d2x vs. f6 with velvia: anybody's got a link?

Discussion in '35mm Cameras' started by Gianni Rondinini, Sep 2, 2005.

  1. yesterday evening i was at my photo club with some friends, to choose
    some slides for a public projection.
    one of the friends told me he recently read an article in which the
    magazine said that they tried to print a file taken with a d2x and a
    couple of scanned velvias --one 50 and one 100f--, which were scanned
    both with a nikon coolscan 5000 *and* a drum scanner. the velvias were
    made to the same subject just after the digital shot, with an
    equivalent lens. the print were made on the same professional printer.
    the result was that the scanned velvias gave a worse
    resolution/quality than the d2x shot.

    now, i can believe or not to this test, but would like to read
    something more --i can remember of the 1ds-mkii that got same
    resolution as a provia 100f on the 'net, but 1ds has a full frame
    sensor and 4+ mpx more than d2x--.
    do any of you have a link to the article?

    please, don't call me troll since i don't want to open a flame: i'm
    just trying to gather some more material for my curiosity. i'm
    shooting velvias and provias with my --bought used-- f5 and will go on
    with this for at least another 200 rolls --which i have in my
    fridge--. since i do own a d100, i also know there are no ways i can
    beat my velvias with my digital ;)
    it's just to read and know...

    thanks in advance.

    Gianni Rondinini, Sep 2, 2005
    1. Advertisements

  2. IMHO you can't compare film and digital in general. For digital, the
    resolution is a hard limit. For film, small details are rendered with
    reduced contrast, but if there is a lot of contrast to begin with,
    small details may remain visible in a print. With digital cameras, there is
    always the risk that small high contrast details result in aliasing errors.

    However, for most subjects a good digital camera with about 8 Mpixels or
    more is enough to create the same impression of sharpness as 35mm color film.

    The following is a q&d scan of a Provia 100F frame on a LS-4000 with maybe a
    bit too much USM (warning 7MB file)
    Philip Homburg, Sep 2, 2005
    1. Advertisements

  3. Gianni Rondinini

    Gordon Moat Guest

    Okay, a few things to consider. The top optical resolution from D2X tests
    seems to be near 60 lp/mm maximum. Velvia 50 is not a film one would
    often choose for high resolution, but it is popular. The newer Velvia
    100F should be a little better than the old Velvia, but the colour
    balance is different. Choosing either Velvia is a bias towards colour,
    not resolution. I have seen recent tests of Astia 100F that showed 80 to
    100 lp/mm capability.

    Next item is the scanners. Sorry to bash a scanner, but the Coolscan 5000
    is just okay. Drum scanner can mean nearly anything, from very old Howtek
    or Leaf, to a brand new ICG, or a five year old Heidelberg Tango.
    Regardless of which scanner, it will only be as good as the operator, and
    the software. In other words, getting that 80 to 100 lp/mm off the film
    can be tough. It would be reasonable to assume that a CCD film scan, or a
    poorly done drum scan, might be near or less than 60 lp/mm (D2X best).

    Use either camera hand held, and the resolution will drop. The D2X has in
    camera sharpening, so it might appear to be a better, or more contrasty,
    or just sharper final image. Whether the CCD film scan, or the drum scan,
    were sharpened, and if that was done competently is another variable.
    There are more things to go wrong with film than there are from just
    downloading an image from a D2X.

    Last item is the printer. It can very well be that the D2X might be a
    better match for the printer. It could also be that the film scans were
    not well matched to the printer. Again, likely easier and less steps to
    get a good print from a D2X.
    I don't have a link to that particular article. The only recent film and
    direct digital comparison articles with prints I have seen were with the
    Leica R9 and the same camera with the Digital Module R. These have
    appeared in the last two issues of LFI (Leica Fotographie International).
    I recommend getting a copy of one of these, rather than trying to view on
    the internet. Forget about judging printed image quality from internet
    articles, you will not get a real idea of capabilities.
    Gianni, I would not accuse you of trolling, since you have posted here
    quite often. If you are happy with Velvia, then my suggestion is to leave
    well enough alone and continue using it. If you do all your photos off a
    tripod when using mirror lock-up, then maybe something higher resolution
    would be something to investigate. If you take photos mostly hand held,
    then forget about improving resolution much at all, regardless of whether
    you use an F6 or a D2X.

