DSLR v Consumer Image quality

Discussion in 'Digital Photography' started by oink, Feb 22, 2005.

  1. David J Taylor, Feb 24, 2005
    1. Advertisements

  2. Stacey wrote:
    []
    This is not a binary group - posting images would be against the rules.

    David
     
    David J Taylor, Feb 24, 2005
    1. Advertisements

  3. Mark² wrote:
    []
    But you still have the penalties of size, weight, bulk and cost. I still
    haven't see a DSLR with a 36 - 432mm f/2.8 image stabilised lens weighing
    less than 600g, for example.
    Simply that for many people the DSLR is not the best solution. I'm not
    taking either side, simply saying that DSLRs have both plus points and
    minus points, and different people will weight those pros and cons
    according to their own needs and aspirations, I think it has been
    established (by reports here) that for up to 10 x 8 inch prints the best
    of the non-DSLRs can be in the same image quality class as consumer DSLRs.

    You pays your money and takes your choice.

    Cheers,
    David
     
    David J Taylor, Feb 24, 2005
  4. That doesn't follow.

    At any given _instant_ the eye will be gathering a very detailed image from
    the central, small x% of "your field of view", but this isn't what you
    "see". In normal vision, the eye is constantly scanning all over the
    scene; the brain accumulates this series of "highly details small-FOV
    snapshots" and "integrates" them to provide the "virtual" image that we
    "see". Think of a tight searchlight scanning around a dark scene -- each
    spot that is lit adds to the "picture" that our brains build up.

    If you stop/limit eye movement by either staring at fixed point (as in the
    experiment "chrlz" [Charles?] describes), or by _gently_ pressing against
    the sides of your eyes, then the brain only gets the same image over and
    over again (the searchlight has stopped moving), and you can only see a
    small area in detail. (And, as an aside, you can notice depth-of-field as
    well: normally, as the eye is roaming around a scene, it constantly adjusts
    focus so you "see" with a very deep depth-of-field. If you fix on one
    spot/object, then you can notice that objects at other distances are
    blurred).

    As an even further aside, the notion of printing just the centre of an
    image at high resolution _could_ be applied to improving the frame-rate of
    3D computer-generated images (e.g. Quake/Doom engines): You would need a
    sensor that tracked where the eye was looking; that portion of the screen
    would be rendered in full detail, but everything else (the bits that would
    be in _that_instant's_ peripheral vision) could be rendered with much lower
    resolution (and hence faster). As your eyes tracked around the full image,
    the area of high detail will follow.



    Regards,
    Graham Holden (g-holden AT dircon DOT co DOT uk)
     
    Graham Holden, Feb 24, 2005
  5. I never meant to suggest _my_decision_ was a "one size fits all", just to
    give another data-point. I'll admit that Ken Rockwell's pages tend to be
    much more dogmatic about such things: while I found/find them useful, you
    do have to take his "passion" into account when reading.

    Obviously all the factors you mention need to be considered on an
    individual basis; but if you're already into the price/size range of
    something like an 8700 (which is larger/more expensive than a "normal"
    p&s), then it's not _much_ of a step to a dSLR. A dSLR _won't_ be the
    right camera for _everybody_, but most/all people who are about to buy an
    8700-like camera should at least _think_ about what a dSLR offers (and its
    drawbacks), and at the very least consciously decide to reject a dSLR,
    rather than reject them out of hand.

    Again from my experience: When I first started looking, I went into a
    camera shop to have a look at and ask about the 8700. The assistant asked
    if I'd thought about a dSLR, and I dismissed this out of hand with
    something like "No, fiddling around with multiple lenses etc. is getting a
    bit too serious for what I want.".

    After seeing Ken's and others' pages, I reevaluated my opinion and
    (possibly narrowly) came down in favour of a dSLR (and I [still] think that
    was the right decision for me).

    However, even if I _had_ decided to stick with the 8700 (after actively
    rejecting a dSLR), if I had made that choice at the beginning (in the
    camera shop), I would have been at fault: not so much for making the wrong
    decision, but for using the wrong decision-making _process_.


    Regards,
    Graham Holden (g-holden AT dircon DOT co DOT uk)
     
    Graham Holden, Feb 24, 2005
  6. Graham Holden wrote:
    []
    Yes, it was Ken's pages I was referring to.

