p.1 #1 · DotTune: New AF tune technique, no photos required
Forgot to post this on the Nikon forum here. It's a new AF tune technique that is fast, accurate, and doesn't require you to take any photographs or even to use AF.
My original dpreview thread describing the method is here (I'm "horshack" on dpreview). That thread includes the background behind the technique.
Procedure
1. Enter Live View and establish critical focus on a high-contrast subject. For cameras with poor LV clarity (D800/E), you may want to take a photograph and evaluate it on the computer to confirm critical focus. I recommend the focus chart here.
2. Leave the body and lens set to AF and configure your body for back-button focusing if not already configured so. This is necessary to disable AF engagement for shutter half-presses, which you'll be using in steps 3 and 4. To configure back-button focusing, set the "AF Activation" option to "AF-ON only". For the D4/D800 this is option a4, the D3/D3s/D700 it's option a5, and for D600 it's option f4 and D7000 option f5 (for D600/D7000 the AE-L/AF-L button will serve as the AF-ON button). Do not set the body or lens to MF as an alternative to back-button focusing; doing so will increase the confirmed focus range and make DotTune inaccurate.
3. Look through the viewfinder to see if the camera thinks the subject is in focus. You'll need to half-press the shutter while doing this to keep the metering/rangefinder from going to sleep. If you see the green arrow pointing to the left, then the camera+lens is back-focusing, so you need to decrease the value of AF tune (negative adjustment). If you see the green arrow pointing to the right, then the camera+lens is front-focusing, so you need to increase the value of AF tune (positive adjustment).
4. Repeat step 3 until you get a green-dot from the rangefinder.
5. Depending on focal length and subject distance there will be a range of AF tune values which produce green-dot confirmation. For example on my D800 w/50G f/1.4 @ 5' subject distance, I get the green dot from -6 to -14. Conceptually the optimal AF tune value should be the value that's the middle of the range, which in the above case is -10. So you'll want to establish the range of green-dot values by progressively increasing the AF tune values until the camera stops showing the green-dot (it'll alternate between the green dot and an arrow when it's at the margin of the range)...and repeat that procedure in the opposite direction to find the other end of the range. Use the midpoint of that range as your final AF tune value.
Notes
When establishing the precise range of green-dot AF tune values you may need to examine the rangefinder for several seconds. I've found that for the extreme margins of the AF tune range the camera may take a few seconds to flicker the arrow/green-dot.
The canon version of this technique is here, which can also be used on Nikons.
Edit 2/10/13: Decided to call this technique "DotTune", to have a quick and searchable moniker when referring to it. It's a bit more identifiable than just "new AF tune technique"
p.1 #3 · DotTune: New AF tune technique, no photos required
That would work well, if it weren't for the fact that there is considerable play in the green focus dot. But at least it will get you into the ballpark, if the lens is way off.
p.1 #6 · DotTune: New AF tune technique, no photos required
I guess for a non-photo adjustment purpose, I'd be more inclined to trust the live view focus and my own eye.
That is, allow the camera to autofocus, then switch the camera to manual, enter live view, and see what focus adjustment is needed to achieve critical focus. Change the values, allow the camera to autofocus, check live view... etc.
Of course, it's still going to be better to take an actual picture so you can look at it blown up on the monitor or feed it into software, but this is a good way to make adjustments in the field.
p.1 #11 · DotTune: New AF tune technique, no photos required
eSchwab wrote:
For those of you saying that the green focus dot isn't completely accurate, READ THROUGH AGAIN and try it. It works if you read carefully.
I did read it, but the logic "well, the dot is fuzzy, so the correct value must be in the middle of the range" isn't reasonable.
There's absolutely no reason to believe that the focus indicator is precisely half off in one direction, and half off in the other.
p.1 #12 · DotTune: New AF tune technique, no photos required
binary visions wrote:
I did read it, but the logic "well, the dot is fuzzy, so the correct value must be in the middle of the range" isn't reasonable.
There's absolutely no reason to believe that the focus indicator is precisely half off in one direction, and half off in the other.
I can humbly offer you only one reason to believe: it works. It arrives at the same AF tune value as other techniques including LensAlign and Focal. And it has worked for everyone who has tried it so far on my original dpreview thread. There's a lot of deep thinking that went into the technique, some of which is described in links on that thread.
p.1 #14 · DotTune: New AF tune technique, no photos required
snapsy wrote:
I can humbly offer you only one reason to believe: it works. It arrives at the same AF tune value as other techniques including LensAlign and Focal. And it has worked for everyone who has tried it so far on my original dpreview thread. There's a lot of deep thinking that went into the technique, some of which is described in links on that thread.
Well, the engineer in me says anecdotal evidence is not sufficient... but it's fair to say that it has worked for others so it's not worth dismissing.
I just have trouble with the assumption that the focus indicator - which is known to be slightly fuzzy with its indications - can't possibly have a bias in one direction. I mean, it's basically saying that the autofocus system is prone to manufacturing imperfections, but the focus indicator is not, and is always the same every time on every camera.
p.1 #15 · DotTune: New AF tune technique, no photos required
binary visions wrote:
Well, the engineer in me says anecdotal evidence is not sufficient... but it's fair to say that it has worked for others so it's not worth dismissing.
I just have trouble with the assumption that the focus indicator - which is known to be slightly fuzzy with its indications - can't possibly have a bias in one direction. I mean, it's basically saying that the autofocus system is prone to manufacturing imperfections, but the focus indicator is not, and is always the same every time on every camera.
I agree, anecdotal evidence is not sufficient; I offered it only to encourage you and others to approach it with an open mind and give it a try. The engineering proof behind the technique is outlined in links on the dpreview thread, which includes dispelling some preconceived notions about how AF tune works and why it's even needed.
As for the focus indicator and rangefinder, much of what people perceive as slop and fuzziness in it is actually due to the underlying phase detection system itself and how it interacts with different targets and lighting. Those same vagaries manifest in the performance of the AF system as a whole. The fact that the focus indicator and rangefinder accurately manifest those phase detection peculiarities is a discovery that has lots of exciting possibilities for reverse engineering the entire AF system.
p.1 #16 · DotTune: New AF tune technique, no photos required
I'm sure I'll give this a shot at some point; it's too convenient to not try it.
Regardless of whether it works or what my initial impressions are of it... I did not express appreciation for the research and writing that went into it - which was rude, so thank you for an interesting approach to this fine tuning (I have read your other thread extensively).
p.1 #19 · DotTune: New AF tune technique, no photos required
Worth a try. Just one note of caution. According to the Nikon user manual for the D800/E, not sure about the D600 and D4,
AF Lenses
Do not use AF lenses with the lens
focus mode switch set to M and
the camera focus-mode selector
set to AF. Failure to observe this
precaution could damage the
camera or lens.
I know that the OP didn't suggest doing this; however it's something to keep in mind just in case.
?