11MP digital or medium format film?

Discussion in 'Digital Photography' started by Beowulf, Aug 22, 2004.

  1. Beowulf

    Beowulf Guest

    Do the 11 megapixel digital cameras approach the resolution of a medium
    format film camera? I have a Canon 10D digital, 6.3 megapixels, and a
    Canon Rebel Ti film camera. I would like the greater resolution of a
    medium format camera for larger finer grained prints-- but perhaps digital
    is approaching that with 11+ megapixels?
    ~Beowulf

    --
    "It said it needed Windows98 or better installed, so I installed Linux."
     
    Beowulf, Aug 22, 2004
    #1
    1. Advertising

  2. "Beowulf" <> wrote in message
    news:p...
    > Do the 11 megapixel digital cameras approach the resolution of a medium
    > format film camera? I have a Canon 10D digital, 6.3 megapixels, and a
    > Canon Rebel Ti film camera. I would like the greater resolution of a
    > medium format camera for larger finer grained prints-- but perhaps digital
    > is approaching that with 11+ megapixels?


    11 MP? True resolution? Hell, 35mm blows it out of the water. Effective
    resolution? Probably. How are you scanning your 35mm film? I use a
    Minolta 5400 (41 MP) and a Pacific Image PowerSlide 3600 (17 MP). Either
    blows 11 MP out of the water... but are you wanting to totaly get rid of
    grain? Then you go with a Digital. That is not however actual resolution.
     
    Robert Meyers, Aug 22, 2004
    #2
    1. Advertising

  3. Beowulf

    Annika1980 Guest

    >From: "Robert Meyers"

    >11 MP? True resolution? Hell, 35mm blows it out of the water. Effective
    >resolution? Probably. How are you scanning your 35mm film? I use a
    >Minolta 5400 (41 MP) and a Pacific Image PowerSlide 3600 (17 MP). Either
    >blows 11 MP out of the water... but are you wanting to totaly get rid of
    >grain? Then you go with a Digital. That is not however actual resolution.


    I also have the Minolta 5400 and I disagree with your analysis. I don't have a
    1Ds to compare against, but there isn't that much (if any) more detail in the
    huge scans from the 5400 that there is from the 6.3MP Totally Digital D60.

    Yes, it creates bigger image sizes, but as you hinted at, the grain is enlarged
    as well.
    If I take similar shots with each it is hard to spot any ("real", not grain)
    detail in the film scan that wasn't in the digital shot.
    I would suspect that an 11MP 1Ds shot would drown it.
     
    Annika1980, Aug 22, 2004
    #3
  4. "Robert Meyers" <> wrote in message
    news:...
    > "Beowulf" <> wrote in message
    > news:p...
    > > Do the 11 megapixel digital cameras approach the resolution
    > > of a medium format film camera?

    SNIP
    > 11 MP? True resolution? Hell, 35mm blows it out of the water.

    Effective
    > resolution? Probably.


    I agree, that is correct for lower ISO film and quality lenses used to
    make the images. A tripod or flash will certainly help as well.
    The scanned resolution of 35mm film seems limited to approx. 80-85
    cycles/mm, where high end (near) full frame '35mm DSLR' sensors max
    out at roughly 50-60 cy/mm, so in absolute resolution they're no
    match.

    The lack of graininess in the DSLR image allows a fair amount of
    postprocessing that will visually bridge the gap a bit. On the other
    hand, noise reduction programs like Neat Image can be successfully
    applied to film scans, and lose very little resolution, so the final
    results can be very close.

    Medium format images in general have about twice the linear size of
    35mm film, so need only half the magnification, which more than
    offsets the slight resolution loss due to non-flattness of the film
    and often lower resolution (because they are larger) than top 35mm
    lenses. Film graininess is also less of an issue because of the lower
    magnification. So MF is potentially about twice (rough estimate) as
    high in resolution for the same output size as a DSLR sensor image.

    Physically smaller sensors need more magnification for equal sized
    output, so they lose more resolution yet.

