Where should shake compensation occur?

Discussion in 'Digital Photography' started by Ockham's Razor, Oct 19, 2006.

  1. I have been reading about shake compensation in cameras and in lenses.

    Which is better or more important?

    Ockham's Razor, Oct 19, 2006
    1. Advertisements

  2. Ockham's Razor

    Hebee Jeebes Guest

    In the body. It maybe a smidgen less effective but cost wise for the
    consumer it is better than what Canon and Nikon does by raping their
    customers with the high cost of stabilized lens and by putting out limited

    Body stabilization is the future, sooner or later they will all have it.

    Hebee Jeebes, Oct 19, 2006
    1. Advertisements

  3. Ockham's Razor

    Bob Williams Guest

    "IF" both systems were equally effective, it would make more sense (from
    the consumer's standpoint) to have shake compensation on the sensor
    rather than on the lens.
    At the present time, I think popular opinion is that IS in the lens is
    more effective than IS in the sensor.
    Bob Williams
    Bob Williams, Oct 19, 2006
  4. Ockham's Razor

    Pete D Guest

    They could of course be because Canon and Nikon both have good systems of IS
    in their lenses, currently there are less choices for in body that work as
    well as the in lens systems, this is slowly changing with the introduction
    of better in body systems such as in the Pentax K10D. The in body systems
    may of course never quite catch the in lens systems but if they can offer
    70-80% or more performance from the in body systems then I am thinking that
    many people will see that as good enough. The Pentax K10D is possibly as
    good as the in lens systems but until production models arrive it is
    difficult to evaluate, that said the K100D SR appears to be a solid IS
    systemthat is giving 2-3 stops for any lens attached, gotta be happy with
    Pete D, Oct 19, 2006
  5. Ockham's Razor

    jpc Guest

    The Lens. Maybe at some point it will get better, but right now
    many of the cameras with onboard IS have noise pickup problems.
    Your image maynot be so blurry but it will be noisy.

    jpc, Oct 19, 2006
  6. Ockham's Razor

    Hebee Jeebes Guest

    Not by much and like with all new technology it will improve. In body IS is
    where its going to be.

    Hebee Jeebes, Oct 19, 2006
  7. Is this true with long lenses that stick out a good way from the front
    of the camera? Since one is holding the camera is will "shake" thru
    lesser arcs than the front of the lens.
    Ockham's Razor, Oct 20, 2006
  8. Rotation is what you need to measure....

    David J Taylor, Oct 20, 2006
  9. Ockham's Razor

    Pete D Guest

    Source? Samples?
    Pete D, Oct 20, 2006
  10. Ockham's Razor

    jpc Guest

    You can see the effect for youself like I did. To duplicate
    my measurements

    1- Go to -http://www.imaging-resource.com. For an IS camera go the the
    sample images in the review of the Canon S3IS. Find the low light
    Dave Box images--they are near the bottom-- and download the full
    size jpg of the 400 ISO 1 footcandlle test image. Do the same for the
    Fugi E900, Oly SP-350 and Canon EOS-5D cameras.

    2 Go to http://rsb.info.nih.gov/ij/download.html and dowmload
    ImaaeJ. It a free and good quality image analysis package develiped by

    3 Open the S3IS image in ImageJ. We're interested in the 6 tone grey
    scale in the the MacBeth color target and the longer Kodax step scale
    to the right. Use the line profile tool--the line icon on the ImageJ
    control bar- and drag a line down the Kodax step scale. Hit cnrl-K-and
    you will have a noisy staircase plot of various light value with the
    noise at each value superimposed on the steps. The ratio of the value
    is the S/N at each light intesity

    4 Do the same with the other 3 images. The EOS-5D graph will show the
    low nosie of a excellent (and very expensive) full frame DSLR. The
    E900 will show you what you can expect for a camera with aggressive
    (and detail bluriing) noise reductions software,. Finally the SP-350
    which doesn't use any hardware or software tricks will give you its
    S/N straight up.

    5--Compare the four plots. There won't be any question that the S3IS
    is by far the noisest of the four camera.

    6--Where's the excess S3IS noise coming from?

    Drag another line across the darkest and top patch on the MacBeth
    target. Only this time do a cnrl-H to generate a histogram. The
    histograms of the three non IS cameras will look normal with all the
    values bunched together in a noisy bell curve. The S3IS will have up
    to a dozen randomly spaced and isolated spikes in adittion to the
    normal histogram. These spikes are cause by electical pickup--most
    likely from the IS piezoelectric motors.

    7--And of couse you can do similar test on any other cameras you're
    interested in. The review site has similar images for hundreds of
    other camrras they've reviewed over the last decade.

    jpc, Oct 20, 2006
  11. Specifically, angular rotation. All of the camera moves with the same
    angular velocity regardless of the distance moved by any one part.

    Actually, by holding the body and having a long lens, the long lens
    sticking out there acts as a stabelizer. So for two lenses with the same
    focal length, the longer one will tend to result in the least shake due
    to the the stabelizing effect. People tend to thing that a long lens
    causes shake because long lenses have a much longer focal length thus
    increasing sensitivity to shake. But the physical length of a longer
    lense actually reduces the degree of angular movement and angular
    Grant Robertson, Oct 20, 2006
  12. oops, I missed out a key word. "angular".

    [Off-topic, what other sorts of rotation are there?]

    It depends how it's supported - I found that long, heavy lenses were more
    difficult to hold simply because of the weight of the lens - the arms get
    tired after a while. Weight is one reason I am happier with my Panasonic
    FZ5, which is compact and weighs just 300g. Of course, I could carry a
    tripod, but that means more weight!

