First, I do see some softness in your sample images, but it's only noticeable when the images are high magnified and it's hard to say why it's occuring.
Something I noticed in a couple of your images is flare at the bottom of the image frame. Are you using the lens hood? Is there a filter on the lens? Is it possible your camera has a light leak (failing light seals around the mirror, for example)? When working on a tripod and not with your eye to the viewfinder, are you covering the eyepiece? These things all might have some effect on auto focus precision, in some instances.
Other stuff... some of it repititious, I'm afraid:
1. You'd be better off starting your own, separate thread discussing 100-400mm lens focus, instead of continuing to tag onto a thread about the 400/5.6L. Other people who might be helpful may be missing your questions, because by adding them to someone else's post here, your concerns are ending up buried several pages into a thread about a çompletely different lens.
2. We don't know you or your skill level or how careful your techniques. If your images had the EXIF data attached, it would help because we could look at that and might see something causing the problems. Very often camera or lens "flaws" come down to somthing the photographer is doing. If you "save for the web" when you make your images, that will strip off EXIF data and is usually the reason it's missing. Instead, simply "save" the image. Some image hosting sites or the softwares used to upload to them also might strip off EXIF (but they also might offer a means of preventing this).
For example, the B29 image (my dad flew one of those)... what shutter speed? That's a difficult subject, going past you at 160 MPH or faster, as well as somewhat distant... The 5DII isn't the greatest camera at tracking moving subjects, but what focus mode were you in? AI Servo? One Shot? What focus points were you using? All points/auto select? Single point? If single, which one?
3. Really critical focus tests should be done on a tripod, with a good target set perfectly parallel to the camera sensor plane. The target can be as simple as a weathered fence, brick wall or newspaper page taped to a wall.
4. HAVE YOU TRIED MFA? With 5DII camera you can dial in some Micro Focus Adjustment to calibrate fine tune focus accuracy of this particular lens on your particular camera. To do so, the target should be about 25X to 50X the focal length being tested. So 100mm x 25 / 25 (to convert to inches) gives about 100 inches or roughly 8 feet minimum (but at least the lens' minimum focus distance away). If testing 400mm focal length, the minimum distance for testing needs to be about 32 feet.
When testing a zoom and setting MFA to calibrate it, 5DII's verions of MFA only allows for a single adjusment. This may be a problem with a zoom, if the lens is focusing significantly differently at different focal length settings. If it's only slightly different, just use an average. Or make an adjustment that favors the focal length you tend to use most.
Canon user manual for 5DII has more detail on how to calibrate your lens yourself, using MFA. This is a manual adjustment procedure. If you want, there are softwares available that can automate the process... one of these is Reikan FoCal (http://www.reikan.co.uk/focalweb/). You'll have to decide if it's worth the cost.
5. When doing focus tests on it's very important to eliminate all other possible variables as much as possible... to try to rule out anything else that can effect image sharpness, so that you are truly testing focus alone.
Use a remote release or the self-timer, so you aren't touching the camera at the moment the shutter is released.
In some shutter speeds (1/30 and slower), use Mirror Lockup so there is no possibility of internal camera vibrations. (When shooting, you can use Live VIew instead, for the same effect as MLU... However Live View uses a separate method of focusing, so would not show you the accuracy of the primary focus system, as seen in the viewfinder.)
If using the 100-400mm on a tripod, you MUST TURN OFF IS at the switch. (It is one of several Canon IS lenses that do not "self-detect" when on a tripod and turn IS off themselves. As a result, when there is no movement at all, IS in this particular lens (and the EF 28-135 IS, EF 300/4 IS and a couple other lenses) can CAUSE movement, which will appear as blur in images.
Good light, good subject contrast for the AF to lock onto are important, too. You don't want to use too high an iSO, since that will also cost some image resolution that might be mistaken for a lens problem
Set your 5DII to One Shot and only use the center AF point.
6. We've asked some questions which you don't appear to have answered and that have bearing on lens sharpness and focus. For example... DO YOU HAVE A FILTER ON THE 100-400mm? If so, remove it and redo your tests. This lens doesn't like filters! Even good filters on a 100-400mm seem to have some effect on image sharpness. I don't know why this occurs, but many 100-400L users have been very pleasantly surprised how good and sharp their lens suddenly became... after they removed a "protection" filter from it! Filters effect all lenses to some extent... maybe very little, maybe a lot under certain circumstances... but the 100-400mm in particular for some reason just seems to be very sensitive to adding any filter.
7. Almost all lenses are sharper stopped down a bit from their max aperture. HOWEVER, f22 you say you are using is a bit of overkill, and actually starts getting into an optical effect called "diffraction", that reduces fine detail and softens images. The full frame, 21MP 5DII is "diffraction limited" at f10 (uncropped 8x12 print of image). f11 it hardly shows up... f16 is generally okay, too... but much beyond that you start to see more fine detail lost to diffraction. The larger the image is viewed, the more noticeable it becomes (8x10/8x12 is generally used as a standard, to allow apples to apples comparison). Read more about diffraction here: http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm
Beware of being too overly critical. Looking at a 21MP image at 100% on most modern computer monitors is equivalent ot making a five foot wide print and then viewing it from only 18 inches away. OF COURSE it will look bad!
Looking at your images... yes they appear a bit soft. Backed off to the equivalent of a 13x19" or 16x24" print, they actually look pretty good. So the softness is only slight and it may be possible to improve with MFA.
If CAREFUL testing shows the lens consisitently focuses differently with different focal lengths and/or with closer or more distant subjects, which MFA cannot correct, especially if it's an older lens or was bought used, it may have some wear and tear, need to be sent in for service and calibration.
We're trying to help... and the sample images help use help you. But lacking EXIF data on the images and not seeing any answers to some of the questions we've sent back to you, we can only do so much.
***********
Alan Myers
San Jose, Calif., USA
"Walk softly and carry a big lens."
GEAR: 5DII, 7D(x2), 50D(x3), some other cameras, various lenses & accessories
FLICKR & EXPOSUREMANAGER
