New 20D needs lenses

Discussion in 'Digital SLR' started by Dale, Dec 26, 2004.

  1. I was dumb! Thank you.
    Well, that's true, but still
    element far away element close 1 element close 2
    / / /
    ,-'/ // /
    ,-' | - | |
    ,-' ( / ( ,-(
    --------{ ---{ ---{
    '-, ( \ ( '-(
    '-, | - | |
    '-,\ \\ \
    \ \ \
    element used to element used to element merely
    the corners the corners, moved closer
    path changed

    moderate acute very acute angle more obtuse angle
    angle at borders at corners of the than 'far away'
    element case

    Now, since you reflect more and more light as the angle gets
    more acute ... simply moving the element (and adjusting or not
    adjusting the light's path to that element) can easily change
    the amount of light passing through. Ok, it'll not change
    much in the central part or a lens ...

    How would they paint the inner walls if not by refracting
    light that should have passed through, or by assuming their
    lenses are closer/larger than they are?
    Even mirror lenses may have glass lenses behind the mirror.
    (and some may be filled with glass instead of air --- no, don't
    ask me why, I'd assume someone thought that the glass-silver-air
    border reflects better than air-silver. Which may be true only for
    rather accute angles.
    My guess is that the mirror, not being perfect, does not reflect
    100% of the light --- which leads to a somewhat darker image than
    you'd assume by the f-stop alone.

    Wolfgang Weisselberg, Jan 6, 2005
    1. Advertisements

  2. Dale

    John Francis Guest

    Let's take a simple view of things. The light entering
    the front element has a circular cross-section, with an
    effective diameter depending on the aperture selected.
    That light continues on down the lens barrel. Light
    destined for the central part of the image continues on
    straight down the lens barrel. But light that will end
    up in the corners of the frame is entering at an angle.
    For wide apertures not all of the light beam from the
    front element falls on the next element (and the further
    away from the front element that next element is moved
    by the zoom mechanism, the less light actually hits it).
    John Francis, Jan 6, 2005
    1. Advertisements

  3. Dale

    Ed Avis Guest

    The distance from sensor to lens is surely just the same as any other
    Canon EOS body... _except_ for the special EF-S lenses which have a
    different mount, moving the lens closer. But these lenses are
    presumably manufactured to tighter tolerances (or at least whatever's
    needed to make them work well with the 300D and 20D).
    Ed Avis, Jan 8, 2005
  4. Dale

    Ed Avis Guest

    Is it possible to directly compare sharpness of lenses at very
    different focal lengths - to say that this 500mm lens is sharper than
    some other 50mm lens? I understand that MTF charts can be produced
    for any lens and give a numeric comparison, but it doesn't seem that
    meaningful since you could not normally photograph the same object
    with the two very different lenses.

    If (for example) a given 300mm lens gets better MTF scores than a
    given 35mm lens, does that mean that in practice if you photographed
    the same subject with both lenses, moving much closer to the subject
    with the 35mm lens and a long distance away with the long lens, then
    the lens with better MTF score would usually give a sharper picture?
    Ed Avis, Jan 9, 2005
  5. That's correct. Producing an MTF for each lens is not the most
    difficult thing, but eliminating camera shake is much more of an issue
    with the 500mm. Also the perspective will be different if you attempt
    to frame the same subject in a similar fashion.
    Probably yes, depending on output size and assuming perfect camera
    technique, but again perspective will be different because of the
    changed shooting position. A sharper picture can mean more than the
    ability to visually resolve fine detail though. Higher contrast of
    spatial frequencies that are important for the impression of sharpness
    is also important.

    Bart van der Wolf, Jan 10, 2005
  6. Santa was very good and brought a 20D. Now the next challenge is what
    You probably need a very wide angle lens for the family shots. My
    suggestion is to get a good general purpose zoom (28-135 IS is great),
    and then a fixed wide angle for when you need it.

    Dr. Joel M. Hoffman, Jan 17, 2005
  7. 28-135, 17-40, 50 f1.8, 85 f1.8 , 400 f5.6 (or 100-400 which doesn't get
    Also, depending on how large you intend to print the final pictures,
    you may not need such a long lens. If you only want 4x6 or even 8x10,
    you can crop the digital image (effectively bringing you closer to the
    action) and still get a good shot with a shorter lens.

    Dr. Joel M. Hoffman, Jan 17, 2005
  8. 28-135, 17-40, 50 f1.8, 85 f1.8 , 400 f5.6 (or 100-400 which doesn't get
    Has anyone really looked at the quality of the digital sensor on the
    dSLR's to determine at what point buying better glass doesn't make any
    difference? The 28-135 IS is a great lens for what it does.

    Dr. Joel M. Hoffman, Jan 17, 2005
  9. I am waiting for a new 20D. I had used one for a week. I never tried
    Others here have shown that 1600 (and 3200) are digital manipulations
    of 800, so if you shoot raw, there's no point in setting the ISO
    higher than 800. (Still, 800 is pretty high, especially when you
    consider how grainy most 800 ISO film is.)

