New scanner discussions (Epson 4870 and Minolta 5400)

Discussion in 'Minolta' started by Robert D Feinman, Feb 19, 2004.

  1. I've added some image examples to my discussion of the Epson 4870, just
    scroll to the end to see the new images.
    In addition I've added an additional discussion about getting the best
    colors from the Minolta 5400 when scanning color negatives. This time
    using SilverFast scanning software.

    Follow the tips link on my home page.
    Robert D Feinman, Feb 19, 2004
    1. Advertisements

  2. Robert D Feinman

    JR Guest

    Nice comparison. The only reason I have not made a move to medium
    format is affordable professional scanning. My 5400 does an excellent
    job on 35mm and I didn't want to buy a new system AND a new scanner. Bu
    the way your Silverfast section link is broken. But maybe you can tell
    me which you prefer?

    JR, Feb 19, 2004
    1. Advertisements

  3. SNIP
    The link to the Silverfast/5400 page doesn't resolve, unless changed to all
    lower case.

    Bart van der Wolf, Feb 19, 2004
  4. I just received my 4870 Yesterday. It replaces my Perfection 3200 and first
    impressions are that it still takes over 30 minutes to scan a 4"x5"
    negative! In all seriousness... If this is "the ideal scanner for
    Professional film scanning" (Epson's words). I probably ought to change my
    occupation! The trade off - and there always seems to be one- is that along
    with the increased scan quality comes an increase in noise (grain).

    The DMax of 3.8 might impress some people but to me, I want to see a 100 ISO
    film scan without monster grain evident everywhere. I want to see a scan
    that allows me to print out at the same or better quality as when I enlarge
    from the negative.

    I want all this to happen at a speed somewhat equal to or faster than doing
    it with chemicals. I'm afraid this scanner may turn out to be very little
    different than the one it replaces. If that happens, I've wasted the cost of
    a new lens on the thing! I'll report more as I get further into it. Not
    impressed so far.

    Biggest disappointment to date is the Digital ICE I bought it for in the
    first place (seems) to only work with the rather crude Epson interface. I
    much prefer Silverfast but it doesn't pop out and hit you in the eye,
    exactly how to use the Silverfast interface and Digital ICE. Maybe it does,
    maybe it doesn't. Hopefully it will. More as the adventure unfolds.

    Douglas MacDonald, Feb 19, 2004
  5. Robert D Feinman

    Alan Browne Guest

    I don't know if Vuescan supports that scanner, but Vuescan's IR channel
    procesing seems to work well (it is not ICE, it is Vuescan's use of the
    same IR channel for ICE like correction.

    I have yet to see a really good review of an Epson scanner. I'm not
    saying this to say they are bad, it's just to say that people who buy
    them tend to be disappointed. IOW: the expectation/performance is
    innorinately high. Whether this is the fault of Epson pushing too hard
    or the purchasers expecting too much...

    I recomend comp.periphs.scanners for true expertise on the subject.

    Dmax and grain are not directly related. Dmax gets you the "color"
    range and density. Grain is a function of resolution and in some cases,
    grain aliasing, for which there are strategies.

    Nikon's 8000/9000 are of course very well received and regarded, and I
    believe the equivalent Minolta MF sized scanner is just as good (as the
    8000, anyway), but not sure if any of these work to 4x5".

    I'm not sure why you mention the 5400 in your subject and/or not sure
    where it becomes part of your posting (if at all). the 5400 is limited
    to 35mm film (or smaller if you care to work out getting it into the
    film holders...).

    Alan Browne, Feb 19, 2004
  6. It wasn't me who mentioned the 5400 but the OP I replied to.
    Alan, I think you have a lot more experience than me with scanners so
    whatever you say I take seriously. I have bought a lot of flatbed scanners
    over the past year for clients ranging from Highly regarded Professional
    Photographers to people just wanting to scan some photos for Emailing.

    The most dissapointing aspect of FB scanners when scanning film has always
    been dust and small scratches and the somewhat limited way the software
    handles it. Digital ICE IMO is the only process able to correct dust as it
    scans. Canon use some cocked up affair that is useless and Silverfast is
    hard to use for anyone not familiar with European methods but just the same,
    very good when you do get the hang of it.

    Having said that, I was mildly satisfied with the Epson 3200. Way more
    satisfied than with any Canon scanner I used. I have many satisfied
    customers with my large and very large photos on their walls which I
    produced from scanning 6cm x 9cm Fuji NPS negatives on that scanner. I
    always had problems with dust and when Epson released the 4870 with Digital
    ICE, I just had to have one.

