35mm film scanner questions

Discussion in 'Scanners' started by Allan, Mar 13, 2006.

  1. Allan

    Allan Guest

    I know there has been discussion on this before, but I now find myself in
    the market for a film scanner. I expect to scan about 5 films per week.

    Also, I have older Nikon bodies and lenses. Am I right in thinking that with
    a good film scanner I can expect similar photograph quality to say the Nikon

    What are the current thoughts on this - models, features etc?


    Allan, Mar 13, 2006
    1. Advertisements

  2. Allan

    george Guest

    Definitely not with what I've got (up to 3200 dpi)...my D200 wins hands
    down. I don't know if the orphaned 5400 II might or the Coolscan 5000 (but
    I'd be very surprised if either one could).
    george, Mar 13, 2006
    1. Advertisements

  3. Allan

    rafe b Guest

    No experience with a D200, but lots of experience with
    film scanning in general (35mm, MF, LF) and with a
    Canon G2 and 10D. Bottom line, I wouldn't bother
    shooting/scanning film in any format smaller than MF
    these days.

    IMO: scanned 645 (MF) roughly matches a Canon 5D
    and/or Nikon D2X. Scanned 6x7 (MF) might be matched
    by a Canon 1Ds MkII.

    95% of users will get better results from a Canon 20D
    than from 4000 dpi (Nikon Coolscan) scanned 35mm film.

    You've got to be really hard core to want to shoot and
    scan 35mm film these days. And if you're going to do it
    at all (scanning film, that is) with 35mm or MF, forget
    about doing it with an Epson film/flatbed scanner (yea,
    even the 4990.) It'll have to be a dedicated film
    scanner, 4000 dpi or better.

    rafe b
    rafe b, Mar 13, 2006
  4. Allan

    m4w3y3 Guest

    A 3200 dpi film scanner will produce around a 4480 x 3045 pixel file from a
    35mm frame (which is the equivalent of a 13.5Mp digital camera). It will
    produce that file form film used in a 35mm SLR or a 35mm point and shoot.
    There are similar anti-noise filters and grain-melting filters that make the
    scan as clear as anything that comes from a dSLR costing many times more
    than the scanner.
    However, my scanner will do 6400dpi which can give me a 160Mg TIFF file
    (8800 x 6000 or equal to a 53Mp digital camera) from a 35mm frame.
    I can scan 24 slides in a single batch and, once the files are saved, I have
    exactly the same editing capabilities in Photoshop (or other photo editing
    software) as any digital camera user. Of course, I can also scan medium
    format and large format (4x5) negatives too...at 6400dpi. (which is 25600 x
    32000 pixels or 820Mp).
    I think the greatest advantage that I find with scanning is the ability to
    rescan a great shot at a higher resolution. My first film scanner was
    1200dpi. When I got a 6400dpi scanner, I could take my best shots and rescn
    them at the higher resolution. If those were dSLR shots and I bought a
    higher resolution digital camera...there is no way that those older shots
    can take advantage of the new resolution. Sure, there are those phony
    interpolation programs but that is simply injecting false data into the
    scene to expand the pixel count...a higher resolution scan actually pulls
    more information from the frame.
    For affordable scanners...look at Microtek..they have a good range of very
    well reviewed scanners.
    m4w3y3, Mar 13, 2006
  5. Allan

    rafe b Guest

    I think so far we all agree, but I wouldn't want
    to overstate the case. With careful shooting and
    careful scanning, the results will be close and
    might even favor the film in some regards --
    but not by much.

    Relative "ranking" of the results would probably
    be subjective, and of course there are so many
    real variables that nothing definitive can really be

    Just for kicks, I'm going to make one or two 16x24"
    prints from 35mm scans this evening and have
    another look. It just so happens that all my prints
    at that size are from 10D captures. Just like my
    Nikons, my 35mm film scans aren't getting much
    of a workout these days either.

    One thing's for sure, the digital capture is a hell of
    a lot less work, at every step of the way. That's
    why I say one really needs to be hard-core to
    shoot and scan 35mm these days.

    rafe b
    rafe b, Mar 13, 2006
  6. Allan

    rafe b Guest

    What 6400 dpi scanner are you using?

    Is that 6400 dpi optical resolution, or interpolated?

    Most of us who have worked with both film scans
    and digital have learned that comparing pixel-counts
    from these two methods is rather meaningless.

