Testing for back focus

Discussion in '35mm Cameras' started by D-Mac, Mar 23, 2006.

  1. D-Mac

    D-Mac Guest

    In another group someone raised the spectre their camera may have an error
    in its auto focus.
    I slapped together this page to help make a few tests to see if its the lens
    or the camera responsible for it. Then the thought occurred to me that it
    might be useful for others struggling with soft pictures.
    D-Mac, Mar 23, 2006
    1. Advertisements

  2. G'day Doug,
    Just one little thing where I disagree with what you have said. Where
    you say "If the error is present in both auto focus and manual focus,
    the problem is in the lens. " - I disagree with that. Just how exactly
    can the lens be the source of a focus problem? The focus might not agree
    with the focus scale printed on the lens, but the lens should still be
    able to get correct focus no matter what, since the body is measuring
    focus (through electronics in AF, or the eye of the viewer in MF) and
    adjustments being made accordingly. If it is impossible to gain anything
    remotely close to focus then that is bad lens optics, but if it simply
    has the wrong focus point then it is a problem in the body. Please
    correct me if I'm wrong.
    I would think that focus errors in AF would indicate the AF sensors
    aren't at at identical distance from the lens as the film/sensor plane.
    IOW the sensor is either slightly out of position or the mirror is
    slightly out of position. Errors in MF otoh are caused by errors in
    mirror position or the viewscreen position (or a bad operator).
    Anyway, that minor quibble with your post aside, I have an old manual
    body that I suspect has a slight focus error. To use your method, am I
    correct to assume that you print a scale on the paper, set the paper at
    a 45 degree angle to the camera, and then focus on the centre point?
    Graham Fountain, Mar 23, 2006
    1. Advertisements

  3. D-Mac

    me Guest

    Good point- it all boils down to the point that some cameras define the
    focusing distance as the film plane to subject and others seem to expect
    the lens to subject distance to be the 'focus' distance. there can be
    a large difference. This has been the subject of debate on rangefinder

    me, Mar 23, 2006
  4. D-Mac

    Cheesehead Guest

    With film cameras (like the Pentax LX and others)
    the mirror bumpers would wear down and change the apparent focus

    I was in a local camera store last year and a pro, using 3
    Canon DSLRs, mentioned to the sales staff that all 3 bodies focused
    on different points. And he had tested them in fixed positions from a
    tripod, using one lens, so distance, lighting, and focus point was
    always identical.
    Apparently there is a problem with the precision of af sensor
    but there are other matters as well. Perhaps very minute differences
    sensor offset? It doesn't take much @ f3.5 or f2.8 to see a difference
    in DOF center.

    Cheesehead, Mar 23, 2006
  5. D-Mac

    D-Mac Guest

    That is how I use the test, Graham. My over simplification is often
    grounds for disagreement. It is not totally uncommon for lenses to be
    made with an element out of place. In this case the lens can most
    certainly be the focus problem.
    D-Mac, Mar 23, 2006
  6. D-Mac

    Scott W Guest

    This just does not make sense, if an element is out of place then a
    lens will have very bad resolution but a lens can not be the cause of
    back focus. Auto focus is a feed back system, the camera adjusting for
    best focus on the focus sensors. Manual focus counts on the focus
    screen being the same distance from the lens as the sensor. In either
    mode the lens does not enter into it.

    Scott W, Mar 23, 2006
  7. D-Mac

    D-Mac Guest

    Yes, I know Scott, It's hard to believe a lens could be responsible for
    not focusing properly but sometimes they are and this type of test can
    help determine if you send a bunch of lenses back for calibration with
    your camera or not.

    It wasn't that long ago when Professional Photographers did these tests
    before ever taking a picture in earnest. Today, for some reason I can't
    understand, people buy a camera and put 10 year old lenses on it, just
    expecting they'll get sharp and clear pictures and when they don't,
    scream bloody murder about sloppy quality control.