    Bart
     
    Bart van der Wolf, Aug 22, 2004
    #4
  5. Beowulf

    Alfred Molon Guest

    Robert Meyers <> wrote:

    >11 MP? True resolution? Hell, 35mm blows it out of the water. Effective
    >resolution? Probably. How are you scanning your 35mm film? I use a
    >Minolta 5400 (41 MP) and a Pacific Image PowerSlide 3600 (17 MP). Either
    >blows 11 MP out of the water... but are you wanting to totaly get rid of
    >grain? Then you go with a Digital. That is not however actual resolution.


    Well... see here for what you get from a Minolta DSLR with a Tamron 28-
    200 lens:

    http://www.ddde.de/F21_35.jpg

    This was scanned with a Nikon LS50 slide scanner at 4000 dpi. We are
    (resolutionwise) at the level of 1-2 MPixel. The Tamron 28-200 probably
    isn't the best lens around, and I guess that with much better lenses the
    resolution you get improves. Perhaps you can make it to 6MP.

    In any case I'd also be curious to know what the effective resolution of
    a MF frame is.

    The area of a MF frame is about four times the area of a 35mm frame,
    isn't it ? Then, taking into account that MF lenses are less sharp than
    35mm lenses (because they don't need to be so sharp), could we say that
    with a MF frame you get three times the resolution of a 35mm from, i.e.
    3 x 6 = 18MP ?
    --

    Alfred Molon
    ------------------------------
    http://groups.yahoo.com/group/Olympus_405080/
    Olympus 5060 resource - http://www.molon.de/5060.html
    Olympus 8080 resource - http://www.molon.de/8080.html
     
    Alfred Molon, Aug 22, 2004
    #5
  6. "Alfred Molon" <> wrote in message
    news:...
    SNIP
    > Well... see here for what you get from a Minolta DSLR with a Tamron

    28-
    > 200 lens:
    >
    > http://www.ddde.de/F21_35.jpg


    That image is hardly representative of what film has to offer.

    The image suffers from camera shake. The zoom lens may have added it's
    own limitations. Had the image been good, then 4000 ppi isn't enough
    to resolve all detail.

    A small (6.35x6.35mm, or 4.7% of the ful frame area) crop from a 5400
    ppi scan (un-retouched, no grain reduction, no sharpening) shows a
    much more detailed image, and that one was handheld:
    <http://www.terrapinphoto.com/jmdavis/SE5400-Wbd-Crop.jpg> for the
    crop, and
    <http://www.terrapinphoto.com/jmdavis/SE5400-Wbd-Overview.jpg> for the
    overview.

    There are other examples of scans on that site
    (http://www.terrapinphoto.com/jmdavis) but many suffer from "operator
    limitations" that would also impact a Digicam image.

    Bart
     
    Bart van der Wolf, Aug 22, 2004
    #6
  7. Beowulf

    Allan Wind Guest

    On 2004-08-22, Alfred Molon <> wrote:
    > Well... see here for what you get from a Minolta DSLR with a Tamron 28-
    > 200 lens:
    >
    > http://www.ddde.de/F21_35.jpg


    Do you have a larger resolution of this scan online, perhaps the
    unprocessed version? I was expecting better results with a slide
    scanner.


    /Allan
     
    Allan Wind, Aug 22, 2004
    #7
  8. Beowulf

    Mark M Guest

    "Alfred Molon" <> wrote in message
    news:...
    > Robert Meyers <> wrote:
    >
    > >11 MP? True resolution? Hell, 35mm blows it out of the water. Effective
    > >resolution? Probably. How are you scanning your 35mm film? I use a
    > >Minolta 5400 (41 MP) and a Pacific Image PowerSlide 3600 (17 MP). Either
    > >blows 11 MP out of the water... but are you wanting to totaly get rid of
    > >grain? Then you go with a Digital. That is not however actual

    resolution.
    >
    > Well... see here for what you get from a Minolta DSLR with a Tamron 28-
    > 200 lens:
    >
    > http://www.ddde.de/F21_35.jpg


    That image is a blurry mess, which is a shame considering the nice scene.
    Film can do far better than that example.
     
