"Moire Effect" - Camera to Offset Lithography - Important Defect?

Discussion in 'Digital SLR' started by Mark Conrad, Apr 19, 2006.

  1. Mark Conrad

    Mark Conrad Guest

    I am seriousely contemplating buying a new 12.8 MP Canon.

    Would appreciate advice from actual owners of that camera.

    Contemplated usage will be mainly for producing offset lithography
    phamplets and possibly larger 8 x 11 inch "slick" brochers.

    My setup will be an Intel-based Mac using Photoshop (or other raw
    editor) for editing the raw digital image from the Canon.

    I have heard that the Canon has a fairly bad Moire effect, but I guess
    all digital cameras in that price range are afflicted with the
    Moire-effect problem, when offset lithography is used.
    (colored bands in areas that should be plain, like flesh tones)

    According to Hassleblad press releases, their newest 39 MP camera will
    incorporate Moire-effect reduction techniques, however that camera will
    be too costly for most users. I heard $43,000 (US).

    I am essentially new to digital photography. The Canon 12.8-MP might
    be my first digital camera of any kind.


    Mark-
     
    Mark Conrad, Apr 19, 2006
    #1
    1. Advertisements

  2. Mark Conrad

    w.beckley Guest

    Well, for my experience, I've shot some rather moire-prone subjects(
    fences, plaid patterns on bedsheets and shirts, through window screens)
    with my Canon 20d (8.2 megapixel) and I've seen not a single instance
    of bad moire.

    I primarily shoot motion picture stuff, and I'm known in my circles as
    a rather moire-conscious cinematographer anytime I shoot on a video
    format (most of the past year and a half). I'm confident that my 20d
    didn't produce any moire... it isn't something that I simply "didn't
    notice".

    You've, rather specifically and yet rather esoterically, isolated the
    Canon 5d as your camera of choice. It is a great camera that I've used
    a bit (and hope one day to buy), and in my limited experience, it also
    shows no unusual rate of moire artifacting. More to the point, however,
    I've extensively used my less-advanced 20d and had no problems. Maybe
    someone here can provide evidence that the 5d has a large moire problem
    that the 20d lacks, but I highly doubt it.

    I've processed all of my photos in Adobe software (first Adobe Camera
    Raw in Photoshop CS; lately in Adobe Lightroom) on a Mac (Power PC, but
    is *that* making the difference?). As far as I can tell, you've got a
    fine setup for a Canon 5d, and it is a great camera to be using. Go
    forth and take some fantastic pictures. Worry only about your own
    skills as a photographer and as a RAW processor. "Expose right" on the
    histogram, overexpose your RAW files a bit, and bring them down in
    post, and you'll get some breathtaking results.

    That's my input on the setup you ponder,

    Will
     
    w.beckley, Apr 19, 2006
    #2
    1. Advertisements

  3. Consider a Sigma SD10 which has a Foveon sensor - no moire artifacts with
    that camera http://www.foveon.com/article.php?a=70
     
    Peter A. Stavrakoglou, Apr 19, 2006
    #3
  4. Mark Conrad

    Colin D Guest

    Moire is not a problem with Canon cameras, thanks to a fairly aggressive
    Anti-aliasing filter, but the camera with the reputation for moire is
    the Nikon D70, which has a less aggressive filter to try to maintain
    sharpness. While the ex-camera image from a Canon appears less sharp
    than from a D70, unsharp masking in Photoshop will restore all the
    sharpness you want, and you don't have the disadvantage of risking
    moire.

    Of course, if you will be shooting half-tone printed images, then moire
    is almost certain to show up with any digital camera - except perhaps
    the Sigma with its Foveon chip, but the color results from Foveons
    leaves a lot to be desired. If you want accurate color, specially skin
    tones, a Foveon won't do it.

    Colin D.
     
    Colin D, Apr 19, 2006
    #4
  5. I have a 5D and have never had this problem unlike a Nikon P/S of my
    wife's which does.

    Even if you get them, they can be fixed in post processing. See:

    Photoshop LAB Color: The Canyon Conundrum and Other Adventures in the
    Most Powerful Colorspace (Paperback)
    by Dan Margulis

    for an example of how to fix it.


    --

    "It looked like the sort of book described in library
    catalogues as "slightly foxed", although it would be
    more honest to admit that it looked as though it had
    been badgered, wolved and possibly beared as well."

    _Light Fantastic_
    Terry Pratchett
     
    John A. Stovall, Apr 19, 2006
    #5
  6. Mark Conrad

    Mark Conrad Guest

    Thanks very much for your comprehensive post about the Moire effect.

