35mm film scanner questions

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

  1. Allan

    rafe b Guest

    Actually, nobody with a brain claims digital has
    more resolution than film. It's generally the other
    way around. Trouble is, most of the extra resolution
    that's available from film is buried in noise and grain.
    Some of us came about it from the other direction.
    I got my first film scanner in 1998, and my first
    digital camera in 2002. And I haven't given up
    on film in general -- just 35mm film.
    Nobody with a brain refers to interpolation as "false detail."
    Or are you talking about the effects of inadequate anti-
    aliasing? Nobody with a brain would claim that an image
    captured on film is object truth, either. I've never seen
    grain in a real sky.
    You need to polish up your Photoshop skills, it would seem.

    Well, at least on that we agree.


    rafe b
    www.terrapinphoto.com
     
    rafe b, Mar 14, 2006
    #21
    1. Advertisements

  2. The macro lens might be just as good as the lens in a film scanner, but
    that doesn't make a DSLR plus slide copier the equal of a film scanner.

    Most film scanners use a linear array sensor, not an area array. That
    allows them to measure red, green, blue, and often infrared
    transmittance individually at every pixel position, instead of using a
    Bayer filter sensor. Measuring RGB independently increases colour
    resolution and avoids colour/luma separation errors. Measuring infrared
    is the basis of automatic dirt-removal methods. And getting linear
    arrays with 5000-6000 pixels across the height of the frame is not
    difficult, while DSLRs with comparable resolution are nonexistent.

    Then, some scanners use an illuminated slit light source that moves
    along with the sensor array (or the film moves and the light source and
    array are stationary, which amounts to the same thing). With this
    configuration, most of the flare light scattered by the lens from the
    illuminated line on the film is scattered outside the sensor and
    vanishes. Another way of looking at it: most of the flare that any one
    pixel in an area-array sensor like a DSLR would see with an area light
    source is not present with a line light source, because almost all of
    the film area is dark at any given time.

    Then there are drum scanners that illuminate only a small disc on the
    film at a time, not a line. They get even more optical flare reduction
    from this - essentially all flare light is rejected.

    The slower pace of scanning and digitizing also may allow the scanner to
    capture more bits per pixel. That, plus the better flare performance of
    the slit-illuminated optics, means you can see further into the darker
    areas of the transparency or negative.

    And of course the scanner provides full control of contrast and colour
    temperature. The colour temperature adjustment, in particular, is
    generally done by adjusting illumination or exposure time, *not* scaling
    the data after the A/D converter, so you get the full A/D converter
    range (unlike a digital camera).

    Finally, film scanners come with sofware that understands how to convert
    negative density back into original-scene light. If you photograph a
    negative with a DSLR, you have to do this yourself, and it's *not* just
    a matter of inverting the channels to get this right.

    All in all, a DSLR is not capable of matching a well-designed film
    scanner.

    Dave
     
    Dave Martindale, Mar 14, 2006
    #22
    1. Advertisements

  3. Allan

    rafe b Guest


    A proper film scanner has to be accurate in one dimesion: a line.
    Everything about its design takes advantage of that.

    A camera has to be accurate in two dimensions: an area.
    Everything about its design (at least as far as image-formation goes)
    revolves around that fact.

    This is also why film scanners are better at extracting detail from
    film than optical enlargers.


    rafe b
    www.terrapinphoto.com
     
    rafe b, Mar 14, 2006
    #23
  4. Allan

    AZ Nomad Guest

    How did you arrive at that theory? The pickup may be a line, but
    the axis that moves the pickup is another dimension. Make either one
    sloppy and you've got a bad scan.



    ditto for a scanner.
     
    AZ Nomad, Mar 14, 2006
    #24
  5. Allan

    rafe b Guest


    Think image circle. The sensing element in a scanner -- the CCD --
    only needs to fit neatly along the diameter of that circle. Whereas
    with film, the entire *area* of the film has to fit in that circle. So
    with a CCD, you can design for a smaller image circle, which
    almost always results in a better lens overall. Furthermore, the
    image is always at the center of the image circle, in in the axis
    of the film or CCD travel.

    Correction for light falloff with incident angle only needs to
    be done in one dimension, and is done scanline-by-scanline.

    The CCD on a good filmscanner has more pixels in its one
    dimension than most square or rectangular arrays have in
    either dimension.

    Example 1: Eos 20D: sensor: 3520x2344 pixels, but a Nikon
    Coolscan 5000 sensor is at least 4000 pixels long, and those
    pixels are aligned with the *short* axis of the frame.

    Example 2: Eos 1DsMKII sensor: 4992 x 3328 pixels, but
    a Coolscan 8000 or 9000 CCD sensor is 10,000 pixels long.
    Again, those 10,000 pixels are aligned with the width of the
    film, so scans of 6x7 or 6x9 cm actually produce images of
    The ultimate extension of this theory is a drum scanner, which
    really only images a point -- not even a line. It is, in effect, a
    microscope objective poised over the film, and imaging one
    microscopic circle at a time.


    rafe b
    www.terrapinphoto.com
     
    rafe b, Mar 14, 2006
    #25
  6. Allan

    AZ Nomad Guest

    You're ignoring the mechanism for positioning the drum and the sensor.
    Any inaccuracy in these mechanisms will ruin your image.
     
    AZ Nomad, Mar 15, 2006
    #26
  7. Allan

    rafe b Guest


    I did not mean to suggest or imply that "slop" is
    acceptable anywhere. I'm simply saying that the
    one-dimensional nature of the sensing element is
    important to the design in many ways -- and has
    advantages over imaging the original in two
    dimensions at once (as in an area CCD sensor.)


    rafe b
    www.terrapinphoto.com
     
    rafe b, Mar 15, 2006
    #27
  8. Allan

    rafe b Guest


    I've been scanning film since 1998 and shooting
    with digital cameras since 2002. No preconceived
    ideas here -- just lots of experience. I've probably
    used (and mastered) more different film scanners
    than you can imagine.


    rafe b
    www.terrapinphoto.com
     
    rafe b, Mar 15, 2006
    #28
  9. Allan

    HvdV Guest

    In addition to the ones mentioned by Rafe, there are at least two more
    factors which makes life easy for the scanner lens: fixed magnification, and
    not much need for high aperture. And of course no zoom. So indeed, a good
    scan lens can be cheaper than a, loosely speaking, 'equally good' camera lens.

    -- Hans
     
    HvdV, Mar 15, 2006
    #29
  10. You suggested that a slide copier was just as good as a film scanner for
    capturing images from film. Now are you saying that all you want is to
    capture what that slide looks like when it is projected or printed?
    The projected/printed image is inferior to the real world in a number of
    ways, and inferior to what a digital camera would capture if it had been
    used in place of the film camera.

    The job of a scanner is not to "enhance" the image, it's just a matter
    of capturing as much of what the original camera and film saw as
    possible. If a slide copier plus DSLR does a poorer job at this, then
    it makes a poorer film capture device, doesn't it?
    Well, the image from a good scanner doesn't have the effects of lens
    flare reducing detail and contrast in the shadow areas that you'd get if
    you optically projected the slide. And it doesn't have the compressed
    contrast you'd get from printing. But why would you *want* those
    distortions in a film digitizing device?

    Dave
     
    Dave Martindale, Mar 16, 2006
    #30
  11. Yes, it looks like CCD noise to me. But I'm just killfiling larger and
    larger bigpond netblocks to avoid him.
     
    Philip Homburg, Mar 16, 2006
    #31
    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.