    Mark M, Aug 22, 2004
    #8
  9. "Bart van der Wolf" <> wrote:
    > "Alfred Molon" <> wrote:


    > > http://www.ddde.de/F21_35.jpg

    >
    > That image is hardly representative of what film has to offer.


    Really.

    > much more detailed image, and that one was handheld:
    > <http://www.terrapinphoto.com/jmdavis/SE5400-Wbd-Crop.jpg> for the
    > crop, and
    > <http://www.terrapinphoto.com/jmdavis/SE5400-Wbd-Overview.jpg> for the
    > overview.


    But even that is pretty pitiful on a pixel-per-pixel basis comparison with
    images straight out of a dSLR. I'm not convinced you've got even than 2700
    dpi of usable* resolultion there. (My usual rant: the information in the low
    contrast tail of the MTFs curve does't significantly contribute to apparent
    image quality.)

    *: Usable means when printed at 300 dpi.

    > There are other examples of scans on that site
    > (http://www.terrapinphoto.com/jmdavis) but many suffer from "operator
    > limitations" that would also impact a Digicam image.


    They all look pretty ugly to me. I really don't see anyone getting over 9MP
    of dSLR quality pixels from a 35mm frame. (Where "quality" basically means
    the width of sharp transitions. Grain noise even after downsampling 5400 dpi
    scans to 2700 dpi is pretty bad.)

    David J. Littleboy
    Tokyo, Japan
     
    David J. Littleboy, Aug 22, 2004
    #9
  10. Beowulf <> wrote in
    news:p:

    > Do the 11 megapixel digital cameras approach the resolution of a
    > medium format film camera? I have a Canon 10D digital, 6.3 megapixels,
    > and a Canon Rebel Ti film camera. I would like the greater resolution
    > of a medium format camera for larger finer grained prints-- but
    > perhaps digital is approaching that with 11+ megapixels?


    You will get many kind of answers to this question. The main
    reasons for the diverse opinions are (1) it is not really
    possble to compare film and digital on equal ground and (2)
    there are lots of myths.

    Problem one: You can (with very good film and very good scanning)
    force very high resolution out of film.

    Problem two: What you can get under the absolute best circumstances
    (i.e. problem one above) is not typical for film.

    Problem three: Film has much larger granularity than it has
    resolution. For digital (on the other hand), resolution and
    granularity is almost the same.

    Problem four: Digital has a regular square grid and film results
    in a random pattern.

    I think it is fair to say that even with a 6 Mpixel DSLR you get
    the same kind of smoothness as you get with a medium format film
    camera. So - if smoothness is your aim - you are already there.

    I also think it is fair to say that a reasonable film and
    scanning technique do not give you much more than 6 Mpixels
    of resolution from 35 mm film.

    Hmmm .... 11 Mpixels and medium format and resolution. I don't know.
    Four times 6 Mpixels is 24 Mpixels. So --- more than that it is not.
    I assume that 15 or so would be enough for most cases if you want to
    get medium format quality. But that is a guess.

    A last remark is that in some cases film grain is beautiful/useful.
    This is true at least for B&W prints. In this case, the smoothness
    of the digital camera is not an advantage.


    /Roland
     
    Roland Karlsson, Aug 22, 2004
    #10
  11. Beowulf

    Bill Hilton Guest

    > Do the 11 megapixel digital cameras approach the resolution of a
    > medium format film camera?


    We have a six Mpix body (10D), two 8 Mpix bodies (1D Mark II's) and an 11 Mpix
    body (1Ds). Also shoot 645 and 6x7 cm film (Velvia mostly, some Provia 100F),
    which we scan with a Nikon 8000 (occasionally with a drum scanner).

    To my eye the 10D prints at 12x18" aren't as good as 35 mm prints the same
    size, Mark II prints are getting very close to 35 mm print quality (call it
    pretty much a draw) and the 1Ds prints are clearly smoother and nicer looking
    than 35 mm prints.