    I don't claim my test is scientific or electronic, just able to
    determine if the camera or lens is responsible for poor focus. You don't
    have to use it, you know?
    D-Mac, Mar 23, 2006
  8. hmmm...sorry but I'm gonna exercise my right to disagree on this one. A
    dud lens might not be sharp, but it can't cause forward or back focus.
    The camera/operator will adjust the lens until the image on the AF
    sensor/viewfinder is at it's sharpest focus for the subject distance. If
    the subsequent image on the film/ccd is focused for subjects at a
    different distance, then the fault can only be in the camera - the
    distance from the lens to the AF sensor/viewfinder is different to the
    distance from the lens to the film/ccd.
    If the lens is at fault, it may not be able to achieve sharp focus at
    all, or it may not have an even focal plane, but that will still show in
    the viewfinder if the viewfinder is at the right position.
    But... I have heard (but I don't know for sure) that the Canon AF system
    has a flaw that can cause focus problems due to faulty lens.
    Apparently, Canon cameras don't provide constant feedback to keep
    adjusting the lens until the image is in focus, but instead work out how
    much the image is out of focus, and just tell the lens to adjust by that
    amount. If the lens overshoots or undershoots (because of excessive play
    or stiffness), then the resulting focus will be out. By doing focus this
    way Canon can get faster AF than others (and this part of it sounds
    true, because their AF is quick), but the focus can be less accurate if
    not everything is calibrated perfectly. I'm not sure if this is the way
    canon do focus, because to me that sounds a stupid way of designing
    something, but the fact they tend to focus faster with less hunting than
    other brands could indicate that this is true. I heard this piece of
    information from someone who was very indoctrinated by Nikon so it could
    be BS - not sure. It would seem to make sense though with the number of
    times I've heard about Canon focus problems.
    Graham Fountain, Mar 24, 2006
  9. D-Mac

    Matt Clara Guest

    Scott made some excellent observations, and instead of responding to those,
    you simply chose to condescend to him.
    Nobody's taking your word for gospel, so if you aren't here to discuss your
    assertions, you might as well not be here at all.
    Matt Clara, Mar 24, 2006
  10. One thing worth mentioning is that most systems focus with the lens
    wide open. Focus shift when the lens is stopped down does seem to
    be a real possibility.
    Philip Homburg, Mar 24, 2006
  11. D-Mac

    Scott W Guest

    If there is any real focus shift stopped down the lens would have
    really bad resolution when wide open.

    Scott W, Mar 24, 2006
  12. Some things that "seem", can't.
    Floyd L. Davidson, Mar 24, 2006
  13. I don't know. Every stop you close the aperture halves the amount of light.
    So, The difference between stop x and stop x*1.4 is averaged with the
    effects of just stop x*1.4.

    To make things worse, I think most AF systems try to use the 'edges' of a
    cone of to focus. So you may be focusing with the worst part of the cone of
    light if you later stop down the lens.

    But I have to admit, I never saw any focus shift in my manual focus lenses.
    Philip Homburg, Mar 24, 2006
  14. D-Mac

    Scott W Guest

    If the outer part of the lens focuses at a different point then the
    center you have spherical aberration. If it is bad enough to make a
    noticeable shift in the focal point then the lens is going to be really
    bad when wide open.

    Scott W, Mar 24, 2006
  15. D-Mac

    Colin D Guest

    Yes, that is broadly correct. Read this about Canon's focus system:


    As you can see, the focus sensors tell the lens how far to shift. The
    lens then shifts the prescribed amount, and that's it. It does worry me
    somewhat that the camera is deemed to be in focus when really it is only
    somwhere within the dof, and not necessarily at the same place with
    repeated focusing. But, with FTM (Full-Time Manual focus) as embodied
    in ring-type USM lenses you can touch up the autofocus if you want.

    Note the piece about the focus sensors being bigger than the marked area
    in the VF. Strong contrast outside the marked area but within the
    actual focus chip 'vision' may do unexpected things, which is why
    shooting charts at a 45 degree angle is likely to be wrongly focused.
    To do it properly, I would mount the chart on a swivel arrangement,
    acquire focus with the chart square to the camera, disengage the AF,
    rotate the chart to 45 degrees and then shoot.

    But, I have to say, my 300D with the 17-85 USM IS gives me very few oof
    images. Mostly, the subjects have enough depth that exactly where the
    focus falls is not that important, since I am employing dof principles
    anyway. If I shoot people, specially individual people, I sometimes
    touch up the focus on their eyes, but more often than not, the camera is
    on the mark by itself.

    Colin D.
    Colin D, Mar 24, 2006
  16. D-Mac

    D-Mac Guest

    Is this your opinion again Matt?
    It seems to me that there are a bunch of jokers in this group - Scott w
    is one of them - who profess to know everything there is to know about
    everything photographic and don't in fact have much of a clue. The part
    which makes me cynical is when these jokers don't believe people because
    they themselves, don't even try to understand the problem, just don't
    understand the solution.