    Now for some technical issues. The D2X is realistically about half frame
    35 mm. What this means is that the edges of the image circle from a lens
    will be cropped. Since almost all lenses have less resolution at the
    edges, this means a D2X crops out the lower resolution areas, which also
    means that the images can appear quite even in resolution and exposure
    from centre to edge of the frame. Since it is half frame 35 mm, then for
    the same size print twice the enlargement would be needed compare to full
    frame 35 mm, though the edges should hold up better on the D2X in some
    situations. Compare D2X images to Canon 1Ds MK II full frame images, and
    sometimes the D2X images seems a bit better in resolution.

    Even a well done film scan will struggle to get more than 90% of the
    available information off the film. If it was 80 lp/mm, then your best
    case scan might only show near 70 lp/mm. A poorly done film scan might
    only get half that information, or maybe 40 lp/mm. Okay, so maybe you
    think that since full frame does not need as much enlargement as half
    frame, then you can get away with the lower overall resolution. Recall
    that the edges might not be as good, so maybe too much of the film
    resolution was lost in the scans, more than could make up any

    Last item is the printer resolution, which is actually the most
    important. Using a somewhat common 300 ppi file for printed output (304.8
    ppi for metric printers), we find that only implies near 6 lp/mm at the
    final print size. More on this stuff in a couple articles at:


    Anyway, if you took the D2X best of near 60 lp/mm, then figured the
    printed file 6 lp/mm, then about only a 10x enlargement should show no
    lack of resolution. In reality, many people can rarely differentiate that
    small detail information, so maybe even a 20x enlargement might not show
    a lack of resolution. Dot gain is another factor, in that some small
    details just will not appear until some very large enlargements are
    printed. The real factor that many judge as resolution can actually be
    accutance or sharpness, and in those judgements the direct digital
    methods can prove a bit better.

    I almost have to laugh about this obsession some people have with high
    resolution. I recently saw the printed exhibit images for the GigaPXL
    Project, which is a system claiming 1000 MP resolution. Indeed the
    resolution was so high, the front desk at MOPA offered magnifying glasses
    to anyone wanting to see more detail in the images. These are the sorts
    of images that some on this group would likely wet themselves in
    excitement at seeing; I recall comments of "details you can walk right up
    to in really large prints". In all honesty, I found these images to be
    the most boring I have ever seen in any exhibit anywhere. The crowd
    seemed equally unimpressed, and it failed to hold anyone's attention for
    long. People who like this sort of thing must be the most boring idiots
    on the face of the planet . . . I would dismiss any comments such people
    might have about resolution, image quality, or photography.

    Contrast that GigaPXL Project images to the other hall at MOPA, which had
    a Steve McCurry exhibit. Even the 30" by 40" prints in that exhibit held
    your attention. This was not because of higher resolution, but because
    they were more compelling, and more interesting images. The colours were
    also better, though I am not sure what films Mr. McCurry used. There were
    maybe two of the Steve McCurry images I wished had more detail or
    resolution, though in those you could tell that a too short DOF was the
    cause of soft areas in those images, not a lack of resolution due to his
    choice to use film.
    Gordon Moat, Sep 2, 2005
  4. thanks to both of you for your replies. i got also a couple of replies
    via email: i wanted to thank you too.

    i didn't know of these lpmm tests. thanks.
    they didn't say anything more than "we scanned the slides with the
    drum scanner used to prepare images for publishing them in our
    magazine." and it's not very detailing about the scanner used...