    []
    Completely agree, Graham. That's what had been rather one-sided (I felt)
    about some of the discussions before. Of course, there is also the
    problem that until one has used both types of camera in anger, one may
    have to rely on third-party reports. I am also guilty of that in that my
    35mm SLR experience is based on film cameras. Using the more compact
    small-format cameras with flexible viewfinders has been a revelation
    (particularly the swing-body Nikon Coolpix series), and I would not wish
    to go back.

    Cheers,
    David
     
    David J Taylor, Feb 24, 2005
  7. oink

    Larry Guest


    My only reason for wanting to go to a DSLR is "Shutter Delay".

    The improved picture quality most DSLR cameras would give me is a "Side
    Bonus" thrown in with the deal.

    Even the very fastest of the "Digicams" arent fast enough to make capturing a
    moving horse at full extension of his stride into a matter of timing. (as it
    is with an SLR)

    Its this functionality that (so far) has not become a reality in the
    "digicam" breed.. Perhaps it never will be there. Then again maybe it will
    get there, but Im no longer willing to wait.. I've missed too many shots, and
    I dont want to miss any more.

    The DRebel that I tried had the timing I want, but it had instantaneous dust
    problems. (as it turns out, after experimenting with my film slr cameras and
    working with other cameras in the same place, the dust problem was probably
    inherant with that particular camera and not DSLR cameras in general)

    Depending on price structure, I may well go with the "Old" Rebel now that I
    have convinced myself that the dust problem has a cure. The old DRebel with
    the firmware hack is probably just what I need. (6mp is enough I think for
    what I'm doing).

    I have looked at the Canon 20D and though it is GREAT it is more Camera than
    I need, and its bigger, heavier, and more expensive.. If a Horse were to
    collide with me and my camera in the show ring (thats how I lost my last
    working Yashica 35mm slr) my insurance company has already told me they wont
    cover it, so the loss would be VERY dear.

    I looked at, held, and used the E-volt, then brought the memory card home,
    and was VERY dissapointed with the picture quality. (the same subject matter
    would have been at least as good with my Sony 828 at the same ISO).

    I have also tried the D70, but didn't see an advantage (for me) in the '70
    over the Rebel. (Im pretty sure, if pressed, I would say the '70 is better,
    but I own lenses that fit a DRebel...none for the '70)

    There are reasons OTHER than high ISO noise (or the lack of it) for moving
    away from "digicams" and into a DSLR, interchangeable lenses dont even enter
    the equation (for me). I have always done my "showring" shooting with one
    lens at a time, usually some sort of zoom, depending on what I have at the
    time. (first choice for inside the ring is 50mm-150mm equivalent f8 to f2.8
    if I can get it.)

    Anyone who has "Horse Show Experience" knows that you dont want to be
    changing a lens in the ring... You need one eye out for the horses at all
    times and one eye for the view-finder.

    At the risk of wasting bandwidth, I'll give my analogy here:

    Being IN the horse show ring is like being INSIDE the fence at an Auto Race..
    its more dangerous than being outside the fence.

    Then add to that:
    Most of the time, the driver (on a horse) is not actually in control.
    He only has the horses cooperation, and that is a big difference. The horse
    can choose not to cooperate at any time.

    Due to the way a horses vision works, you can get run over because the horse
    didn't see you. OTOH if a horse panics, he can run you down whether he sees
    you or not, It really depends on how YOU react to the horse as he charges
    down on you.

    That show ring (150' by 225') that looks so huge when empty, seems REALLY
    crowded when you put 24 to 36 horses, 2 judges, 2 ring stewards and a
    photographer into it.

    Change lenses??? No thanks...

    All of this is only here to show--- Everybody has a different perspective,
    and none of them is wrong.
     
    Larry, Feb 24, 2005
  8. oink

    Skip M Guest

    I'd have to say that a DSLR isn't the best solution for the overwhelming
    majority of camera owners. Most people buy cameras merely to record
    moments, "I was there" or "He did this." Detail, color rendition, any of
    the other things that an SLR type camera is capable of are moot points to
    them.
    This is not meant to be insulting, it just is the way things are.
    Otherwise, nobody would by the 2-3mp $100-200 digital cameras, and this
    newsgroup would have several million subscribers.
     