    You asked about "resolution" ... if you shoot test targets you'll probably find
    that the 1Ds doesn't even match the full resolving power of 35 mm film, yet
    because the files are so smooth and noise-free they *look* better when printed,
    especially with larger prints.

    As for medium format, I've upsampled the 1Ds files to 54 Mpixels and printed
    20x30" @ 300 ppi and the prints are incredibly smooth, much better than I
    thought I'd get. In some ways they look better than 645 prints at that size.
    I've seen 1Ds prints as large as 40x60 that looked good. A lot depends on how
    you upsample and how well you can sharpen (use edge sharpening if you know
    how), but done right you can really blow it up.

    I haven't shot the exact same scene with the 645 or 6x7 and the 1Ds (we've been
    doing mostly wildlife with the 1Ds and Mark II since we got them, not
    landscapes) but since I don't like to print the 645 larger than 20x24" (which
    is a stretch ... 16x20" looks excellent to me though) I'd expect the 1Ds prints
    to start approaching MF quality, or at least 645 MF with the gear I have
    (Pentax). I'd be really surprised if they match 6x7cm prints though, but my
    6x7 lenses are exceptionally good (Mamiya 7 II). Again, this is mainly due to
    the smoothness of the images, the resolution isn't as high but the net effect
    is the prints look beautiful since there is no distracting "grain".

    In September I plan on taking all three systems (1Ds, 6x7 and 645) to shoot
    aspens in Colorado and will try to shoot the same scenes with each body as
    often as possible (the 6x7 lens selection is limited to short-tele's and wider)
    so I can do some more exact comparisons.

    Bill
     
    Bill Hilton, Aug 22, 2004
    #11
  12. In article <>, Bill Hilton
    <> writes
    >which we scan with a Nikon 8000 (occasionally with a drum scanner).


    When do you feel necessary/appropriate/desirable to use the drum
    scanner?

    > A lot depends on how
    >you upsample and how well you can sharpen (use edge sharpening if you know
    >how), but done right you can really blow it up.


    What do you mean by "how you upsample? Can one upsample in more ways
    than one? What do you mean by "edge sharpening"? Do you mean
    filter/sharpen/unsharp mask in Photoshop, but only after "find edges"
    etc., or other selective techniques for USM? Or do you use
    filter/sharpen/sharpen edges in PS?

    >In September I plan on taking all three systems (1Ds, 6x7 and 645) to shoot
    >aspens in Colorado and will try to shoot the same scenes with each body as
    >often as possible (the 6x7 lens selection is limited to short-tele's and wider)
    >so I can do some more exact comparisons.


    Bill, you left out large format. LF scanned with, perhaps, a drum
    scanner, would this remain the last bastion of film, in your view? What
    do your instincts/experience tell you?
    --

    nobody
     
    nobody nowhere, Aug 22, 2004
    #12
  13. "David J. Littleboy" <> wrote in message
    news:cgavla$jm7$...
    >
    > "Bart van der Wolf" <> wrote:

    SNIP
    > > <http://www.terrapinphoto.com/jmdavis/SE5400-Wbd-Crop.jpg> for the
    > > crop, and
    > > <http://www.terrapinphoto.com/jmdavis/SE5400-Wbd-Overview.jpg> for

    the
    > > overview.

    >
    > But even that is pretty pitiful on a pixel-per-pixel basis

    comparison
    > with images straight out of a dSLR.


    Pixel-per-pixel yes, as I said, but since there are so many more
    pixels available in a filmscan, they can be scaled smaller than the
    relative few of a digicam. And the scans were not processed for noise
    and sharpness, whereas most dSLR images are (either in camera, or by
    Raw-converter).

    > I'm not convinced you've got even than 2700 dpi of usable*
    > resolultion there. (My usual rant: the information in the low
    > contrast tail of the MTFs curve does't significantly contribute
    > to apparent image quality.)


    That is only a valid argument *before* sharpening. After proper
    sharpening, the high spatial frequency modulation is raised to levels
    that are clearly visible. That requires the prior application of noise
    reduction on film scans, because otherwise that noise would become
    visible in output as well.