    Furthermore, there's not much correlation
    (unfortunately) between scanner dpi ratings and
    their actual resolving power.

    For example: I own a Nikon film scanner rated
    at 4000 dpi, and an Epson flatbed/film scanner
    rated at 4800 dpi. The Nikon's scans are far
    sharper, even though they're at lower resolution
    than the Epson's.

    rafe b
    rafe b, Mar 13, 2006
  7. You'll get higher resolution probably (if you use the right film,
    lens, tripod, and scanner). But you probably won't be able to enlarge
    pictures as large and have them look good.

    Plus, of course, you'll spend a lot of money on the good scanner, and
    a LOT of time scanning. Plus the cost of film and processing. (The
    D200 is pretty expensive itself, of course).
    If you're seriously looking to compete with a good DSLR you'd better
    look at the Nikon Coolscan 5000 ED.
    David Dyer-Bennet, Mar 13, 2006
  8. Except it doesn't work out that way so your comparisons are exceedingly
    wide of the mark. For a start, the information you are pulling off of
    the film at that resolution is at a very low contrast - due to the
    vanishingly small MTF of the film, not to mention the original camera
    optics. Similarly, you gloss over the noise with some glib "anti-noise"
    filter statement without recognising that no filter can distinguish
    between signal and noise, so reducing noise also reduces the signal - of
    which you have already precious little anywhere near the limiting
    resolution of your scanner. Meanwhile the dSLR image has contrast and
    SNR in abundance right up to it resolution limit.

    In your scanned film image, the MTF at the nyquist frequency of the
    scanner is likely to be of the order of 2-3% at best, based on the MTF
    curve for Provia. By comparison, the MTF of the dSLR sensor at Nyquist
    will be around 65% - a figure that Provia could only produce at about
    35cy/mm, or just under 2000ppi.

    So most of those 160MB in your Tiff files are redundant information but,
    because it is mainly noise, doesn't even compress efficiently. It
    certainly isn't the equivalent of the 53Mp digital camera that your
    exceedingly naive assessment claims. More like 15Mp - if you are very
    lucky, and probably less than half that.
    Kennedy McEwen, Mar 14, 2006
  9. Allan

    m4w3y3 Guest

    I have found that whenever there is an attempt to directly compare digital
    and analog images, there is always something that renders the comparison
    'invalid'. I wonder if this is an attempt to protect digital from
    unfavorable comparisons.

    For example, my scanner will capture an image at a full 24bits/channel...I
    don't know of any digital cameras that come even close to that...what can we
    get...8 bits/channel? maybe more (theoretically with so-called RAW formats)
    and yet, I will bet that someone will argue that the reduced bits/channel
    figure of digital is somehow...'better'.

    For another example, the resolving ability of film and film camera lens has
    been measured in lpmm (line pairs/mm) for years. There are, therefore, years
    of data on lens and film resolution measured in lpmm...but digital cameras
    resolving potential is NEVER stated in lpmm thus making direct comparisons
    impossible. Instead, digital camera resolution is measured in LPH (or Lines
    per Picture Height)..a completely incompatible unit of measurement with no
    conversion factor. They say that this is to take into account the different
    sizes of digital sensors in spite of the fact that lpmm is used for ALL
    sizes of film. I think the reason for this is that digital cameras just
    wouldn't compare very favorably to film when the same units are used.

    But, people do claim that the results from their digital cameras are
    'better' than the results from film. I am wondering if there isn't some
    interesting psycho-engineering of preception going on here. Digital images
    with artificial sharpening, averaging out 'noise', and all the various
    enhancements that go into producing the image that fools the eye. Much of
    this is based on a set of assumptions (and you hear them quoted in many of
    these discussions) such as 'normal viewing distances' which means 'Don't
    look too close at my digital print!' On the other hand, I can look closely
    at my film prints or even through a 5x loupe and still see detail.

    One thing that digital people obsess over is 'noise' and they (and the
    manufacturers) go to extraordinary lengths to remove any hint of noise even
    if it means the loss of detail. It seems that, psychologically, a noise-free
    shot is perceived as clearer even if it contains less detail. Much of what
    people 'think' they see in digital images could be more illusion than
    reality...so don't look too closely!
    m4w3y3, Mar 14, 2006
  10. Really? And what do you do with this amazing dynamic range which, at
    7.2 is more than a thousand times more than the density range of the
    best film?

