Minolta Scan Elite 5400 is SLOW with ICE!

Discussion in 'Digital Darkroom' started by drjedsmith, Dec 29, 2005.

  1. I recently sold off my Nikon Coolscan V because I didn't like the way
    it did B&W film...and switched to a Minolta Scan Elite 5400 (version
    I) for its B&W performance.<BR><BR>
    I must say, the quality of scan from the 5400 is indeed impressive.
    But try turning ICE on when scanning a color film, and my goodness!
    I will be here a month of Sundays now. Wish there would have been
    some write-ups about this speed difference - maybe there is, and I
    just missed them.<BR><BR>
    At any rate, I just figured I'd warn the next person who's debating
    this...if you're going to scan a lot of color, get either the
    Coolscan V or the 5400II...<BR>
  2. jtk


    You made a big mistake dumping the Nikon.

    B&W is superb with Nikon V if you use Vuescan, or use the Nikonscan neg-as-pos workaround.

    Your problem with B&W was Nikonscan software, not the machine. Nikon V also accepts the Scanhancer if one uses the FH-3 film holder, which holds strips very flat. 2.5 minutes WITH Ice or Vuescan's infared.
  3. Jedidiah, if you are multi-tasking (say using Photoshop at same time as scan), and you have the capability, use the firewire connection. If you don't have the hook-up, consider adding it. It makes MSU much less subject to slowdown, when you have other programs open.

    If you are scanning color slides, perhaps try turning off auto-exposure, at least if your workflow involves post-processing with Photoshop or Vuescan. Your times will then be a constant 5~6 minutes. Color neg film is a different beast, and will take a minute or two longer, per scan. There are a few "novel" approaches to color neg scanning. For some, check out this thread:


    If you missed the reports of slow scan times with Elite 5400 and ICE, you could *not* have researched in much depth. LOL

    My current (big) project is slides. I have 3 slide holders on hand, and they run thru at 20~25 minutes per. That is manually focussing each frame, at time of scan. I've currently got 15 rolls scanned, which took me about a month. I did a bit of post-scan clean-up as I was scanning. But the scanning process was getting so far ahead, that I've just shut it down, and have just been Photoshop cleaning my scans, for the past 6 weeks or so. In short, if you intend to do post-scan cleaning and processsing, the speed (or lack thereof), will not be the limitting factor.

    And don't sell the scanner short. The grain dissolver is sweet, as is the delicious 5400 dpi results. Be patient and give it a chance.
    FWIW, I only output 16 bit linear of my slides (with auto-exposure off, and exposure sliders almost always zeroed, apart from the occasional, grossly underexposed slide). I post process these files thru Vuescan, employing them as Vuescan raw files.
  4. Slow scanning with Digital ICE is a processing issue, not the scanner. A faster computer will shorten this time. If possible, use the scanner software as a stand-alone program instead of as a Photoshop plug-in. Photoshop takes too much of the computer's resources just to remain standing.
  5. og


    Although I also recommend the stand-alone program for scanning (stability and memory issue), slow scans with ICE/GD/Negatives are not a CPU issue: all you need for those is a 800Mhz CPU (the reported 100% CPU usage doesn't mean anything).
    Check also this thread (end of first message).
  6. jtk


    Eric's not entirely wrong to worry about connections, but the discontinued 5400 is regularly reported to run 4-6 times slower, with Ice, than Nikon V which is in turn may be half as fast as the Nikon 5000 (5000 @ 1 minute, V @ 2.5 minutes, 5400 @ 10+ minutes).

    If these numbers seem wrong, time your own and report the results...

    My Nikon V and former Minolta 5400II scanned at about the same speed, 2.5 minutes with Ice. My humble Sony 2.8G/1G Pentium IV / USB2 is probably typical.

    Someone recently said the old 5400 is very fast without Ice, perhaps a minute with silver B&W...perhaps faster than the new machine? The Nikon V takes about 1:45 without Ice (or infared) with every type of film. If you're only scanning silver B&W and if you think the old Minolta's light source is special, it might be a good bet at the right price, though the IV might be even better if one installed a Scanhancer...since you can get them for around $250 with used warrenties.