    Dr. Joel M. Hoffman, Jan 17, 2005
  10. Pardon my butting in with ignorant comments but I just realized that a
    Only a fast lens. The wider the aperture (that is, the small the
    number), the less DOF you get. Changing the ISO doesn't change the

    Dr. Joel M. Hoffman, Jan 17, 2005
  11. Excellent explanation. However, under limited lighting, the wide angle
    Without a tripod, you can get sharper shots under the same lighting
    conditions with wider lenses. To consider an extereme example, at
    f/30, a 300mm lens will give you nothing but blur hand-held, while a
    35mm lens will give most people a crystal clear shot.

    Dr. Joel M. Hoffman, Jan 17, 2005
  12. Excellent explanation. However, under limited lighting, the wide angle
    This is also generally true, but the consequences of this optical fact
    are that telephotos lenses tend to be slower, while wide-angle lenses
    tend to be faster. An f/3.5 300mm lens is expensive, while an f/3.5
    28mm lens is cheap.

    As for idential f-stops at different focal lengths, see my last post.

    Dr. Joel M. Hoffman, Jan 17, 2005
  13. Dale

    JPS Guest

    In message <%7SGd.690$>,
    Better glass is always better, unless your camera is lacking an
    anti-aliasing filter, like the Sigmas SDs and the Kodak 14MP cameras.

    Even if the camera's pixel spacing and AA filter play an important role
    in limiting resolution, a better lens will give better pixel-to-pixel
    contrast as well as truer overall contrast.
    JPS, Jan 17, 2005
  14. Dale

    JPS Guest

    In message <7cSGd.693$>,
    Not exactly. I said that the Canon 10D, 300D, and 20D, when set to
    "1600", are actually internally amplifying for 800, and doubling the RAW
    values upon output.

    "3200" on the 10D and 20D is digitally manipulated 1600-level
    amplification. So what you get is this:


    100 1x 1x
    200 2x 1x
    400 4x 1x
    800 8x 1x
    1600 8x 2x
    3200 16x 2x
    If you shoot 1600, you might as well shoot at 800 with -1 exposure
    compensation, because you get an extra stop of headroom (of course,
    flash compensation needs to be equally offset, if flash is used). On
    the 10D (and possibly, the 300D as well, but I haven't really tested any
    300D files), an additional benefit is that you avoid the ridiculous
    stripes of even numbers that have been offset by 1 to give some odd
    numbers in the data; perhaps to fool histograms. I have blurred dark
    frames at ISOs 1600 and 3200 from the 10D, and they show distinct zones
    of darker and lighter areas due to this striping when
    histogram-equalized. The 20D gives even RAW numbers throughout the
    image at ISOs 1600 and 3200.

    The amount of gain available at "ISO 3200" on these cameras can not be
    accomplished with the camera set to ISO 800, so "3200" is not as
    redundant and useless as "1600". In my experience, images shot at
    "3200" are cleaner looking than ISO 800 images under-exposed by 2 stops.

    IMO, Canon went too digital-minded in trying to create a smooth
    transition in image quality from one ISO to the next.

    Had they used pure amplifaction all the way through the range, image
    quality would have deteriorated very rapidly as you went to 1600 and
    3200, so they limit the deterioration by using some math, but they could
    have done 11x amplification for ISO 1600, with 1.45x scaling
    instructions in the RAW data for the converters, and we would probably
    have an ISO 1600 worth using. Or, they could do that for the casual
    user, and have a custom function to turn on literal amplification for
    more technical users who see through the numbers.
    JPS, Jan 17, 2005
  15. Dale

    JPS Guest

    In message <InSGd.696$>,
    That's only moderately wide, though. When you start going wider than
    that, the lenses start getting slower again, at a given price point.
    Most of the affordable fast lenses you see are in the 50 to 100 mm
    JPS, Jan 17, 2005
  16. Is this a roughly-right way of looking at it:

    The aperture of a 28mm f/1.0 lens will be 28mm.

    The aperture of a 300mm f/10 lens will be 30mm.

    Therefore, in some crude way, the amount of glass (and therefore, roughly,
    the cost) will be comparable.

    The aperture of a 300mm f/2.8 lens will be about 100mm and will therefore
    need pieces of glass at least three times the diameter of the two examples
    above. This would be _at least_ 9x the cost (the area is squared),
    probably more because the lens elements will be thicker, and probably need
    better machining.

    Graham Holden (g-holden AT dircon DOT co DOT uk)
    Graham Holden, Jan 18, 2005
  17. Dale

    Alan Browne Guest

    I won't research prices to see the correlation, but that is roughly right.
    Further, of course, is the volume effect. As the prices of these longer/faster
    lenses go up, there is less market, less volume so the price is pushed up even
    further. A Canon 1200mm f/5.6 is over $118,000. I believe you can negotiate a
    price break if you order 3 or more.
    Alan Browne, Jan 18, 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.