    Whilst the tiff format will support huge bit depths, there aren't too
    many applications that will interpret anything more than 16
    Kennedy McEwen, Mar 14, 2006
  11. LPH is a far better way to measure resolution, since it requires no
    conversion between formats. (Film comes in a lot of sizes, from Minox
    through 4x5, and some of us use several of them.)
    The P&S dcams are far better than film in terms of lp/mm for actual imaging.
    (Zeiss uses blue light and microfilm to get some insane numbers, but slide
    films have almost no practically usable response at 50 lp/mm while most P&S
    dcams show strong contrast at well over 100 lp/mm.)
    In my 6x7 vs. 5D tests, the 6x7 does better at high contrast detail. This is
    very noticeable in urban architectural shots where there are signs and fine
    architecturel detail everywhere from nearby to infinity. But for landscape
    and nature shots, where the contrast in the detail is a lot lower, there's
    no significant difference. 35mm is, of course, a pitiful joke compare to the

    David J. Littleboy
    Tokyo, Japan
    David J. Littleboy, Mar 14, 2006
  12. Nope; mostly it's attempts to protect film from unfavorable

    The important comparison to make is taking the kind of pictures you
    want to take, and making the size and kind of prints, web displays,
    slide presentations, or whatever other uses you want to make of your
    photos, and *then* comparing the digital and film versions. And of
    course you have to be equally competent at the relevant skills for the
    two sides, to make the comparison fair. Ideally the visual
    comparisons of prints should be made "blind", that is the person
    viewing the prints shouldn't know which ones came from which sources.
    I don't believe your 24 bits per channel. Some scanners achieve 14
    bits per channel (which are stored in 16-bits-per-channel files). I
    can imagine modern professional drum scanners perhaps achieving 16
    bits per channel of real data, maybe. What scanner are you using
    In measuring film camera and lens results, people mostly reported lpmm
    on the film, as determined by microscopic examination. Given the very
    wide range of digital sensor sizes, I think the explanation given
    makes sense, and lpmm doesn't make much sense for that use.
    And do you *care* about the detail you can only see through a 5x
    loupe? This is one of those decisions each artist needs to make;
    there is no right or wrong answer, really.

    Your choice of the phrase "artificial sharpening" shows me that you
    have a strong film bias, and are ignorant of digital imaging.
    Um, I hate to break this to you, but stuff you see through your eyes
    is heavily processed. The measure of image quality is how it works
    with the human visual system -- that is, the subjective experience of
    people looking at your pictures. Can you make much of an argument for
    any *other* standard of quality? (The only one I can think of is to
    be prepared for when we meet aliens, whose visual systems work very
    differently from our own. Okay, that's a reasonable argument; it just
    won't get very many people following you, I don't think.)

    I see lots of people complaining about the "plasticy" look of
    overly-noise-reduced images, rather than everybody running out and
    eliminating every hint of noise as you suggest.

    But I think you're *right* that a noise-free shot is perceived as
    cleaner than a noisy shot. And also, something you didn't quite say,
    that many pictures can have lesser detail and still look first-rate
    (especially if not compared side-by-side with the same picture, just
    as clean, but with more detail).

    But go back to my second paragraph. If you're in serious doubt (as
    many reasonable people *should* be) about whether digital or film will
    be better for some particular aspect of your work -- make the test.
    Take some pictures both ways, print them to the sizes you normally
    use, with the care you normally use, and show them around to a bunch
    of friends. Don't tell them which are which. Borrow equipment to do
    this if necessary. Possibly get help in post-processing and printing,
    if you're not an expert on both film and digital printing. See how
    *you* like the pictures, and how people you show them to react to
    them. Then make an informed decision about which will work for you at
    the moment. And you should then be reasonably confident that you're
    doing the right thing. At that point, laugh at people who claim
    you're wrong -- at least unless they can point out problems in your
    technique, or changes in technology since you made your tests.
    David Dyer-Bennet, Mar 14, 2006
  13. Allan

    m4w3y3 Guest

    Well, 48bit color amounts to 16 bits/channel so, that should be an easier
    number for you to accept. Still twice what I can expect from a digital
    camera. The other problem is that the act of scanning a film image degrades
    the image since the scanner can't hope to capture all of the detail. One of
    the errors that people who try to compare alaog images to digital images is
    that they think that making a scan of the analog image puts everybody on the
    same level...but it doesn't. It is a ploy to compare a degraded copy of a
    film image to an enhanced original digital image. It's a setup.
    m4w3y3, Mar 14, 2006
  14. Allan

    m4w3y3 Guest

    If that is true why is there resistance on the part of digital people to a
    fair comparison?
    m4w3y3, Mar 14, 2006
  15. Allan