    It gets worse when, to see if these jokers actually bothered to look
    before they leap, I discover the faults this page helps identify are
    described on Google searches in more sites than I can count.

    One google search returns dozens of sites listed with a description of
    how Canon auto focus lenses actually focus. This is different to other
    makers. Scott never bothered to look and see if it was possible for a
    Canon lens to "look" focused whilst not actually being in focus before
    he started on me. Is your poke at me part of an endemic disease with
    Americans? Just can't take your finger off the trigger long enough to
    load some real ammunition?

    The camera does not focus, the lens does. The camera can have an error
    in the allowable tolerance between the lens flange and the sensor. This
    is the only time a Canon SLR camera can have a mechanical influence in
    focus. This is known as back focus.

    Canon DSLRs have one of the shortest flange to sensor distances in the
    Industry and as such are prone to needing critical re-adjustment before
    they are 'spot-on'. Very few Canon DSLRs are able to accurately focus in
    the middle of the DOF, direct from the factory. The use of lenses with
    f/2.8 as the largest aperture and programming to prefer auto and program
    exposures with smaller apertures overcomes this design issue.

    Canon AF lenses use a unique method of focus which is also prone to
    needing critical adjustment before it is 'spot-on'. Canon DSLRs
    therefore are fairly unique in that you can sometimes see a sharp image
    in the viewfinder but obtain a soft image because the lens can actually
    be out of focus while giving you a sharp viewfinder image.

    Couple the two possibilities for soft images from a Canon DSLR and you
    get a situation where you might "think" it's back focus error causing
    your soft images when in fact it could well be the lens causing it.

    I imagine even you could see the possibility for disappointment when
    sending your camera in for back focus adjustment and still having soft
    images when it came back because you mis-diagnosed the cause.

    Those fortunate individuals with a swag of lenses might pick up on the
    problem a little faster than most but this page was not intended for them.

    It might surprise even you that some of the older Canon AF lenses with a
    reputation for soft images on DSLRs, can actually be calibrated in
    harmony with the camera to fix the issue.

    My method of identifying which is causing the problem may not be perfect
    but it has stood the test of time and works a lot better than the
    "newspaper-on-the-wall" trick often bandied around the groups.

    Is yours and Scott's reaction why no one now bothers to post any
    information in this group anymore? If you ask me, your and his reaction
    of starting a rumble instead of looking to see it it had any substance
    or even relevance to your personally, is worse that the troll activity
    everyone attributes to the fall in posts here. Keep it up and pretty
    soon you'll be able to call this the Matt, Al and Scott group.
    D-Mac, Mar 24, 2006
  17. D-Mac

    Skip M Guest


    That's what I understood, too. And why some manual focus aficionados used
    to refer to Canon's AF as standing for "Approximate Focus."
    Skip M, Mar 25, 2006
  18. OK, so basically what I was told was right then. I'll agree with Doug
    then that the lens can be at fault. To me that sounds like an
    extraordinarily stupid way of doing AF. Yes it is quick, and it does
    seem to be mostly ok, but the system is just begging for errors to crop
    up. Perhaps this is just part of Canon's built in obsolescence and a way
    of making sure people keep buying new stuff. Pretty much every other AF
    system uses constant feedback, which does result in slower focus, but
    far more accurate, even if the lens has started to develop wear.
    Graham Fountain, Mar 25, 2006
  19. D-Mac

    D-Mac Guest

    This and a few other highly susceptible "features" in Canon cameras is
    why I have gone off in so many directions trying to find an alternative.
    The Nikon D2X is the nearest I've yet found to a well made and reliable
    camera which takes the type of photos I want.

    The Fuji S3 is also capable of really good capture but it has more
    "questionable" design features than Canon have. I might have put up with
    the rest of Canon's less than great features and embraced the 1 series
    if it had not been for the unreliability of Canon's auto focus and my
    ever decreasing eyesight which cannot focus manually with their setup.

    One of the most annoying things this AF system does is when you have
    (for example) shot a few pics inside a church and as the couple emerge
    into daylight and you need to use fill flash, the first shot or two
    doesn't focus at all.

    It is a fact of life for some of us that we rely on the camera we use to
    focus because for one reason or another, we cannot see sharpness in the
    viewfinder as we can with MF cameras. Sometimes we need to have both
    eyes open so as not to miss the action. When this happens, the potential
    for "dud shots" from an autofocus system prone to errors rises
    dramatically. Way too far for professional use.
    D-Mac, Mar 25, 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.