Autofocus Confusion

Here's a simple test to see if your camera's autofocus coincides with its manual focus. (The other problems that you are describing have nothing to do with what some people claim to be experiencing with front or back focus).

Set the camera up on a tripod in front of a TV set. Manually focus the camera so that you see the interference patterns (moire) that forms between the TV screen and the fresnel lens in the camera's viewfinder. This is most easily seen in a TV image with lots of clear area, like a sky.

The camera is perfectly focused on the screen when the moire pattern is strongest. (You may have to move the camera a few inches one way or another to see the pattern clearly. Also, you will see the pattern over a moderate range of focus points but it is clearest when perfect focus has been achieved).

Now, turn on autofocus and half-press the shutter release so that the lens autofocuses.

If the moire pattern is identical between manual and autofocus setting then your autofocus is working perfectly. If not, you might want to have the camera checked.

Do this test with several different lenses and at several different lens to TV distances before drawing any conclusions though.

Michael

[Jim D]

Thanks for the suggestion Michael.  I will give these tests a try and report back for the benefit of others.  

Actually, Ray, a thorough discussion of this topic is probably way outside the scope of this forum.  From my engineering training, though, I can make a fairly good guess that this is a classic case of control system theory.  What I mean is that you have to consider the camera's computer algorithms and the lens used as a feedback control system.  When the sensor (phase detector) in the camera measures an error, the system moves the lens one way or the other to minimize the error.  When it sees a small error, it has to stop the lens movement.  In all likelyhood, it will overshoot and have to make a small reverse correction.  Maybe several before it converges to a point that the error is below the acceptable threshold.  I have noticed my lens make a large initial movement and then a couple small corrections before announcing lock.

I think the challenge to the camera designers is that the lens, which is an integral part of the system due to its mass and interitia, is different from model to model and between brands.  The nature of the motor in my 28-135 lens might be the same as in the 70-200, but the mass of the glass would certainly not be.  Thus, stopping the lens movement "on a dime" as they say, would be more challenging.  Imagine accelerating your car and applying the brakes at just the right moment and with just the right pressure so that your front tires stopped within six inches of a line that was 100 yards away every time.  

I have a suspicion that some of the front or rear focus observations (I will not call it a problem yet) are due to several factors and we may never come to a simple conclusion due to the multitude of variables.  For instance:

1.  Which side of focus the system starts from, or more precisely, ends on.  Since depth of field has a different rate of change on one side of focus than the other, then the error seen by the system would be larger on one side when the lens comes to rest causing a subsequent corrective movement back.  If the error is within tolerance, then the system is satistifed and WYSIWYG.  
2.  The multitude of lenses available.  My guess is that each lens calls up a set of constants stored in the camera memory that affects how the AF system algorithm responds with THAT lens.  If I were Canon, I would not waste too much time on testing third party lenses to determine their constants.  This may be why some people have reported greater problems with non-Canon lenses.
3.  The focal ratio of the open lens itself.  For a given physical overshoot on an f 2.8 lens, the error would be larger because the DOF is shallower and the system could try to further minimize the error.  Perhaps this is why I've heard it said that the AF systems do not work well with lenses above f 5.6.   The optimum point may be very broad for the error tolerance the camera is programmed to accept.

[Ray]

An interesting point was made in that article by Bob Atkins, or at least in the discussions that followed, that focussing accuracy is related to the speed of the lens. Professional type bodies are designed to focus within 1/3rd of the DoF of the maximum aperture whereas consumer type bodies tend to focus somewhere within the entire DoF of the aperture.

Whatever the actual fractions and however one defines DoF, it would seem that a faster lens will always tend to focus more reliably, all else being equal, since the Dof at F2 for example is considerably less than DoF at F3.5, whatever the quality of the body the lens is attached to.

[Jim D]

Not at all Pete.  The 28-135 just seemed like a good value for a first Canon lens for me and covers a very useful range with the crop factor.  

On the zooming issue, I am not sure you can blame it on the lens.  You can observe on manual focus that the focus is broad at the 28 end, so if you try manually focusing near the wide end, odds are that you will have to make an adjustment if you zoom to 135 (unless you use the distance scale and not your eye).  But the converse is not true.  Manually focusing at 135 and zooming back does not require even the slightest touchup of focus, at least not on my lens.  So you can conclude that IF you could precisely set the focus at 28mm, then it would be also perfect at 135mm.  I doubt that manually any of us could do more than hit it randomly once in a while.  The AF system, on the other hand, uses a sensor and some measured parameter like phase (whatever that means!) which is hopefully much better than our eye at detecting focus.  As in most electronic systems, the limit as to what you can achieve is only limited by how much money you throw at it.  It would not bother me if the pro cameras had more R&D, more sophisticated sensors, more computing horsepower, and tighter tolerances behind their AF systems.  As consumers, we each decide what level of performance we are willing to pay for.  

I tried this test several times on my 10d and 28-135:  Let the camera one-shot AF at the 28mm end on a good test target with the center sensor selected.  Then without touching the focus, zoom out to 135 without pressing the shutter again and evaluate the focus.  Mine was nearly perfect maybe 25% of the time.  The remainder only required the slightest movement of the focus ring.  I would not be upset at all if the same lens on a 1Ds got it perfect 90% of the time.  For the difference in price it should.

[chesty]

One thing I have always done is to select only one dedicated focus point.  The dead center of the screen.  I don't like the camera deciding which point to use, no matter how many of them there are.  (the more there are the more it will screw up in my book)

I have used this on my 1NRS, and now my 10D, I have not had any problems on focus with the 10D using Canon lenses.

I would find it hard to believe that Canon would release yet another camera with auto focus problems.

[Jim D]

Chesty:

I was not trying to imply that having different H/V sensitivity between the AF points is a problem.  It's just the way it is....the way Canon chose to design it.  If I am correct, I believe I read in another post that all the Canon cameras have some differences between the AF zones.  What would be unfortunate (in terms of design, not performance) is if Canon did not let you choose a preferred point and forced  AF to use the 7 point method exclusively.  They are smart enough not to do that in a camera of this class.  Let's give them credit for designing a brilliantly engineered camera and bringing it to the market at a reasonable price.  

Just to put a wrap on this topic, and the issues I mentioned in the first post, Canon and I decided that my unit probably needs a little tweak, so I have it and the lens in the service center now for evaluation.  So far, the experience has not soured my opinion of either the camera or the company because I understand the difficulty of bringing cutting edge technology to market in such a short timeframe.

[jiacone]

Well, I must say I recieved my Canon 10D around 3 weeks ago- my fisrt set of pictures which was taken on a 10 day tip to Italy was a little disappointing - a lot of my shots looked like they had a soft focus to them, but most of these were of buildngs and still shots. Since then I have taken the camera to my son's soccer game and roller hockey class - I am now convnced that I have a similar focus problem to the one being described in this thread. When putting the camera in servo mode and shooting action - there is about 60% of my shoots were out of focus.
 I guess it's going back to Canon