    D-Mac Guest

    I have some Nikon ED 5000 film scanners in my business. I also have some 5D
    Canon cameras and a few Olympus "E" series DSLR cameras. I'll stick my neck
    out here and say that it is absolutely impossible to get a 30" x 40" print
    from a 35mm (any 35mm) film from any method - optical enlarging or
    scanning - that will equal or come close to equalling the sharpness and
    detail in a print that size enlarged from an Olympus E300, 8 megapixel DSLR
    image. The proof hangs on the walls of many happy clients of mine. Start
    considering large DSLR images and the gap just widens.

    There is constant misinformation spread about the process of Interpolation
    as applied to digital camera images and I won't get into it again other than
    to say if you really don't want to believe me when I say I do it every day
    for a living, that's the problem of small minded people who refuse to
    believe what they themselves cannot do, can actually be done. The pic on the
    index page of my canvas site is a picture of one such enlargement.

    It's not rocket science. NASA have been doing it for years. The single
    limiting factor with film is not how many lines per inch it can resolve but
    how many it can produce in a print. Optical enlargements have to contend
    with degradation of the light between the lens and the paper. Scanned film
    has to contend with not just grain but the texture of the film itself and
    the way in which a film image wraps around the individual grains and
    produces anomalies impossible to remove.

    this picture: http://www.photosbydouglas.com/film-digi.jpg shows
    dramatically and unretouched, the difference between a 100 ISO domestic
    Kodak film processed at Kodak's own "Pacific" lab and scanned on an ED 5000
    and a 20D image. No retouching here, no fluffing the results. The 20D
    incidentally had a back focus error later fixed by Canon which is the reason
    for the soft image. The film was shot in a Nikon F90 using a Nikon lens.
    Plenty of people have downloaded that image (without my permission) and
    enhanced the film part but that's not what this discussion is about. These
    are camera/scanner direct images cropped for the Internet.

    There is no chance a Nikon 35mm film scanner is ever going to consistently
    produce scans to equal the quality of a D200 image - ever. Anyone who say it
    will hasn't got any experience to qualify such statements.
    D-Mac, Mar 14, 2006
  16. Allan

    Allan Guest

    Well, I seem to have started quite a discussion. Thanks for all the replies.
    I will give the film scanner some serious thought

    It looks like I should just forget about it and invest the money into DSLR


    Allan, Mar 14, 2006
  17. Your film bias is showing again. There is no such resistance. It's
    just very hard to do in a repeatable objective fashion. Most people
    don't care enough to take the trouble; they can find out what works
    *for them* with much less work than doing the first-rate overall
    study, and that's all they care about.
    David Dyer-Bennet, Mar 14, 2006
  18. Me too. I upgraded my scanner 3 years after I got my DSLR, and
    something like 8 years after I got my macro lens.
    David Dyer-Bennet, Mar 14, 2006
  19. Any film scanner worth the name provides all of the features on your
    dSLR - and more. How, for example, does your dSLR distinguish between
    defects on the film surface, such as dust, dirt, or scratches and real
    image content? How does it remove these defects from your final image
    without affecting other similar image features? How does it correct for
    the uneven transmission of the macro lens across the field? How does it
    pull out shadow detail in the slide that is at densities of 3 or more
    without posterising them? Sure, some of these can be fixed with a lot
    of hard work on the dSLR output, others can't - yet, after a little
    practice, a decent scanner will wrap them all up in a simple single push
    button process.
    Aha - now it is clear that you have no idea what a decent film scanner
    is capable of, because you spent your cash on a "one-size-fits-all"
    solution rather than specialising.
    Kennedy McEwen, Mar 14, 2006
  20. This seals it for me!

    One of the first things most folk do when they get their first dSLR is
    compare how it performs against their tried and trusted film SLRs.

    Folks - I think we have been quite successfully trolled. Congratulations
    Marbing, you got most of the regulars in the group with that one .

    Your 6400ppi 24-bpc scanner should have been a clue to most of us that
    your pro-film position was just a wind-up.
    Kennedy McEwen, Mar 14, 2006
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.