20D background weirdness

Discussion in 'Canon EOS' started by kraig_cuddeford, Aug 29, 2005.

  1. I've seen this in a couple of my pictures, it ammounts to vertical and horizontal image patterning in the slightly out of focus bacground. I'm wondering if it's an artifact of the AA filter or zoom lens or both? has anyone shot the 100-400L with a film body and find the vertical, horizontal patterning or image doubling of the out of focus background?
    00DOFu-25417684.jpg
     
  2. Is this consistent (ie. has happened more then once?)
     
  3. When you shoot through a screen this is bound to happen.
     
  4. It's showing up consistantly in photo's with objects that have strong thin lines and edges in the out of focus background and not necesarrily vertical and horizontal only, I see some blades of grass in there at a bit of an angle with double images parallel to their angle. I've only noticed the effect in several photo's and I'm not shooting through a screen, are you talking about the AA filter?
    00DOIP-25419584.jpg
     
  5. The above example doesn't show as much, but it's there a little. Here is another example with stronger doubling. I'm finding it to look as distracting as bad rubber stamp cloning, if you know what I mean.
    00DOIz-25419784.jpg
     
  6. Are these JPEGs right out of the camera, or if not, what RAW converter produced them? Have you tried another RAW converter on the same files to see if the same thing happens? I've occasionally seen a similar texture in some of my 20D images, converted with ACR in Elements 3. I haven't identified a pattern as to when it does or doesn't show up (type of scene, ISO, lens used, etc.) but then again it hasn't showed up all that often, nor is it particularly noticeable even at 100%, so I haven't been trying to track it down.
     
  7. The AA filter has no microstructure in the XY plane (i.e. the plane of the sensor) so any
    blurring it performs is homogeneous in that plane. It has several parallel layers of exotic
    materials (like lithium niobate and IR reflective or absorptive material) is intended to shift
    some of the light ~1 pixel vertically or horizontally (one layer does each). It is designed to
    remove moire?and there's no physical way for it to create it. The doubling and tripling of
    fine details in the OOF region I see with telephoto lenses (just reviewed some and quite a
    few with the 100-400 have them) especially when they are way away from a relatively
    narrow zone of focus. It is probably the characteristics of the lens that contribute to
    "bokeh" that are at play here. The vertical and horizontal pattern I've never seen before
    except on oversharpened NTSC video.....Can't help with that

    Andy
     
  8. OK that makes sense, if the doubling is part of the bokeh of the telephoto zoom, I figure the pattern in the first shot is caused by continuation of the doubling from the strong lines in a fence's vertical boards. The fence is just outside of the frame in the photo but not out of the lenses view-- due to the 20D crop factor.

    Just one more question for ya Andy:

    Have you had the chance to review any non zoom telephoto's?
    If you have reviewed a non zoom telephoto, do you notice similar effect in the bokeh of the non zoom?
     
  9. Look at this:

    http://www.vanwalree.com/optics/bokeh.html
     
  10. Thanks for the link. figure 6 is definately what I'm seeing with this. That artifact seems unavoidable with the lens being a zoom. ahhhwell. I'm still very impressed and happy with my first L series..
     
  11. Kraig
    On the 300 2.8IS I looked hard but never found one as obvious as with the 100-400 - the
    background twigs, blades of grass were reasonably uniformly blurred, sometimes the blur
    was brighter at the edges of the blur, but it didn't break into two or three blurs!
    Andy
     

Share This Page