    I have a very shaky, and probably incorrect, understanding of Moire,
    even though I Googled it to death.


    Brevity is nice, however the only way I know how to explain my present
    understanding of the Moire effect is the long-winded way below, forgive
    me.

    Start of Long Winded Explain of Moire Effect
    *******************************
    My present idea is that it all starts with the fact that a digital CCD
    has rows of ordered sensor cells, one row red sensitive, next row
    green, next row blue. This pattern of rows repeated across the CCD.

    I think the Moire problem can arise even when photographing a plain
    surface with no pattern on it, such as a light brown wall.

    As we know, light brown can be simulated by a mixture of colored dots
    in a lithographed image, like the images we see in ordinary magazines.

    This simulated-brown-dot I will call a "sim-dot", actually a
    collection of smaller dots.

    These lithographed sim-dots are in a regular pattern on the paper.

    If the paper image size is an exact multiple of the CCD image, no Moire
    effect will be present, because each red, green, blue row of CCD cells
    are trying to "influence" one row of sim-dots on the paper image.

    _If_ the paper image is _not_ an exact multiple of the CCD image
    size, a situation exists where there is periodically a surplus row of
    red cells in the CCD trying to "influence" the color of the paper
    sim-dot, _then_ a bit further on in the image there will be a surplus
    row of _green_ dots in the CCD trying to "influence" the paper
    sim-dot, _then_ further on in the image there will be a surplus row
    of _blue_ dots in the CCD tryng to "influence" the paper sim-dot.

    The net result will be repeating bands of screwed up color all across
    the image of the light brown wall, the "Moire effect".

    How close the screwed up bands of color are to each other will be a
    function of how far off from "exact ratio" the sizes are.



    The reason Moire _can't_ occur when the litho' image is an exact
    multiple in size of the CCD image:

    Assume there is one extra row of red cells in the CCD trying to
    "influence" the color of one row of litho' sim-dots.

    red green blue red green blue red <-- one extra red

    Because the image and paper sizes are _exact_ multiples of each
    other, the above 'excess' of red cells will _always_ be an excess of
    red cells, making for no change in color of our light brown wall, i.e.
    no "banding" of a slightly different color.

    In other words, no slight shifting to a different 'extra' color - - -
    or if very slight shifting _does_ occur, then it would not be
    noticable because the whole image of the wall had to be traversed
    before the 'different' excess color appeared.

    For example, no situation like this can arise, with exact ratios:

    green blue red green blue red green <-- extra green, not red

    ******************************
    End of Long Winded Explain of Moire Effect


    Like I said, ny understanding of the Moire effect might be wrong.

    Mark-
     
    Mark Conrad, Apr 19, 2006
    #6
  7. Mark Conrad

    Scott W Guest

    You will not edit the raw file but rather convert it and then edit it.
    The convertion process
    does looks a bit like editing but it is not, you are adjusting how the
    raw file is converter
    not changing the raw data.
    You should not get any moire problems in fleash tones, it take a
    pattern in the scene
    to get a moire patten. I use a 20D and get a noticable moire pattern in
    maybe 1 out of 1000 shots, a lot depends on what you will be
    photographing.

    Given the fairly high cost you might want to rent one for a while and
    see how you get along with it.

    Scott
     
    Scott W, Apr 19, 2006
    #7
  8. Mark Conrad

    tomm42 Guest

    Don't know if you are over analizing a digital image, are you
    photographing images that are made with offset lithography or using
    digital images in offset lithography. With the former you should be
    concerned, with the latter there shouldn't be a problem. Yes several
    cameras are know for moire the Nikon D70 for one and the Kodak DCS
    cameras fore another. I have used a DCS 760 (don't photograph a small
    plaid shirt with this camera) and printed offset from the files with no
    moire, even going to 18x24 offset prints of a baseball team, lots of
    flesh tones. Also almost 100% of news photos are now digital, I work
    for a hospital magazine that requires digital to avoid scanning costs.
    Is offset lithography significantly different from these applications?
    I really would like to know.

    Tom
     
    tomm42, Apr 19, 2006
    #8
  9. Mark Conrad

    Mark Conrad Guest

    Does your wife's Nikon have the problem even in the plain non-patterned
    parts of images, like the sky?

    Are the Moire patterns still there when printing the final image to
    regular photographic enlargement paper?

    I would really like to know the answer to those questions, as it would
    save me a tremendous amount of time running my own tests.