    Skip M, Feb 24, 2005
  9. This demonstrates your printer is the limit to your perceived
    sharpness, not your eyes.
    On a high contrast target, the eye can do much better. For
    example, laser printers: many people can see the difference
    in edges printed from a 300 dpi printer versus a 600 dpi printer.
    Many can also see the difference between 600 and 1200 dpi printers.
    Then people say photographs do not have high contrast edges.
    Yes they do. For example snow-capped mountains, shadow boundaries,
    tree branches against the sky or shadows, etc. So if the eye can
    resolve 600 lppi, the angular separation of two black lines
    at 600 lppi with a target at 10 inches is 6 arc-minutes.
    (lppi line pairs per inch) (The centers of two black lines
    has a separation of 1/600 inch.) The specification
    for visual acuity is 0.6 arc-minute.
    See: http://clarkvision.com/imagedetail/eye-resolution.html
    Your printer needs to resolve 600 lppi cleanly if you want
    to reach the limits of human vision; most do not.

    This gets back to what I said earlier: printer technology is
    limiting what we see in printed images. But slowly the
    standards will be raised. The results above also indicate
    why large format contact prints look so impressive: the
    detail is there, and humans can see it!

    Roger
     
    Roger N. Clark (change username to rnclark), Feb 24, 2005
  10. oink

    Owamanga Guest

    Even 200% is too much.

    I critique all my images at about 33% - so the whole frame fills my
    screen. The decision to keep/dump is going to be made on tonal
    quality, lighting, subject's expression, focus, exposure accuracy and
    composition. I can see all of this at 33%. If I feel focus may be
    slightly off, then I'll go in further, but *never* more than 100%. The
    only other thing the 100% view shows me is the level of noise present
    in the image that may affect my choice & method of sharpening, but
    doesn't affect my decision not to print it any more than the
    appearance of grain would hold me back from doing a big enlargement.
    The argument is that when comparing image quality from two cameras,
    it's meaningless discerning differences that are only apparent at 200%
    when these do not show up on the print. From that perspective it *is*
    necessary to make a print - this is a very simple way to eliminate any
    differences between the two cameras that are irrelevant.
    When doing a comparison, to weed out what's relevant and what's not,
    yes, we do require the losses that printing introduces. The print is
    the final aim here. It's the point of doing this.

    Anyway, preposterous? Get real. A big generalization, I'll accept.
    There are exceptions, but for the vast majority of users:

    *** We photograph to print. ****

    If the problem can't be seen on the print, the problem doesn't exist.
    It's that damn simple. The print is the *only thing that matters*.

    Go ahead, buy a 100Mpix camera when one comes out, at normal viewing
    difference, the printed 8x10s will look no better than what we can get
    from a 6Mpix camera today.

    Lots of people do it, every day - scrutinizing a 6Mpix image at 200%
    that is destined for an 8x10 print or smaller. This is like critiquing
    a Rembrandt with a microscope - unrealistic and unnecessary.

    David LittleBoy claims his printer can resolve to this level, and I
    don't disbelieve him. But if he's culling prints based on some problem
    that is only visible at 200% magnification, or by finely studying the
    intricate details of the print from a few inches away, then he too,
    needs his head tested. Flaws of this nature, in the real world, *do
    not matter*.

    Yes, there are exceptions. An aerial photo for example, you would
    strive for the upmost image quality, sharpness and highest resolution
    because people *will* study those things with a magnifying glass. But
    for the vast majority of subjects, at these levels of magnification,
    all we can see is noise, artifacts, scanned film grain and the pixels.

    I can't believe that a photographer would really dump an image that is
    otherwise perfect due to a nasty artifact that can only be seen at
    200% magnification, this is just *insane*. Other attributes in the
    image are always *far* more important.
     
    Owamanga, Feb 24, 2005
  11. oink

    Owamanga Guest

    Absolutely. More people need to step back and realize what they are
    doing .
    I don't even think it matters for a 5x8 foot print. Take a look at the
    quality of print used on one of those huge underground/subway
    bill-board style posters. I guess they are about 20ft wide and 15ft
    high. The dots are the size of your fingerprint (slight exaggeration),
    but you are standing the other side of the track - at least 10ft away.

