Best at limiting distortion at widest end - 24-105 f4L or 24-70 f2.8L?

Discussion in 'Wedding and Event' started by steve george, May 11, 2007.

  1. I find the distortion at the 24mm end of my 24-105 a shot killer on a regular
    basis - sometimes I think it's not that far off what I used to get with the
    15mm on my Bessa L :)

    In all seriousness I find I have to be careful where I take it past 30mm in any
    situation where I want straight lines to be straight as the curve can be so bad
    I have church steeples apparently toppling on couples and pillars made out of
    jelly and bowing under the weight of the roof.

    Has anyone who's used both lenses noticed a significant improvement with the 24-
    70 f2.8L? This is on a 5D...

  2. There is, very much so.
  3. Thanks Ben, that's great. The f4 is a constant frustration in dark churches that forbid flash - I've compensated for this with fast primes, but lens swapping is a bind. Now I'm going to a two body arrangement it makes sense to ditch the 24-105 if I can fix two problems in one go...

    24-70 f2.8 and 70-200 f2.8 - going to have to get down the gym but worth it I think.
  4. Try PTLens. It costs $15 and corrects lens distortion in about 5 seconds. Works very well with my 24-105, and no cropping is required for correction of distortion at the wide end of that lens.
  5. I've found that primes tend to distort less, and keep more of the frame in focus. Have no
    idea why.
  6. Thanks for the tip Chris - I try to limit my post processing image correction though. I'd rather get it as right as possible in camera and spend PC time after doing enhancement rather than correction...and if a lens choice gets me closer to that then in the long run it makes sense to me :) There's also the faster f2.8 that I'm finding I need more and more over f4. I'm repeatedly finding myself at 1600 iso and f1.2 to f1.8 on primes for no-flash ceremonies in dark churches and rooms...f2.8 would bump up the ISO compared to the primes but give me the invaluable flexibility of a zoom in that kind of situation.

    If the 24-70 is better at minimising distortion than the 24-105 then it's a win-win. If it was worse then I'd stick with what I've got most likely.

    I agree primes control better - at a guess it's easier to optimise a lens for a single focal length than for multiple focal lengths. I love the 85mm f1.2 - it was the lens that first attracted me to Canon.
  7. To be honest, even my 17-40L is better at 24mm for both distortion and vignetting, I use it a lot. I also use PTLens, on every single image, it's built into all my actions, but PTLens has been optimised for distances of approx 20 yards, it doesn't 'fix' close up, certainly helps but doesn't fix.
  8. What do you mean by distortion. True distortion causes straight lines to curve inward (pincussion) or outward (barrel) from the center. One lens may have less distortion of this sort than another, however, the effect is usualy negligible on human subjects and scenery, just buildings or objects with obviously straight lines. So-called "fisheye" lenses, have exaggerated barrel distortion, to increase the field of view at the expense of causing straight lines to bow outwards.

    Wide angle lenses also exaggerate convergence, which is a function of distance and field of view (focal length) and not actually distortion (the effect goes away if you look at a print from a distance proportional to the focal length). Your choices are to move back and/or use a longer focal length and to keep the camera level in all planes (to avoid vertical convergence). Sometimes convergence will cause straight lines to look curved. This is an optical illusion - check it with a ruler sometime.

    Another type of "distortion" causes heads to become egg-shaped near the edges and (especially) corners of the frame. Besides convergence, this is the "distortion" normally associated with wide-angle lenses. It is is caused by the way a rectilinear lens renders 3-D objects in the focal plane, and increases with the angular field of view. It occurs when spherical objects (e.g., heads) are mapped onto the flat film. Like convergence, this effect disappears if you view the print from a distance proportional to the focal length. Your only option is to move further away, and/or choose a longer focal length.
  9. I totally agree with Edward and have always tried to stay parallel to the ground to avoid vertical convergence when using ultra wide lenses.
  10. I don't have the 24-70, but the 24-105 does pincushion like a bear. The 24-85 was actually better in that respect (and as sharp) as the 24-105. For a really great comparison of sharpness, distortion, and contrast. click on:
    Roll over the images to see the two lens tests. See how the bottom of the 24-105 curves upwards like a smile?
  12. I have both lens: will try to do some tests and post. I have noticed with my copies:

    1. The 24 on the 24-105 is a little wider than on the 24-70.

    2. Not exactly on topic, but the color of the 24-70 is noticeably warmer than the 24-105. Of the two, I think the 24-105 is more neutral.
  13. Was just going through a card of raws and see I have something already. img_2420 is the 24-105 and img_2422 is the 24-70. Both are at 24mm.
  14. and 24-70 at 24mm:
  15. Here's a couple of full res test shots. Both at 24mm, with manually set exposure set to 1/125 and f5.6, iso at 400. In both case, autofocus on the pic on top right corner of tv. IS was on on the 24-105 shot. I'm a little dismayed by the softness away from the center on the 24-105 shot. Anyway, first the is the 24-70:
  16. And the 24-105:
  17. If you get a couple of explorer windows (or Tabs in IE7), one for each image, and then navigate around each image by just clicking the scroll bars, the 24-70 is the clear winner for sharpness as soon as you get away from center. And for chromatic abberation in the upper left corner. And for light fall-off.

    The difference in color is pretty apparent also. Not sure which is more accurate, but the 24-70 is much warmer.

Share This Page