    I am not yet dissapointed with it although I had to replace the power supply
    in my PC before it would scan neagtives (yeah- figure it out!) I have only
    scanned one frame and it was so grainy/noisy as to be unsuitable for it's
    purpose. The 3200 did much better in that regard on that frame. Early days
    yet but I do know using flat bed scanners for film is an artform few of us

    I will take your advise about vuescan, If they don't support the scanner
    now, they will shortly, I'm sure. Although I personally am having problems,
    none of my clients who already bought them have called me with complaints.
    One can only presume it's like the guy who bought an EOS1 from me and came
    back complaining it didn't take very good pictures! It's up to the
    Photographer to make the gear produce stunning photos. I'll report more as I
    learn more.
    Douglas MacDonald, Feb 19, 2004
  7. Robert D Feinman

    Gordon Moat Guest

    I just got a brochure from Epson, with that scanner, and two others. The
    other two are the ones they suggest for professional work, at least for
    those who want more than inkjet prints. The other two are the 1680
    Professional, and the 1640XL Graphic Arts model. Both have a true
    focusing system, rather than the fixed focus of the 4870 Pro.

    Now, having stated that, I should mention that I know "professional"
    photographers and graphic designers that use the 3200 and other similar
    lower specification scanners. Bottom line is that if it gets the work
    done, makes the clients happy, and leaves room for profits, then it is
    "professional". Not many of us can afford a Creo iQSmart flat scanning
    I really have to wonder about the various DMax and DMin claims. I have
    yet to see anything other than high end systems, or drum scanning
    systems, that report a true value. It might help if there was an
    industry standard that all the manufacturers agreed to follow, but it
    has not happened.
    Best of luck with your gear. I hope you find a nice working solution.


    Gordon Moat
    Alliance Graphique Studio
    Gordon Moat, Feb 20, 2004
  8. Robert D Feinman

    Alan Browne Guest

    I'd love to see a scanner standard related to commonly available targets
    (Kodak et al).

    I don't know if that is as good as you would need, Gordon, but something
    -Drange for a dozen color points on the reference
    -Drange for 11 white-grey-black points
    -color accuracy (not sure how to define this).

    Would be a great help.

    Alan Browne, Feb 20, 2004
  9. Gordon Moat wrote:

    It was my impression that these were theoretical figures, having to do with
    information capacity rather than sensitivity; 16 bits per channel per pixel
    yields a DMax of 4.8, and that sort of thing. Is this not the case?

    Bill Tallman
    William D. Tallman, Feb 20, 2004
  10. I downloaded vuescan as per your recommendation Alan.
    I've got to say it really is THE scanning software for this scanner. Full
    report coming up on in a week or so. The current version of
    vuescan doesn't support IR dust and scratch removal but the developer of the
    software already has a test in progress to provide it. The file I tried
    unfortunately crashed on attempting a preview but it's early days yet.

    I scanned a 6x9 neg at 4800 dpi and after noise removal plus a little
    Photoshop work, has printed at 23"x13" from my S9000 inkjet with very good
    to excellent results. The full jpg file is nearly 50 Meg in size and has
    enough data in it to get a print five foot long off a Durst Lambda printer.
    That happens next week! Thank you for the suggestions Alan, right on target
    to! My purchase of the scanner is vindicated. If the big print is as good as
    I expect, the scanner paid for itself yesterday.

    Douglas MacDonald, Feb 20, 2004
  11. Robert D Feinman

    Alan Browne Guest

    The declared value by the manuf, yes. Subtract 1.5 bits for noise... so
    a 16 bit becomes a 14.5 bit...

    log(2^14.5)= 4.36 ish

    Various claims for slide film put the Dmax at around 4 to 4.2, so a 16
    bit scanner should manage quite well ... but...

    Alan Browne, Feb 20, 2004
  12. Apparently a couple of the links to my discussions were incorrect.
    I think all are fixed if you need to check them again.
    Robert D Feinman, Feb 21, 2004
  13. SILVERFAST SE or Ai from LASERSOFT IMAGING hard to use ??? NOT ANY

    SilverFast SE & Ai version 6 is the first software to overcome the
    difficulty of operationg sophisticated software by building QuickTime
    movies directly into the software. Each dialog now has a QuickTime
    movie explaining the each function with a large movie.
    If you, for instance, want to know how the Histogram is working, click
    on the QuickTime movie button inside the histogram dialog and the
    movie comes up explaining how to work with the histogram. You do not
    leave the software, the histogram dialog stays open, so what you have
    learned you can immediately apply it right there.
    Whatever you want to know, now the live explanation is right there
    where you are at!!

