the rest of my post. The answer was qualified in the
later statements.

Which post? Which later statements? Someone asked, "Is the quality
comparable to that of dedicated filmscanners?". You said, "No." Not "Not in
most circumstances." Not "No, not at the same consumer level.". Not, "Not
most of the time" or "Most times, no. Rarely, yes.". You said "No." "No." is
very absolute (and inaccurate).

I think that your finding fault with my post is a bit rude. Why ask
for replies if you want to try to pick apart the resopnses that you
don't happen to agree with?

Whether I agree or not is irrelevant. I asked for clarificaitonre an
absolute statement.

Thats the same as someone saying "no" to the question of "Is it
raining?".... and then the questioner says well, maybe not here, but
is is certainly raining somewhere! In that case, yes, my "no" answer
is innacurate.

It's not the same as this at all. Not even close to it.

Remember, the question was "is the quality comparible to that of
dedicated film scanners?" Is "yes" a more accurate answer?

Of course not. It's as inaccurate as a "No" for the same reason -- it's an
absolute when an absolute is not an accurate response. If you (or anyone
else) had said "Yes." I would have asked exactly the same question. If I had
said, "Yes" I hope someone would ask me to clarify.

PS. Whether I like the answer or not is irrelevant.
PPS. It's not personal.

 

[Peter D]

Peter, you asked for opinions, you got opinions.

Actually I asked for something more concrete than opinions. I asked for
verifiable facts.

If you do not like the answers, or the answers are not what you want to
hear
or read, then that is fine.

Regardless of what you think I like or don't like, or what I do or don't
like, it's irrelevant. Facts aren't.
HTH (doubt it will)

 

[Polar Light]

Which post? Which later statements? Someone asked, "Is the quality

comparable to that of dedicated filmscanners?".

That would be me. Most of the info I got seemed to indicate that the quality
obtained from a dedicated unit would be superior to that of a flatbed.

You said, "No." Not "Not in most circumstances." Not "No, not at the same
consumer level.". Not, "Not most of the time" or "Most times, no. Rarely,
yes.". You said "No." "No." is very absolute (and inaccurate).

It looks like, overall, the answer is 'no', but there are always exceptions.
Last year I bought a cheap dedicated 35mm scanner & returned it due to poor
quality. I've never used a flatbed to scan film so I can't compare but it's
likely that a top-end flatbed *could* do a better job than that unit, if
that wasn't the case, most users wouldn't be satisfied with the results from
their flatbeds.

The overall conclusion seems to be that flatbeds are more versatily because
you can scan film in formats larger than 35mm AND prints too. Dedicated
scanners that handle medium format film cost nearly 10 times as much, so if
you have a variety of film sizes they're definitely the way to go (unless
you're wealthy or setting up a service bureau), however dedicated scanners
produce better images from 35mm film.

I have only 35mm film and already own a multi-function
scanner/printer/copier that came free with my PC, I can use this to scan
prints, the 600 dpi resolution is enough in most cases. I find scanning
printed stuff at higher resolutions often only brings out the pattern of the
paper and/or the ink dots rather than more detail.

 

[Hecate]

That would be me. Most of the info I got seemed to indicate that the quality
obtained from a dedicated unit would be superior to that of a flatbed.


It looks like, overall, the answer is 'no', but there are always exceptions.
Last year I bought a cheap dedicated 35mm scanner & returned it due to poor
quality. I've never used a flatbed to scan film so I can't compare but it's
likely that a top-end flatbed *could* do a better job than that unit, if
that wasn't the case, most users wouldn't be satisfied with the results from
their flatbeds.

Most users are satisfied with the results from their flatbeds because
they don't know and any better and/or have never been a position to
compare scans and/or wouldn't know how to get the best from a film
scanner in the first place (most flatbeds except the very top end are
very automated).

--

Hecate - The Real One
(e-mail address removed)
Fashion: Buying things you don't need, with money
you don't have, to impress people you don't like...