    something disappointing is that for "professional results" they make
    compasisons using a consumer scanner. i planned to buy a coolscan
    5000, but i know it's nothing of professional: it's just a good
    consumer product.
    ok. we can consider that with a great slide --say, a provia 100f--,
    you could reach about the same resolution.
    is there a website with the lpmm one can reach with the different
    slides? i mean, for example you said that velvia 50 --what about
    100f?-- wouldn't be often used for very high resolution images, but
    what about other slides? where can i read more about this? i don't
    know the exact terms used for these parameters and wouldn't know what
    to google for.
    yes, being sincere i'm happy of velvias. i started shooting also some
    provias: the colours aren't the same, but every subject has it's best
    film. for portraits, i wouldn't recommend any of those --astia is for
    sure the best--.
    i love projecting my images then with no doubt i'd say that slides are
    the best support for my photos.

    in the last days, i've been working on something i need to print, then
    i used my d100. since i'm interested only in printing a couple of
    images out of some hundreds, digital gives it's best in these things.

    but the thing that "digital beats slides everywhere everytime" is
    getting more and more frequent and begins to be annoying and since i'm
    not a big marketing-lover, i wanted to have some numbers by hand to
    talk about when we are at our club tuesday evening.
    very well.
    indeed, i use the tripod quite seldom. i mean, when it's really
    necessary --macro, very slow times, medium format, so on...--. but i
    shot most of my photos without a tripod.

    i understand.

    well, being sincere and not being a very-ultra-super-wide angle
    lenses, the fact that nikon is going on with their cropped sensor is
    useful for me. i mean: a 300/2.8 is far more expensive than a 14/2.8,
    everything's clear.

    Gianni Rondinini, Sep 6, 2005
  5. I don't know why Gordon Moat calls an LS-5000 a consumer scanner (you have to
    ask him), but if you stay within the limits of the scanner, you should
    be able to get professional results.

    An LS-5000 is limited to 4000 ppi. Depending on your application, that
    may or may not be a problem. If you print those 4000 ppi at a 300 ppi
    output resolution (ie at most 12x18") then the scanner should not be
    a limiting factor.

    If you want to get the most out of a frame, the Minolta 5400 series can
    extract a bit more detail and some drum scanners go even further.
    However, it is unlikely that those extra details result in a
    significantly better print (wrt resolution). The appearance of grain does
    benefit from higher resolutions (or drum scans in general).

    Another issue is that drum scan are better for scanning very dense parts
    of slides. Whether that is useful depends on how much shadow detail needs
    to be made visible.
    is a good starting point for scanners

    The following is table with film characteristics:
    Philip Homburg, Sep 6, 2005
  6. Gianni Rondinini

    Gordon Moat Guest

    You could start with Fujifilm website for Fuji Data Sheets. These will show
    various test resolutions of their films. However, I have never seen anyone
    outside of Zeiss indicate numbers close to data sheet limits. In more real
    achievable controlled conditions, I would expect a best of 75% to maybe 80%
    of the maximum on the data sheet; mostly for transparency films from either
    Fuji or Kodak. With B/W films, it might be possible to do much better than
    100 lp/mm with some films. You still have a problem of scanning that, though
    the higher you start the more you retain.
    I never liked any Provia choices from Fuji, just not a good film choice for
    me. Quite happy with the new Astia 100F. I use more Kodak E films than
    anything from Fuji.
    Feel free to print out and use my articles, if you want. You might also want
    to check out Zeiss and Erwin Puts site <http://www.imx.nl>. Erwin has some
    copies of the articles he wrote for LFI about a year ago, mostly Astia 100F
    and Kodak E100G and GX.

    What I never like about those arguments is that the limits of technology are
    rarely approached by most photographers. We like smaller cameras because we
    can use them hand held. We also tend to like zoom lenses (except me), due to
    convenience. The reality is that any camera used hand held with a zoom lens
    will not get near the possible resolution limits.