    I've done the tests and the effect that sharpening has on the MTF of
    film scans is astonishing. Only a blind bat would not notice the
    effect in output.

    Bart
     
    Bart van der Wolf, Aug 22, 2004
    #13
  14. Beowulf

    Bill Hilton Guest

    >> Bill Hilton <> writes
    >>
    >> In September I plan on taking all three systems (1Ds, 6x7 and 645)
    >> ... so I can do some more exact comparisons.


    >From: nobody nowhere
    >
    >Bill, you left out large format. LF scanned with, perhaps, a drum
    >scanner, would this remain the last bastion of film, in your view? What
    >do your instincts/experience tell you?


    Sure, 4x5 drum scanned prints great, it's just that I don't have a 4x5 system
    :) At one workshop a guy had 8x10" Velvia trannies and even though he only had
    them scanned to a 200 MB file size the prints were incredibly smooooth. Bigger
    is better.

    >> which we scan with a Nikon 8000 (occasionally with a drum scanner).


    >When do you feel necessary/appropriate/desirable to use the drum
    >scanner?


    When we're selling large prints it's nice to have a $50 200 MB drum scan done
    professionally. Or when we're exhibiting something as a large print. The 8000
    does a really good job for the money but the Tango has a bit more resolving
    power and better shadow detail, though Photoshop CS's Shadow/Highlight tool
    helps even this up a bit.

    >> A lot depends on how you upsample (for printing large)


    >What do you mean by "how you upsample? Can one upsample in more ways
    >than one?


    Sure, many ways ... just a simple Image > Image Size (bicubic) or with CS
    'bicubic smoother' for two. Or Genuine Fractals, which some people like. Or
    by upsizing in multiple 110% increments, known as 'stair interpolation'
    (usually this beats Genuine Fractals on normal nature images, I find).
    Sometimes different image structures upsize better with different techniques so
    it's worthwhile to try more than one.

    Here's a run-down on various methods, including a couple I don't have and
    haven't tried: http://www.fredmiranda.com/SI/index.html

    >What do you mean by "edge sharpening"? Do you mean
    >filter/sharpen/unsharp mask in Photoshop, but only after "find edges"
    >etc., or other selective techniques for USM? Or do you use
    >filter/sharpen/sharpen edges in PS?


    Deke McClelland is the first guy I saw mention this, several Photoshop versions
    ago (I think I first saw it in "Photoshop 5 Bible") ... basically you make a
    copy of the image, open a new channel and paste the image in there, run Filter
    > Stylize > Find Edges to outline the edges (which is what you want to

    sharpen), invert it, soften up the mask a bit .. on a typical large file for
    printing I'd run Median with radius 3, Maximum with radius 5, Gaussian blur
    with radius 5, then adjust it a bit in Levels, maybe touch it up with the
    paintbrush by painting white on areas I wanted more sharpening (like eyes) and
    painting black with low opacity on areas that are too white in the mask and
    might oversharpen. Then load this channel as a selection (mask) and sharpen
    through the mask with USM with something like Amt 400%, threshold 0 (since the
    mask is defining the edges you take the threshold out of the equation) and
    watch carefully as the radius is adjusted from between .4 to say 1.2. (Smaller
    numbers on all of these steps for smaller files). I have an action from Bill
    Atkinson that does all this automatically on the L channel in LAB mode and it
    works great.

    If you just use USM it's hard to sharpen the high frequency data (eyes, hair,
    leaves, etc) without adding noise to the low frequency areas (sky, out of focus
    areas, etc), though it's much easier with digital than scanned film because of
    the lower noise. The mask takes care of this for you by masking out the low
    frequency areas.

    Bill
     
    Bill Hilton, Aug 23, 2004
    #14
  15. nobody nowhere wrote:
    > In article <>, Bill Hilton
    > <> writes
    >
    >> which we scan with a Nikon 8000 (occasionally with a drum scanner).

    >
    >
    > When do you feel necessary/appropriate/desirable to use the drum scanner?
    >
    >> A lot depends on how
    >> you upsample and how well you can sharpen (use edge sharpening if you
    >> know
    >> how), but done right you can really blow it up.