    Thanks,

    Mark-
     
    Mark Conrad, Apr 19, 2006
    #9
  10. Mark Conrad

    Mark Conrad Guest

    I see what you mean, any filter that blurs the individual dot pattern
    of the CCD will make the digital image like the image from a regular
    film camera, at the price of causing minor unsharpness. We all know
    that unsharp masking to restore sharpness has its own problems.

    That seems a poor tradeoff, IMO.

    It takes a sharp patterned image from the Canon CCD, and converts it
    into a slightly unsharp image to "mask" the dot pattern of the CCD.


    I am sorely tempted to consider buying a Nikon instead, then forget
    about using high priced offset lithography output for my short runs of
    phamphets, brochures - - - then print the final _sharp_ Nikon images
    to regular photographic paper, where the Moire effect does not exist.

    (because of no regular "dot pattern" on photographic paper, which is
    absolutely necessary to trigger the Moire effect)

    ....according to the Google article explaining the Moire effect.

    Mark-
     
    Mark Conrad, Apr 19, 2006
    #10
  11. Mark Conrad

    Mark Conrad Guest

    You will not edit the raw file but rather convert it and then
    edit it. The convertion process does looks a bit like editing
    but it is not, you are adjusting how the raw file is converter
    not changing the raw data.[/QUOTE]

    Thanks, I obviousely have a lot to learn.


    Wonder where I went wrong in my long winded "analysis".

    Think I will go back to my Google sources, and read them again.



    I know there is a defect with digital that _does_ require a pattern
    in the original image. That defect often shows up when photographing
    fine lines in a drawing. It results in colored "fringes" on the edges
    of the fine lines, which have to be laboriously removed by using an
    editor to remove the colored fringes, pixel by pixel.

    That particular defect never shows up when a smooth blank area of the
    image is being photographed, it _needs_ a "pattern" like fine lines
    in the image.


    I think some here, (probably me), are confusing the above described
    defect with the "Moire Effect".

    I will go back and re-read my original Google sources about the Moire
    effect, to try to resolve where I went wrong.


    As the literature on the Moire effect explained it, the Moire effect
    _only_ occurs when the final image has a dot pattern, like the dot
    pattern in an offset lithography magazine photo, or the dot pattern in
    old fashioned halftone newspaper photographs.

    Print the same image to regular photographic paper instead, and the
    "Moire Effect" defects should _never_ be evident in sky, blank walls,
    lerge areas of skin tones, etc., etc.


    Dunno, I might have read the source articles on the Moire effect wrong,
    I will go back and re-read them.


    Great suggestion. I can then run some tests by submitting my photo's
    of blank areas to an offset lithography shop, being careful to ensure
    that the litho' image size is _not_ exact multiples of the CCD image
    size, and put this entire "Moire effect" issue to rest.

    'Course, the litho' guys will call the guys in the white coats to haul
    me away, when I ask them to process images of blank walls.<g>

    Mark-
     
    Mark Conrad, Apr 19, 2006
    #11
  12. Mark Conrad

    Mark Conrad Guest

    I will be using digital images in offset lithography, not photographing
    images that are made with offset lithography.

    Offset litho' is different in the sense that very large high res'
    images of blank areas like flesh tones are liable to be in the final
    litho' image, whereas newspaper photo's are less liable to be "fashion
    shots" with large areas of skin tones.

    Also, the newspaper editor can change the size of the image in his
    paper, to minimize any Moire pattern by making his final image an exact
    multiple of the CCD image size.

    Clients for extremely high cost fashion images are less inclined to
    approve any change in final image size.



    My Google sources explaining the Moire effect, claim it creates banding
    even in areas that have no patterns, like plain walls, sky, skin tones,
    etc.

    Has something to do with the dot pattern in the camera CCD being
    "out-of-sync" with the dot pattern of the displayed image in the
    newspaper or slick magazine.

    According to the Google source, the Moire effect only occurs when the
    final image is created with a regular dot pattern, as in a newspaper
    photograph, or a magazine offset litho' rendition.

    Print to regular photographic paper that has no dot pattern, and the
    Moire effect should never occur.

    In my mind, this means the Moire effect is different than the colored
    fringes on fine lines, commonly observed when photographing plaid
    shirts etc.
     
    Mark Conrad, Apr 19, 2006
    #12
  13. Mark Conrad

    Scott W Guest

    Test for two cases that give digital camera the most problems, fine
    patterns, like you might find in some fabrics, and detail that is in
    color, such as red lines on a blue background, or the other way around.

    You should not have a problem with moire pattens between the cameras
    pixels and the screening, at least if some care is done in the
    screening. The screening is just a resampling of the digital image and
    this normally does not cause a problem, at least if the person doing
    the screening knows even a little about what they are doing (which they
    might not).