    The Nikon V can use Scanhancer, a light diffuser, like on the old 5400...the 5400II apparently cannot.
  7. jtk


    Pardon me, I meant that Edward isn't entirely wrong. "Erik" is somebody else :)
  8. I feel your pain. Although I love the results that the 5400 puts out (and feel it is superior
    to any Nikon scanner that I have seen), the long scan times for negative film, especially
    with ICE, are a real pain. I solved the problem by switching to color slide film, which scans
    in 25% of the time required for negs. When I want B&W, I convert in Photoshop. This
    solution may not work for you, if you truly want to shoot B&W film.
  9. First things first - thank you to each contributer here. Things are looking up.<BR><BR>
    John, the V is a fabulous scanner for color films. I wish my only problem with the Coolscan V when scanning B&W was the software! It was actually a problem with getting the negatives to feed right. That feed system is a clever idea, and worked perfect for every color film I tried, and several B&W films. But it would not feed correctly with Ilford B&W for anything. It would constantly be half into a frame when scanning, and I would have to manually correct it, only to find it off again. Believe me, I tried several different things such as cutting the negs different on the ends, etc. Nothing worked.<BR><BR>
    Mendel, thank you very much for the tips. For the record, I used to be somewhat of a computer nut, and my computer is not a slouch. I have changed to firewire. BIG IMPROVEMENT. My original reasoning was, "if they dropped firewire from the 5400II, it must not matter - just use USB". Well, I think the firewire helps.<BR>
    For John and others interested, scanning as "color positive" with no auto exposure or image correction in the scan software, I am now getting final scans in the 4.5 minute range with ICE and Grain Disolver on. That's only about 2x the length of my old Coolscan V with ICE; I never timed it using the ICE & GEM option together, but I remember GEM slows it down, so really, I'm down to less than 2X if you compare apples to apples.<BR><BR>
    So far, it seems as if the ICE(4) on the Coolscan V was a bit better at removing the deepest scratches, but the 5400(I) is old tech...I bought it for the diffused light source; I'm sure they must have improved ICE since then. I'm hoping this small caveat will be outweighed by the way it scans B&W w/ the diffuser.<BR><BR>
    About buying VUESCAN: will this improve the quality of my scans, speed up ICE, or is this program just "easier to use"? I can't see any time difference with running Photoshop and importing the scan, or just running the Dimage scan by itself, and then saving as tiff.<BR><BR>
    Lastly, can anyone tell me what the difference is with 16bit linear? Is this better than 16bit regular?<BR>
  10. jtk


    Jed, Vuescan's "light infared" and Nikonscan's lowest Ice both seem literally perfect on V (except of course for silver and perhaps half of my Kodachrome). This is apparently due to Nikon's forth light source, dedicated to Ice/infared.

    Your difficulty with Nikon's automated transport and one particular Ilford film stock may have been a matter of learning curve (in B&W I've used it with Neopan, XP2, and Delta 400). It might have trouble with dark film base, as well...as with fog between framelines.

    Using Nikonscan, SA21 reads framelines, advances and positions without film handling...whereas Vuescan measures distances, doesn't read framelines, and therefore requires adjustment once or twice on every other film strip in my experience: initially annoying, but easy with practice, faster and cleaner than handling film (end frame flatness, especially with C41, suggests FH-3).

    Mendel and I are fanatics on film flatness...he triumphed with 5400 by mounting negatives in slide mounts. By contrast, Nikon's SA21 holds film very flat, but it's not good with ends of some strips... rather than building my own, I finally got Nikon's FH-3 filmstrip holder (fussy, not as innovative as Mendel's solution, but it does what I want).

    V accepts Scanhancer (origin of Minolta's diffuser concept), which uses Nikon V's FH-3 or the slide holder but doesn't work with the SA21 carrier (diffusion confuses the auto-positioning). Erik might sell a bunch of Scanhancers to people with Nikon V if he really wanted to be in that business. I've not tested Scanhancer thoroughly as I'm lazy and find Vuescan's non-softening "light grain reduction" is plenty, at least up to 12X18.
  11. Jedidiah, I found the firewire connection to be slightly *slower* than USB2.0, *but* more-or-less impervious to slowdown if you're multitasking. I invariably have PS running as I scan, verifying focus of each scan is acceptable, and/or cleaning scans.

    My (uneducated) take on 16 bit linear: it is info straight from the scanner. Gamma is left at 1.0 (not adjusted upward, say to 2.2), the image is not inverted (with neg films), and the histogram has not been stretched in any way.

    This sort of file is not ready for viewing, is way too dark, uninverted (if need be), but can be used as a Vuescan raw file (my usual workflow). Alternately, it can be worked in Photoshop, by applying custom profiles and levels/curves.

    As to Vuescan, I'd say don't even bother thinking about it, plunk down your $80us for a pro license. It's an evolving concept, you will have signed up for a lifetime of upgrades. The author, Ed Hamrick, answers emails, and will entertain questions and wishlist items. Just try to keep it to the point and think it through first, the guy is as busy as a cat in a room full of rocking chairs, I think.

    My use of Vuescan now is purely for scan-from-disk, which consists of doing pseudo scanning with the software, utilizing previously output Vuescan Raw File. This is a gamma 1.0 tiff. More typically this raw file would created by Vuescan itself. But in my case, I use the Minolta Scan Utility 16 bit linear output. I prefer the quality of MSU's ICE cleaning (coupled with the Grain Dissolver) to Vuescan's cleaning (either with or without the GD). Also, Vuescan's raw file, by default, retains the infrared data as a separate 16 bit channel. This is an interesting concept, with great potential, but also a pain, and increases file size by 1/3.