    >
    >
    > What do you mean by "how you upsample? Can one upsample in more ways
    > than one? What do you mean by "edge sharpening"? Do you mean
    > filter/sharpen/unsharp mask in Photoshop, but only after "find edges"
    > etc., or other selective techniques for USM? Or do you use
    > filter/sharpen/sharpen edges in PS?
    >
    >> In September I plan on taking all three systems (1Ds, 6x7 and 645) to
    >> shoot
    >> aspens in Colorado and will try to shoot the same scenes with each
    >> body as
    >> often as possible (the 6x7 lens selection is limited to short-tele's
    >> and wider)
    >> so I can do some more exact comparisons.

    >
    >
    > Bill, you left out large format. LF scanned with, perhaps, a drum
    > scanner, would this remain the last bastion of film, in your view? What
    > do your instincts/experience tell you


    Large format (fuji velvia drum scanned) is equivalent
    to about 200 megapixels of digital. See:
    http://www.clarkvision.com/imagedetail/scandetail.html

    I agree with everything Bill said. While in my testing,
    6x4.5 medium format film is equivalent to about 50 megapixels
    and for Velvia ISO 50 film in terms of resolved pixels, and 35mm
    velvia is 16 megapixels, the digital images are smoother (higher
    signal to noise) so have the appearance of looking
    great, with a perception of sharpness. See other articles
    at: http://www.clarkvision.com/imagedetail
    I'm making nice 16x20s from 6 mpixel digital (a lot of
    careful work upsampling, smoothing, sharpening). But
    then applying the same techniques to 35mm velvia film,
    I'm making 25x36 inch prints from 4000 dpi scans of velvia film
    that match digital in appearance, using digital processing
    methods. Large format: I do lots of 30x40 and an occasional
    4x5 foot enlargement, again using the same methods.
    (The large format could go much bigger.) The key to making
    large prints is processing. Just like Ansel Adams spent
    hours in the darkroom making a print, I spend hours at the
    computer enlarging, sharpening, smoothing, adjusting contrast,
    dodge and burn, setting color balance, and repairing defects.

    Roger
    Photos at:
    http://www.clarkvision.com
     
    Roger N. Clark (change username to rnclark), Aug 23, 2004
    #15
  16. "Bart van der Wolf" <> wrote:
    >
    > > I'm not convinced you've got even than 2700 dpi of usable*
    > > resolultion there. (My usual rant: the information in the low
    > > contrast tail of the MTFs curve does't significantly contribute
    > > to apparent image quality.)

    >
    > That is only a valid argument *before* sharpening. After proper
    > sharpening, the high spatial frequency modulation is raised to levels
    > that are clearly visible. That requires the prior application of noise
    > reduction on film scans, because otherwise that noise would become
    > visible in output as well.
    >
    > I've done the tests and the effect that sharpening has on the MTF of
    > film scans is astonishing. Only a blind bat would not notice the
    > effect in output.


    You've said that before. Here's a test. With a tripod and a slow film, shoot
    the same subject from the same place with two different lenses that differ
    in focal lengths by a factor of 1.666. Look at a 900 x 900 pixel crop from
    the shorter focal length image and a 1500 x 1500 pixel crop of the same
    piece of the subject from the longer focal length image. Both should show
    the same thing.

    NeatImage and sharpen both the best you can.

    Print both at 3" x 3" on an Epson R800 with Qimage. (I'd actually downsample
    the larger crop to 900 x 900 and resharpen first.)

    Tell me if you can see the difference. (Or even better, see if your friends
    can see a difference.)

    If you can't, you're right, and if you can, you're exhaggeratting the claims
    for high-res film scan resolution.