    Get a taste of the integrated tutorial movies right here:
    Karl-Heinz Zahorsky, Feb 21, 2004
  14. Robert D Feinman

    MikeWhy Guest

    I'm getting full-frame 3200 dpi scans in about five minutes through
    firewire. While not lightning fast, it's also not molasses slow. (Not sure
    about the five minutes; I just know the 30 minutes is about a magnitude

    I tinkered briefly with it on a USB port. It turns out XP really dislikes
    Intel's USB 2 driver, so I was using USB 1.1 instead. That takes about a
    half hour for a scan. If I had to guess, this is likely to be your problem.
    It might also be swap space problems with Photoshop. It can be any number of
    things, but scanner capability is not one of them.
    Again, I'm getting good clean scans at 2400 and 3200 dpi with a 3200. These are largely
    untouched. (Never mind the aesthetics, okay? Those are
    film/lens/camera/scanner tests. )

    Not sure where Dmax comes into it. Very dense transparencies will give you
    problems, but even bad negs, within reason, scan fine all day long.
    100 megapixel 48-bit scans are double-o-slow in Photoshop. Even so, I find
    it far faster from scan to finished print than chemicals can ever be.

    Tell us more about your setup. If we can't help, maybe one of us can take
    the 4780 off your hands.
    I hear D-ICE makes it about compatible with your USB 1.1 speeds. That is, it
    takes about a half hour for a 35mm scan. I'm presuming ICE isn't what's
    holding up your scan speed; you said "still take over 30 minutes..."
    compared to the ICE-less 3200. It's a shame the included software doesn't
    (seem to) work with the hardware features.
    MikeWhy, Feb 21, 2004
  15. Alan Browne wrote:

    Yep. But, the device has to actually be able to extract the information
    first before it can encode and transmit it. And therein, I take it, lies
    the rub.

    Where did you get the 1.5 bits factor for noise?

    Bill Tallman
    William D. Tallman, Feb 21, 2004
  16. Does it also explain how to get a properly working SF5.5 Ai version for
    the Nikon LS-30 without having to waste money on an endless succession
    of expensive upgrades for the rest of one's life?

    "Full color-corrected 48 bit output with Nikon scanners", they said...
    "Buy now!", they said...

    Ralf R. Radermacher, Feb 21, 2004
  17. Robert D Feinman

    Alan Browne Guest

    Engineers I worked with once upon a time (so I knew it a decade ago, and
    the decade old info had "decayed" ;-) ). When shopping scanners again,
    I called a couple up to ask them about D/A noise figures, and it appears
    pretty much standard practice when designing a digital sampler to
    consider the lowest bit to be noise and the next lowest to be noise half
    the time in computing the system performance. hence 1.5 bits.

    So shine a light through the film, put the detectors on the other side
    connected to the D/A and sample away. Assuming (assume again) that the
    detectors are linear and that the drange is within *their* spec, there
    should be no problem. The cool thing here, vice most digital cameras,
    is you get to sample R,G and B at every point (assuming you have low
    registration error), and so the color information is very deep. Color
    errors would come from the filters, the detector sites (at least all of
    the pixels in a scan row come fromt the same three detectors...), etc.

    Where it is tough, is the blacks (slide), the high order bits are
    useless, the "black" is compressed into the lowest order bits... hence
    the need for more bits overall, just to get the dynamic range in the
    bottom and get out those shaddow details...

    Alan Browne, Feb 21, 2004
  18. Robert D Feinman

    Alan Browne Guest

    Glad I could help. IAC, many would reccomend the same to you, and Ed
    Hamrick's support of the product is exemplary, as you will see if you
    hang out at comp.periphs.scanners where he often drops in with advice or
    discussions on improving scanning and the product.

    The IR channed certainly does work for the Minolta 5400 (and others
    [Minolta, Nikon...]...) so it hopefully won't be a long wait...

    Alan Browne, Feb 21, 2004
  19. Robert D Feinman

    Gordon Moat Guest

    I think that list is an excellent start. Gretag made some good charts,
    which could be flat scanned. There are also the Kodak Q-60 targets,
    which I think includes a 4" by 5" target.


    Gordon Moat
    Alliance Graphique Studio
    Gordon Moat, Feb 23, 2004
  20. Robert D Feinman

    Gordon Moat Guest

    You probably have that correct with what the manufacturers actually
    list. This might explain some scanners with lower stated Dmin/Dmax
    ranges actually performing better.


    Gordon Moat
    Alliance Graphique Studio
    Gordon Moat, Feb 23, 2004
    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.