    Viewers *don't care* about this type of problem, they just stand back
    and admire the subject. People can jump on me and say 'that's not a
    photograph', fine, but my point is advertisers don't care and neither
    do the viewers because the photo still works.
     
    Owamanga, Feb 24, 2005
  12. oink

    Scott W Guest

    I assume the above is a typo and you meant 600 lppi at 10 inches is
    0.6 arc-minutes not 6 arc-minutes.

    My book has more like 0.7 arc-minutes for the limit of the human eye,
    but again this is for 100% contrast detail, and this gets a bit better
    with better brightness.

    But you are 100% correct that if you want the printer not to be limited
    what the eye sees it needs to resolve 600 lppi and not just resolve it
    but do so with close to 100% modulation. And of course the image source
    needs to have its MTF be close to 100% at 600 lppi as well.

    For those of us with old eyes another thing comes into the equation and
    that is what is our visual acuity at 10 inches? My bifocals have 2
    diopters of correction for near vision, this puts my near focal point
    at ½ meter or close to 20 inches. I am near sighted by 4 diopters and
    so can see clearly at 10 inches with my glasses off. At my age most of
    my friends are using bifocals and so their near vision is pretty
    limited.

    My parents who are older yet have a very hard time seeing the
    difference between a print at 150 ppi and 300 ppi. When I tell them
    that the 8 x 10 prints that they make from their 2 MP camera look poor
    they just can't understand it.

    BTW for referance my eyes are 52 year old.

    I think this is why large prints like 20 x 30 are popular, you can
    view them from 20 inches back and not get eye strain. If you really
    want to improve the detail and sharpness of a print you need to get it
    at a very comfortable viewing distance.
     
    Scott W, Feb 24, 2005
  13. I have been concerned for decades with the acceptance of continuing
    lower quality photographic imagery. I thought digital might be an
    exception, because folks seem to really want the best image in their
    digitals.

    I got into photography when 35mm was well established (though my first
    camera was a cheap 120 one). So I was not around when folks bitched
    about the quality reduction with the small film.

    And I saw a sign of sanity when the 110 and 126 formats flopped. But
    folks seem to be happy with APS even at the expense of image quality.
    And, I have seen so many lenses gain acceptance that offer convenience
    over image quality.

    I then saw the Point and shoot cameras come to dominate the market over
    the SLR, which at one time was the primary 35mm type sold. Yeah, we
    have a cheap P&S film camera, but when we want a really good photo we
    drag out one of our good SLRs.

    In summary, I am glad to see this discussion because it shows that there
    ARE other people out there who care about image quality. :)
     
    Don Stauffer in Minneapolis, Feb 24, 2005
  14. In what way is it better? Why does the small sensor of the human eye
    have better resolution than the large sensor of a 35 mm frame?

    The retinal area of the eye is smaller than the 35mm format? News to me.

    I thought area of retina greater than area of 35mm frame, and spatial
    resolution in lp/mm worse over most of that area than most 35mm films.
     
    Don Stauffer in Minneapolis, Feb 24, 2005
  15. oink

    Scott W Guest

    I shoot with both a F828 and a 20D, looking at the screen the 20D blows
    the F828 out of the water. And in fact looking at the photos just on
    screen one would never buy the F828 over the 20D (and maybe they would
    be right) but when printing out 8 x 10 the differnace is pretty small
    and are hard for most people to see. Below is a link to the same scene
    taken with both the 20D and the F828, taken within 5 minutes of each
    other.

    http://www.sewcon.com/photos/compare/Sony F828.jpg
    http://www.sewcon.com/photos/compare/20D with 50mm lens.jpg

    The F828 photo is from raw and with lots of sharpening and no luminance
    smoothing, which is why there is so much noise. The same is true for
    the 20D image but the noise is so low on it to begin with that you
    don't see noise on it.

    When I printed both of these photos out at 8 x 10 I was amazed that the
    two look very close to the same, when on the screen the F828 photo
    looked like crap.

    Now clearly I could get a larger print from the 20D photo then I could
    from the F828 photo, but when limited to 8 x 10 prints the F828 holds
    up very well.

    The sad thing is that people will look at the these two photos on the
    screen and say to themselves "his crazy if he things he is going to get
    a good print from the F828 photo" and they will never bother to print
    out the photos to see for themselves.

