Why do *lenses* front/back focus?

[Class A]

Often one reads about users complaining that lens "X" front focuses and lens "Y" back focuses on the same camera body. In fact, some cameras, e.g., the Pentax K20D, support correction values to be input for several lenses respectively.

I can understand that a particular body produces FF (front focus) or BF (back focus). Misalignment of the sensor or AF unit can be the cause. But this error should be consistent across all lenses.

My understanding of the AF system of a camera is that it is a closed feedback loop and, hence, whatever differences may exist between lenses, the camera should just alter the focus until the sharpest setting is reached.

How can different lenses produce different FF/BF behaviour on the same body?

I have two speculations to offer:

1. The prism-based AF system is thrown off by individual lens colour casts (chromatic aberrations?).

2. Focusing doesn't work why a closed feedback loop only but there is an element of control without feedback and lenses differ in how they implement this feedback free control. This would explain why lenses contain chips that are not only responsible for metering issues and that re-chipping can apparently cure FF/BF issues.

It'll be great if someone could enlighten me what the technical cause for FF/BF lens behaviour is.

[Colin Southern]

Personally, I've never had an AF issue with any of my lenses, leading me - for the very reasons you explain so well - to suspect that it has more to do with the photographer than it does the lens.

What I have observed though is many many many images looking less than 100% because they haven't been sharpened correctly in post processing - personally, I think that it's one of the most widely and completely misunderstood post-processing steps.

Looking forward to reading what others have to say.

Cheers,

Colin - pbase.com/cjsouthern

[McQ]

I suspect that the vast majority of front focus and back focus complaints (on a per lens basis) are largely either (i) very subtle and/or subjective, (ii) the result of slight image quality variations between lenses which are mistaken for front and back focusing or (iii) the result of some sort of systematic focus bias caused by the test itself. This could mean that a telephoto lens might make a camera body's misfocus more apparent than with another wider lens, as one example. I would also think that any systematic focus bias should be almost entirely the result of camera body miscalibration as opposed to some sort of lens to lens variation.

[Raycer]

I had a lens that had focus problems. It was an 18-70 kit lens that came with my D70. It was sharp when I first got it. When it was approximately 2-3 yrs old, I notice some of the pictures had sharp backgrounds and blurry subjects. I assumed it was user error.

This became more frequent and when I checked the focus point on NX, it confirmed my focus point was on the subject. I then used a focus chart from
http://focustestchart.com/chart.html

and sure enough, all the pictures from this lens were back focused. I then tried the same thing with different combination of bodies and lenses and confirmed that the problem was with the lens.

I sent the lens to Nikon and got it back a week later. Did the same tests with different bodies and all the pictures were sharp on the focus spot.

On the list of repairs, Nikon showed "lens recalibration."

Here's my best guess of what happened. First I went to Nikon's website:

http://www.nikon.com/about/technolog.../software/caf/
In the web site above, Nikon says
Then, the autofocus system computes instantaneously how much the lens position should be moved and in which direction, and moves the lens accordingly.

From the website above, it seems there isn't a feedback loop for phase detection system. If the lens is out of calibration, is it possible that the camera thinks the lens has moved to the correct location while its not?

That's my 2 cents.

Cheers
Ray