    So, FWIW, my slide workflow is to:

    1. Scan thru MSU, with ICE and GD (GD not optional once ICE is on), Auto-Expose for slides off, exposure sliders (typically) all zeroed, manually focus just prior to scan, and output 16 bit linear.

    2. Open the file in Photoshop and clean remmants missed by ICE. Note, setting your proof viewing mode (with no files open) to Minolta's Scan Elite 5400 Posilinear profile (or other custom profile) will allow you to "turn on the house lights" on these files with <ctrl> "y". They are really dark, and this is needed, for the darker zones of the images. Tho, it's best to turn this off for the brighter regions: then you *really* see the defects.

    3. Utilize above file for Vuescan Scan-From-Disk, outputting jpegs.
  12. jtk


    "...as busy as a cat in a room full of rocking chairs..."

    :) :) :)

    Happy New Year Mendel!

    Jed, listen to Mendel.
  13. Hi John, all the best for the approaching new year! Hey, why do you keep saying I'm remounting neg strips in slide mounts? I never do that. Just previously mounted slides.
  14. Jed, regarding running MSU stand-alone or as a Photoshop plugin, I always go with the former. MSU has a rudimentary auto-naming function that increments, and will same your files on the fly. It is less of a tax on system resources than running MSU as a PS plugin. And with the plugin approach, I believe it just dumps unnamed files into PS workspace, it's kind of pointless, and prone to loss if something locks.
  15. jtk


    Mendel, I misunderstood you all the way through 2005! Quite an accomplishment, considering your clear writing. Best wishes!
  16. Oliver said "slow scans with ICE/GD/Negatives are not a CPU issue", and was kind enough to support this statement with a reference to a statement he made in a previous post, also unsupported.

    I have an LS5000. With ICE on a 0.5 GHz computer with 2G memory it takes a little over 6 minutes to scan a frame at 4000 dpi, 14 bit depth. The same scan on a 3.6 GHz, dual-core workstation with 2G of DDR2 RAM takes 2.5 minutes and about 4 minutes on a 1.5 GHz processor. All three computers run Windows XP Pro. A 5400 ppi file is nearly twice the size of a 4000 ppi file. This would take twice the time to transfer and twice the processing time.
  17. Correction, I have a Nikon LS4000 scanner.
  18. Mendel and John, thanks again for the replies. I will try these new tips and see what happends. I am certainly more happy with the current scan times than I was last night!<BR><BR>

    Good point, Edward (or do you like "Ed"?). So in reality, the Minolta is not really slow, as the file size is twice as big.<BR>
    Just an observation about windows itself - I would imagine some of the CPU power is helping to speed up the scan time simply because Windows & whatever scanning/image program you have running are using a certain base amount of processing power...the faster the CPU, Ram, BUS, and GPU, the more "power" is left over for crunching out the scan. Of course, at some point in all this, the weakest link is the USB or firewire connection from the scanner...<BR>
  19. I think the speed "hits" with the Elite 5400 are due to both ICE and Grain Dissolver. Grain Dissolver alone will (roughly) double scan time, due to it's reduction of light source brightness. And with the extra computation overhead of ICE, the scan time (roughly) doubles again.

    I've found, as is the case with other jobs in my life, dealing with this time factor is just a matter of finding something else to do to fill the time. There are enough post-scan processes I want to apply to my files (final clean-up, for example), that I can't keep up with the scanner, pokey tho it might be.
  20. og


    Edward, the supporting tests to show the influence of CPU on the Minolta 5400 are described in the two last messages of this 2 years old thread. Here is a quote:
    It is very difficult to estimate how much [CPU Power is required], because the Process reports "100%" automatically (my method : I had 2 programs running at high priority and taking about 55% of CPU. I started the Scan at 5400dpi 16bits with GD&ICE, and the process got 60% of CPU [of my Athlon 2500Mhz] with the same scanning time) => 5400dpi at 16 bits with GD&ICE needs 1500Mhz for fastest results - 1000Mhz should be fine for Negatives.
    The Nikon scanners and Minolta 5400 v1 are really different kind of performers and you shouldn't apply a general recommandation/experience based on one, to the other...
  21. My expeirence is even more painful

    With a 2.4C CPU, 2Gb RAM, USB 2.0, it takes hours to scan a strip of 6 negative images at full resolution. However, I still find the focusing be an important issue. Minolta did replace the whole motherboard for the scanner but it is still a problem.
    Manual focus is not easy. Which image should I look for reference if I am using the focusing knob?

Share This Page