    A moire pattern is one form of aliasing, lines that are jegged is
    another. I can make it show up at will, shoting through a screen at an
    angle usning a larger f number will do it. but it is pretty rare in the
    kind of photos I take. But be aware that you subject matter can have a
    large impact on how often you see it.

    Scott
     
    Scott W, Apr 19, 2006
    #13
  14. I find the colors from my Sigma to be just fine - with the exception of
    skintones as you noted but those are easily corrected. It's also an
    intermittent problem, not a constant one. FWIW, did you view the site I
    linked to? The fabric example used to show the moire artifacts on Foveon's
    site is almost identical to a shirt I have - even the color. I could not
    reproduce much of any moire artifacts using my shirt and a Nikon D50. How
    much of a real probelm is this for most DSLRs anyway?
     
    Peter A. Stavrakoglou, Apr 19, 2006
    #14
  15. Mark Conrad

    JPS Guest

    In message <>,
    http://www.pbase.com/image/37173876
    --
     
    JPS, Apr 19, 2006
    #15
  16. No, but Houndstooth tweed has it very bad.
    I don't print digital prints on "regular photographic paper" as that
    would take a wet dark room. I use Epson Premium Photo Glossy to print
    with in an Epson R1800 and it is still there.


    --

    "I have been a witness, and these pictures are
    my testimony. The events I have recorded should
    not be forgotten and must not be repeated."

    -James Nachtwey-
    http://www.jamesnachtwey.com/
     
    John A. Stovall, Apr 19, 2006
    #16
  17. Mark Conrad

    w.beckley Guest

    I primarily shoot motion picture stuff, and I'm known in my circles as
    I didn't say it was impossible, I said I hadn't seen it in my
    photographs even when shooting various "real-world" subjects noted for
    producing moire.

    I didn't shoot any monitors though, and I don't think I've ever had
    need to do so in photography. I suppose if the original poster is
    shooting alot of monitors he should avoid the 20d and likely every
    other prosumer digital camera on the market.

    Will
     
    w.beckley, Apr 20, 2006
    #17
  18. Mark Conrad

    JPS Guest

    In message <>,
    .... and I didn't say anything. I just linked a photo.
    If it can happen on a monitor, it can happen on any pattern of colored
    dots with the same angular resolution, reflective or source. A sharp
    lens "helps" (I was using the Tamron 90mm Di Macro lens, an extremely
    sharp lens).
    --
     
    JPS, Apr 20, 2006
    #18
  19. Mark Conrad

    Mark Conrad Guest


    Well guys, I went back and checked my Google sources.

    I was wrong in my posts - - - I hate when that happens.

    The "Moire Effect" is totally different than I thought it was.

    Apparently, like everyone was trying to tell me, patterns in the
    original object-being-photographed are necessary to trigger the Moire
    effect.

    I was dead wrong when I posted that there can be Moire defects when
    photographing a plain sky, or a wall without any texture, or broad
    areas of evenly colored skin tones.


    Also, I read a very large number of posts from poor devils who had
    Moire defects in their final images. Those guys often tried
    everything suggested to them by pros, and _still_ could not get rid
    of the Moire defects.

    It seems to be something of a black art, as to what works and what
    doesn't work.

    Even ordinary film cameras are sometimes affected by Moire defects.

    It was pretty well agreed that some models of Nikon are afflicted with
    more than their share of Moire problems.

    Come to think of it, I sometimes see Moire defects on my television,
    when the weatherman wears some types of patterned ties.<g>

    Hope the Canon 12.8 MP camera I am considering buying only has its
    "ordinary" share of Moire problems.

    Apparently this is something we all have to live with, trying this and
    that to minimize Moire defects.



    Sorry everyone, for wasting all your time because of my own inept
    reading of stuff about Moire.

    Mark-
     
    Mark Conrad, Apr 20, 2006
    #19
  20. Actually, the Canon EOS 5D, along with the Nikon D70, has a
    fairly weak antialias-filter and is therefore more subject
    to moire than, say, the EOS 20D or 30D.

    For this reason, I use an image from the EOS 5D to illustrate my
    article on (colour) moire:
    http://hannemyr.com/photo/defects.html#moire

    That said, moire is not something that you'll run into often with the
    EOS 5D. It will only occur when certain subjects (e.g. fine fabric
    with a regular pattern) is undersampled.

    Different RAW converters treat moire differently. If one RAW
    converter gives you moire with a certain subject, re-processing
    the same file with a different RAW-converter will often fix the
    problem.
     
    Gisle Hannemyr, Apr 20, 2006
    #20
    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.