    David J. Littleboy
    Tokyo, Japan
     
    David J. Littleboy, Aug 23, 2004
    #16
  17. In article <>, Bill Hilton
    <> writes

    Thank you very much Bill. I have found the technique you describe below
    in a number of articles by Bruce Fraser, dating from 2001, which I
    downloaded

    http://www.creativepro.com/printerfriendly/story/12189.html

    The irony is that it might be you who referred me to this article in the
    first place!!
    Thanks again.
    >
    >
    >Deke McClelland is the first guy I saw mention this, several Photoshop versions
    >ago (I think I first saw it in "Photoshop 5 Bible") ... basically you make a
    >copy of the image, open a new channel and paste the image in there, run Filter
    >> Stylize > Find Edges to outline the edges (which is what you want to

    >sharpen), invert it, soften up the mask a bit .. on a typical large file for
    >printing I'd run Median with radius 3, Maximum with radius 5, Gaussian blur
    >with radius 5, then adjust it a bit in Levels, maybe touch it up with the
    >paintbrush by painting white on areas I wanted more sharpening (like eyes) and
    >painting black with low opacity on areas that are too white in the mask and
    >might oversharpen. Then load this channel as a selection (mask) and sharpen
    >through the mask with USM with something like Amt 400%, threshold 0 (since the
    >mask is defining the edges you take the threshold out of the equation) and
    >watch carefully as the radius is adjusted from between .4 to say 1.2. (Smaller
    >numbers on all of these steps for smaller files). I have an action from Bill
    >Atkinson that does all this automatically on the L channel in LAB mode and it
    >works great.
    >
    >If you just use USM it's hard to sharpen the high frequency data (eyes, hair,
    >leaves, etc) without adding noise to the low frequency areas (sky, out of focus
    >areas, etc), though it's much easier with digital than scanned film because of
    >the lower noise. The mask takes care of this for you by masking out the low
    >frequency areas.
    >
    >Bill
    >
    >


    --

    nobody
     
    nobody nowhere, Aug 23, 2004
    #17
  18. In article <>, Bill Hilton
    <> writes

    Me again. The Tango is pushing things a little too far for an amateur,
    isn't it? You mentioned the beast before. However, how would in your
    view the Nikon 8000 compare to a say Imacon 343, or thereabouts? (If
    you answered this question before id didn't register with me, presumably
    due to tiredness... :))). Thanks.

    PS: I am surprised that you still don't own a 4 x 5, in particular since
    you produce large prints, even occasionally. They are cheap nowadays,
    and the glass (usually Schneider) is unsurpassed (also expensive, or
    relatively expensive). Presumably is a matter of convenience.

    >>

    > The 8000
    >does a really good job for the money but the Tango has a bit more resolving
    >power and better shadow detail, though Photoshop CS's Shadow/Highlight tool
    >helps even this up a bit.
    >

    --

    Nobody
     
    nobody nowhere, Aug 23, 2004
    #18
  19. Beowulf

    Chris Brown Guest

    In article <>,
    Beowulf <> wrote:
    >Do the 11 megapixel digital cameras approach the resolution of a medium
    >format film camera? I have a Canon 10D digital, 6.3 megapixels, and a
    >Canon Rebel Ti film camera. I would like the greater resolution of a
    >medium format camera for larger finer grained prints-- but perhaps digital
    >is approaching that with 11+ megapixels?


    It probably depends on your expectations, and what you mean by "medium
    format". 645 is a whole different ballpark to 69, for example.

    Here's a datapoint that may prove useful. Last night, I was scanning a 66
    medium format Velvia slide which I'd shot handheld with an old Yashica TLR
    at 1/30 of a second. I was scanning this on an Epson 4870 flatbed, which
    doesn't come close to getting all the detail out of the slide. A 1 megapixel
    version of the image is here:

    http://www.fastfoto.co.uk/Chris/Beer.jpg

    The initial output of the scanner is 100 million pixels. They're not "DSLR
    quality" pixels, but that doesn't really matter for this point.

    Anyway, at the full 100 million pixel output of the scanner, two things are
    apparent:

    1) You can easilly read *all* the text on the label of the beer bottle.

    2) The shot has camera shake (surprisingly little for the circumstances, but
    TLRs are very stable for handheld work).

