LOL Chris! :^D

alphanikonrex said:
Well, I haven't seen any consistent problems, so I guess it's best I leave it be then.

How do these "weird things" work Chris? They sound very interesting! Are you able to adjust the lens without opening it up or anything like that?

Yeah they're great! They're always on me. I think they're called hands.

ChrisLange said:
I would advise not adjusting the lens correction unless you have a particular problem lens that is obviously underperforming with consistently similar errors.

Yea - and that's the key word here, consistent.
The new tool is a constant adjustment to fix consistent errors. There are all sorts of other AF errors possible, but they're all a result of the system (in camera) getting confused by a particularly challenging scene. (Random errors) This is a fix for a lens which doesn't move as far as it is told to.

Well, I haven't seen any consistent problems, so I guess it's best I leave it be then.

How do these "weird things" work Chris? They sound very interesting! Are you able to adjust the lens without opening it up or anything like that?

I would advise not adjusting the lens correction unless you have a particular problem lens that is obviously underperforming with consistently similar errors.

I have some great tools for adjusting my lenses. They're these weird things with five extensions which are, oddly enough, very intuitive to use. They work with every lens too, cool stuff, even manual focus ones!

soap said:
Errors /should/ be very uncommon. In the old days you'd get your system recalibrated if need be, now you can dial in a correction factor. That's cool x10 in my book.

Errors may be uncommon, but do you think it's worth a check? You know, just to be on the safe side?

alphanikonrex said:
Thanks for your responses. So, according to soap, a phase-detect is only done once, right? So if there's an error nothing can stop it from happening. That's annoying.

That's the beauty of it. It's FAST.
Errors /should/ be very uncommon. In the old days you'd get your system recalibrated if need be, now you can dial in a correction factor. That's cool x10 in my book.

alphanikonrex said:
But can't AF-C mess up as well?

It can - and it should, as it is focusing the same way, it just focus multiple times per unit of time. (Don't know the rate).
It /should/ keep splitting the error and get closer, but w/o a malfunctioning lens I can't prove this, and I haven't read any direct commentary on this situation from a trusted source.

ChrisLange said:
This is with AF-S on by the way, not AF-C, which is constantly updating focus readings.

Yea - absolutely. I assumed that was implicit, but you're right - I should have been explicit.

But can't AF-C mess up as well?

This is with AF-S on by the way, not AF-C, which is constantly updating focus readings.

Thanks for your responses. So, according to soap, a phase-detect is only done once, right? So if there's an error nothing can stop it from happening. That's annoying.

But in contrast AF everything should function perfectly, regardless of any lens-errors, based on my understanding. So it does make sense to check my lenses I guess.

When (Nikon SLRs) do phase split detection, the camera measures the difference and tells the lens (or the screw-drive motor in the case of non-AFS) how far to move. A second reading is not made.

Any consistent error in the interior design (or implementation) of the lens will therefore not be caught by the AF system.

Compared to when doing contrast detect AF:
Unlike phase split AF, contrast detect AF is unable to _measure_ how far out of focus the image is, and therefore must focus in and out a bit to see which way, and then continue in the correct direction until contrast decreases (overshoot) and then reverse direction until the peak is found again. This is why phase is so much faster - you can measure the focus error to a high degree of certainty and tell the lens exactly how far to go.

certain lenses don't necessarily always stop at the exact right focus point, you'll notice that the electronic rangefinder in the D300 will stay lit over a very limited range of focus, which, with long and or fast lenses, can cause a bit of error. My D70S nearly never reports correct focus for my 85/1.4 with the electronic rangefinder, I usually have to reshoot two or three pictures (if I have time, sometimes I just have to settle for a softer shot). The F4/F5/F6/D2/D3 series cameras have much better rangefinders inside, with a 3 icon indicator, a green dot in the center and 2 arrows on either side indicating which direction has been over focused. With manual focus lenses on a body without a split image or microprism focusing screen, this is a absolute necessity.

The AF correction will let you tell the CPU to lie to the lens, and focus the extra 10 micrometers for a certain lens to ascertain its best focus. If you care that much at least, 99% of the time it's completely fine.

I know that my D300 has a focus-correction thing that memorizes the proper correction that I enter for every lens.

Excuse my stupidity, but how can a lens's focus be off? Doesn't the camera just spin the lens focusing until the image appears to be in focus (whichever method it happens to be using, phase or contrast detection)? How can focusing be off if the camera it has nothing to do with the lens in particular?

I think I'm not getting something, that's all.