    The other side of the issue is colour, though that is very subjective. Most
    people feel that whatever colour they can see on a computer monitor is the
    limit to the range of possible colours. People who view transparencies on a
    light table often have another viewpoint. People involved in commercial
    printing bring a third (often conflicting) view. Simply put, transparency
    films can often capture colours not possible with Bayer pattern digital
    cameras, and some printing colours will not appear on any computer monitor
    they way they should actually look.
    My guess is that describes how many people take photos. Another resolution
    helper is using strobes or flash. Of course not many do that most of the time
    either, since strobes and flash units add bulk and weight.
    The widest I use is a 20 mm, and though I have used a 300 mm in the past the
    longest lens I currently use in 35 mm is a 180 mm. A 300 mm f2.8 is a
    somewhat special lens. Another one is a 200 mm f2.0. Both those are not very
    common, and never sold in big numbers, so price comparisons are tougher.
    Compare a 20 mm f2.8 to a 180 mm f2.8, and used prices are close, though the
    bigger lenses often cost a bit more.
    Glad to be of help. Try not to get too hung up on the numbers. A good image
    is much more than the technical quality.
    Gordon Moat, Sep 6, 2005
  7. Gianni Rondinini

    Gordon Moat Guest

    Simply put, that scanner is not good in a production environment. Higher volume
    of scans, as might be often needed in a professional environment, can reveal
    that a scanner is too slow, or needs service too often. High end professional
    gear can stand up to a higher volume of scans, and often when service would be
    needed has some user serviceable components, or a quick service from the
    manufacturer (three weeks repair service is not quick and would not help a

    Of course, this is my opinion on this, and not matched by all. I don't consider
    the resolution as the guide of whether or not something is professional quality,
    since 4000 ppi certainly is usable in many published outputs. Indeed, even some
    professionals use the LS-5000 for their work. Other professionals might only
    consider using that if they had two of them on site, since if one breaks or
    needs service they are left with another one working. Take the cost of two
    LS-5000 scanners into consideration, and you can see that an Imacon or low end
    Creo seem to make more sense. Other good production or professional level
    scanners are the Fuji Lanovia Quattro and Dainippon Screen Cezanne.
    Put in a more technical manner, a drum scanner can often have a greater possible
    dynamic range and/or a higher Dmax limit than a film scanner. However, there are
    some CCD film scanners that do well with dynamic range and Dmax.
    Philip, I misplaced the Zeiss link for their film tests. If you have that handy,
    I think Gianni might find those interesting too. Thanks.
    Gordon Moat, Sep 6, 2005
  8. Well, I have to admit that my LS-4000 is rather slow, but whether it is
    too slow depends on the amount of post processing. I'm quite sure that for
    negatives I cannot keep up with the scanner. For properly exposed slides
    it is possible that the scanner is a limiting factor.

    Either very few people are trying to use Nikon scanners 24x7 or they are
    not complaining about it on the Internet (or I never came across their

    The only bit of maintenance I am aware of are dirty mirrors. But that is
    something anybody can do.

    What kind of user serviceable components are you talking about? Does some
    mechanical part wear out too soon?
    Do you have scan times for those scanners?
    Okay, here it is:
    It is page 6 of 'A Virtual Tour of the Lens Production'.
    www.zeiss.com > German Headquarter Divisions > Camera/Cine Lenses >
    Production Tour > production of optical components > Download the Virtual
    Optical Production
    Philip Homburg, Sep 6, 2005
  9. Gianni Rondinini

    Gordon Moat Guest

    Probably just not writing about it much.
    Mirror maintenance (cleaning) could do more harm than good if someone is not very
    careful. It is something best left to an expert, or at least some company who will
    replace it if they make a mistake.
    Mostly the lamps, or with Nikon scanners the LEDs. A few large production scanners
    offer serviceable, or user replaceable lamps. The main issue is not burning out, but
    changing in colour temperature. Mechanically there are stepper motors that go out of
    alignment. Mind you this would be a high volume environment, not occasional use.
    Based on an industry standard of scanning a 6x7 transparency at 300 ppi and 300%
    Creo iQSmart1 - 50 scans per hour
    Dainippon Screen Cezanne - 54 scans per hour
    Fuji FinePix 2750 - 75 scans per hour, with some higher end models like the Lanovia
    doing better.
    Great and thanks. Will have to bookmark that.
    Gordon Moat, Sep 7, 2005
    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.