    I love the 20D for a whole lot of reasons. I do like that fact that I
    can look at the photos at 100% on the screen and they look clean. I
    also love how fast it is and that I can shoot in low light, but if my
    only criteria for a camera was how well does it produce 8 x 10 prints
    there would be little need for it over the F828.

    Scott
     
    Scott W, Feb 24, 2005
  16. oink

    Scott W Guest

    Ok for all the people who claim you can tell how a photo will look as a
    print when viewed at 200% on the screen I have a question for you. How
    many of you use a magnifying glass to view your prints to see how they
    came out? I can just see you with a magnifying glass looking at two
    prints and saying "See I knew this print would be better".

    If you really want a sanity check show your prints to a number of
    people and see what they say. I posted these links before but here they
    are again, the two photos are from a F828 and a 20D, print them out as
    8 x 10 and show them to friends and family and ask them which is
    sharper, but don't give them a magnifying glass to look at them.

    You know what I think you will find, people won't be able to tell the
    difference.

    http://www.sewcon.com/photos/compare/20D with 50mm lens.jpg
    http://www.sewcon.com/photos/compare/Sony F828.jpg

    Scott
     
    Scott W, Feb 24, 2005
  17. oink

    rafeb Guest

    Stacey wrote:


    I've been quite consistent, actually.

    I post scan snippets based on a small, fixed area
    of film (or a portion of a digicam sensor) so that
    we can look at these at 100% without downsampling.

    That's a scientific, practical approach to
    a wide range of real world measurements --
    when it's imractical to measure the whole
    thing, we work with standardized samples.

    It's fortunate for you that your friendly
    phlebotomist only needs a couple dozen
    milliliters of your blood to get a useful
    analysis.

    I don't pretend to speak for "95% of camera
    users" and neither should you.



    rafe b.
    http://www.terrapinphoto.com
     
    rafeb, Feb 24, 2005
  18. oink

    bob Guest

    I declared the image to be 600ppi in photoshop, so the smallest lines
    are 300 lines per inch. I can easily see the 2 pixel lines (150/in) but
    not the 300/in, so I can probably see the 200/in that your book cites.

    But as it says right at the top of the page where the chart comes from,
    most people can easily see the difference between 300dpi laser prints
    and 600dpi.

    Try this test. Print out small text on good paper at 600dpi. Now lay
    that next to an ad at the back of Bon Appetite magazine or the like. I
    don't have any problem at all telling the difference between 600dpi and
    2400.

    I've already stipulated that I can't resolve even 300 lines per inch,
    but yet I can discern the difference in resolution. Of course that's
    very high contrast. But at least some of my photos have at least some
    areas with high contrast, too!

    Bob
     
    bob, Feb 24, 2005
  19. This has been done several times in reality.

    Somebody built a jet fighter flight simulator that used two image
    projectors. The one projector showed a very wide-angle but
    low-resolution image projected onto the inside of a dome screen around
    the simulated cockpit. The second projector produced a much higher
    angular resolution over a much smaller angle, and it was mechanically
    aimed at wherever on the dome the pilot was looking (measured by a head
    tracker of some sort).

    Another version: somebody built a hybrid high res/wide angle computer
    display by fitting a LCD monitor into the middle of a white projection
    screeen. The screen was illuminated by a data projector mounted on the
    ceiling above the user's head. The image on the LCD and the image
    projected on the screen were synchronized to show the same scene at a
    matching scale. So the user would see a very wide-angle view of
    whatever they were looking at (e.g. a map) at low resolution, but they
    could scroll (2D image) or re-aim the virtual camera (3D) image to bring
    any particular area of the image onto the high-resolution LCD area close
    to them. (In this case, the high-res area doesn't track the eye gaze;
    it's fixed and you have to drag the content to it).

    Dave
     
    Dave Martindale, Feb 24, 2005
  20. oink

    bob Guest

    I use an 8x loupe. Or a reversed 50mm f/1.4. With digital I don't
    bother, of course, because they're all the same.
    Not quite: when you use the loupe you say, "Oh dear. The one with the
    better composition doesn't have quite as much detail. What to do..." ;-)
    Yeah, but most people can't tell the difference between color laser
    printer and wet print, either. And that doesn't bother me.

    Bob
     
    bob, Feb 24, 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.