    With long lenses, I would never hold the body alone, as the twist on my
    wrists would be too great. I would have my left arm extended, supporting
    the lens.

    I think whether a longer or shorter lens was better when hand-holding
    probably depends which feels "most comfortable" for an extended period of

    David J Taylor, Oct 21, 2006
  13. Ockham's Razor

    prep Guest

    Depends on how you define `best'... If cost is a major factor, body. Only
    one needed. Ditto for a fixed lens, but that is probably RAA...

    You are building a servo system, and you have to tune it to get the
    results you want. If you want the BEST stabilisation for all lenses
    then the body system has to have fast enough response to handle small
    light lenses, and enough throw to deal with a very long lens. Doing
    that spells MONEY... Also it is way easier to use linear X,Y motion
    rather than rotary x,y that is a better match for real world lens

    The lens systems can be designed to work best with THAT lens, plus
    they can have their optical centre set to give the best result. But
    you have to pay more with every lens...

    Good, cheep, fast. Pick two.

    Paul Repacholi 1 Crescent Rd.,
    +61 (08) 9257-1001 Kalamunda.
    West Australia 6076
    comp.os.vms,- The Older, Grumpier Slashdot
    Raw, Cooked or Well-done, it's all half baked.
    EPIC, The Architecture of the future, always has been, always will be.
    prep, Oct 21, 2006
  14. Yeah, I was just being picky for future readers. But really, if they
    didn't get it from what you said then the only other way to explain it
    would have taken a page or two.
    Grant Robertson, Oct 21, 2006
  15. "IF" both systems were equally effective, it would make more sense (from
    Perhaps a minor point, but in-lens IS gives stabilized images to an
    (optical) viewfinder in addition to the sensor, so it's easier to take
    the picture in the first place.

    Dr. Joel M. Hoffman, Oct 21, 2006
  16. Ockham's Razor

    pixel_a_ted Guest

    If it's in the lens, then the image on the sensor is stable. Wouldn't
    this be an advantage for focus and exposure calculation by the camera?
    Not having experience with either method, I don't know how much of an
    advantage it would be compared to in-camera stabilization.
    pixel_a_ted, Oct 22, 2006
  17. Ockham's Razor

    Alfred Molon Guest

    That was physics, but don't forget biology. It is more difficult to hold
    a heavy object steady for some time, while it is less of a problem
    "freezing" a light object in mid-air.
    Alfred Molon, Oct 23, 2006
  18. I know all about biology. (OK, a lot.) While you may think it is easy to
    hold up the object, you aren't holding it as still as you think you are.
    A big part of biology in this context is perception, and perceptions are
    deceiving. It may be easier to hold the light camera up for a longer time
    without your hand shaking noticeably but the unnoticed shaking of the
    light camera will actually cause more detectable shake in the final
    photo. Besides, compact cameras are often held away from the face and
    SLRs are often held close. This biology definitely generates more shake
    for lighter cameras.

    Biology cannot trump Physics. If you want a test then just get a compact
    camera with manual settings. Set it to a slow shutter speed like 1/5
    second. Set the aperture as small as it will go to increase sharpness.
    Add lights to or darken the room to compensate as necessary to get a
    decent exposure. Take a picture holding the camera the way one would
    normally do so, with hands held out and the camera about a foot and a
    half from your face. Now get a brick. Using rubber-bands fasten the
    camera to that brick with the bottom of the camera against the brick.
    While holding the brick, press the shutter. You aren't allowed to hold
    the brick with one hand and jab at the shutter with the other hand. Press
    it carefully just as you would when taking a normal picture.

    Now look at the pictures. The one with the brick will have less shake.
    Even if you hold the brick for a while and get a little shaky feeling, it
    will probably be a better picture with the brick. Only if you really hold
    the brick for a long time and really get twitchy will it make the shot
    worse. But then, no one is going to hold a heavy camera up and out for
    that long.

    Now for another experiment: Fasten the camera to a short pipe about 3
    feet long and take a picture. That will reduce shake even more than the
    brick even though it is lighter. That is because rotational inertia is a
    factor of both the mass and the distance from the center of rotation. You
    will notice that the shake is only reduced in the direction that the pipe
    is able to resist rotation. The best arrangement would be a T shaped pipe
    which could resist rotation in all directions.

    So, what does all this tell you? The best shake reduction may very well
    come from a large rock laying on the ground rather than the camera
    itself. Just pick up the rock and hold it to the bottom of the camera
    while taking the picture. Or you could attach your tripod or monopod but
    not extend it. Just hold it up off the ground. The additional mass and
    the distance from the center of rotation will help a lot. Hold the camera
    and the tripod. Don't try to support the weight of the tripod with just
    the screw threads in the camera. You may look like an idiot but you will
    know that physics is on your side.

    If you have any doubts about any of this, go to your nearest archery shop
    and ask them to explain why people attach long stabilizers to their bows.
    Grant Robertson, Oct 23, 2006
  19. Ockham's Razor

    Ron Hunter Guest

    It depends on how well the compensation is done. Both in camera, and in
    lens systems work well, IF well implemented. OBviously, doing this in
    the camera has advantages, since one gets the benefit regardless of what
    lens in installed. It is also possible to do some compensation via the
    camera firmware, but this is much less effective. Doing it on the
    computer is even less effective. Given a choice, I would recommend the
    in camera method if purchasing a new camera.
    Ron Hunter, Oct 23, 2006
  20. Ockham's Razor

    Ron Hunter Guest

    And it is a way of adding the feature to an existing body....
    Ron Hunter, Oct 23, 2006
    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.