    The camera shake can be seen on the white-on-blue text on the bottle - the
    letters very obviously have a slight vertical smearing to them, with letters
    like "H" having two distinct crossbars instead of just one.

    Now I don't leave my scans at 100 million pixels, as that would be a huge
    waste of disk space, so I typically downsize them. The way this image
    behaves when downsized to different resolutions is educational:

    - At 36 megapixels (6000*6000), the image quality per pixel is similar to
    whet I get from my 10D at 400 ISO. The text on the bottle is still
    readable. The camera shake is still obvious.

    - At 25 megapixels (5000*5000), the image quality per pixel is similar to
    what I get from the 10D shots I get in good light, with a good lens, on a
    tripod. The smaller text is barely readable. The camera shake is there,
    but difficult to see if you don't know where to look.

    - At 9 megapixels (3000*3000), the image quality per pixel is vastly better
    than anything I've seen from a DSLR, with very sharply defined edges. It
    looks like an over sharpened DSLR image, minus the sharpening artifacts,
    or an SD10 type image, minus the aliasing. In short, "per pixel" at 9
    megapixels, this is vastly better than anything I've ever seen any DSLR do.

    *HOWEVER*, in that same 9 megapixel image, you can't read any of the
    smaller text on the bottle - each letter is only a few pixels wide. Also,
    the camera shake is all but undetectable.

    With a proper film scanner, rather than a flatbed pretending to be one,
    there is probably more detail to be recovered from medium format slides,
    *and* this is a handheld shot, on a 1970s camera, with its Tessar-type lens
    focused closely and wide-open (a situation where this lens type is not known
    to shine). It is, however, shot on very high resolution slide film, namely
    Velvia 50, and results from print film may be a bit softer.

    Having said all that, the conclusion I can draw from this is that using this
    combination of film, camera and scanner, and looking purely from a detail
    point of view, 66 medium format makes 11 DSLR megapixels look to be
    something of a poor relation. A 1Ds might have delivered an image where the
    smaller text on the bottle was just about readable, but it would be a close
    thing. The slide, on the other hand, has lots of detail left at that point.

    However, if you're comparing to 645 medium format, shot on print film, and
    you consider other things such as grain and shadow noise, the 1Ds probably
    looks a bit more competitive.
     
    Chris Brown, Aug 23, 2004
    #19
  20. Beowulf

    Chris Brown Guest

    In article <>,
    Alfred Molon <> wrote:
    >
    >The area of a MF frame is about four times the area of a 35mm frame,
    >isn't it ? Then, taking into account that MF lenses are less sharp than
    >35mm lenses (because they don't need to be so sharp),


    That's not a good assumption. It *may* be true with things like 645 medium
    format SLRs, but many other types of medium format cameras have very
    well-made, simple lens designs which aren't compromised by a) being zooms,
    or b) having to employ large element count, retrofocal designs to cope with
    an SLR's mirror box. I recall reading some lens test results which rated the
    lens in a 1970s Yashica Mat 124G twin lens reflex ('70s implementation of a
    19th century 4-element design) as sharper than Nikon's 50mm f/1.8 35mm SLR
    lens at pretty much any aperture, and there are much better lenses around
    than that.

    The lenses for the Mamiya 7 medium format rangefinder system are supposedly
    some of the sharpest lenses available for any camera system.
     
    Chris Brown, Aug 23, 2004
    #20
    1. Advertising

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

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Guest
    Replies:
    4
    Views:
    521
    Mxsmanic
    May 23, 2004
  2. Guest
    Replies:
    0
    Views:
    435
    Guest
    May 23, 2004
  3. spassig

    Epson Perfection 2400 - scanning medium format film?

    spassig, Oct 13, 2004, in forum: Digital Photography
    Replies:
    1
    Views:
    761
    Gene Palmiter
    Oct 13, 2004
  4. Rick Baker
    Replies:
    6
    Views:
    1,209
    Scott W
    Mar 17, 2006
  5. Progressiveabsolution
    Replies:
    15
    Views:
    727
    Alfred Molon
    Sep 17, 2006
Loading...

Share This Page