D5 auto AF fine-tune: User notes

[Carl16033]

I understand the need for establishing precise focus (step 5 in the Nikon manual) before activating the calibration routine (step 6 from page 312), and I now understand that the process does its calculations based on the exact focus (or lack of focus) that the camera sees at the point that the process is initiated. A half-press on the shutter button accomplishes step 5.

I'm not disturbing the focus after establishing it in step 5. When I indicate I am re-focusing, I'm going through the process described by Nikon, then at the final step (step 8), I'm starting over at step 1 and doing the calibration routine again and again (step 1 through 8, 1 through 8, 1 through 8, etc....). Since doing these steps multiple times follows the procedure prescribed by Nikon, doing so should be acceptable and should result in consistent results,

If I thought that Nikon had given a clear and complete procedure that is repeatable, I wouldn't have written this thread.

I'm not interested in diagnosing issues caused by Nikon's procedure. If you would care to use the procedure outlined in my post "Test Sequence" above, then post your results, I could be of some help to you.

--
Source credit: Prov 2:6
- Marianne

Thank you.

 

[kennebunk larry]

Hi Marianne,

Thank you, thank you, thank you!

You're suggestion of getting out of Live View and moving the focal point off-center and re-centering for solving the "Auto AF fine tune is not available at current focus settings" problem worked for me after much frustration.

Also, your later comment that one should not re-focus in the middle of collecting a set of AF fine-tune values. This solved my problem of getting a wide spread of values with which I could have no faith.

I do have a question. Today was my first attempt at fine tuning. Following all your suggestions, I started with my Nikon 200-500mm at 500mm fl on my D850. I first collected 8 values as follows : -6,-6,-6,-9,-6,-9,-5,-5. The average is -6.5. I then thought to start all over and do another set. This time I got: -7,-9,-8,-8,-11,-11,-10,-8. The average is -9. There was one difference in the set-up. The first set was collected at f5.6. The second at f8. Should the aperture make a difference? If so, is there a preferred aperture to use? Should it be the aperture you're most likely to be using in the field?

Thanks in advance for your response.

Larry

 

[Marianne Oelund]

Hi Marianne,

Thank you, thank you, thank you!

You're suggestion of getting out of Live View and moving the focal point off-center and re-centering for solving the "Auto AF fine tune is not available at current focus settings" problem worked for me after much frustration.

Also, your later comment that one should not re-focus in the middle of collecting a set of AF fine-tune values. This solved my problem of getting a wide spread of values with which I could have no faith.

I do have a question. Today was my first attempt at fine tuning. Following all your suggestions, I started with my Nikon 200-500mm at 500mm fl on my D850. I first collected 8 values as follows : -6,-6,-6,-9,-6,-9,-5,-5. The average is -6.5. I then thought to start all over and do another set. This time I got: -7,-9,-8,-8,-11,-11,-10,-8. The average is -9.

Your peak variation is only +/- 2, so I would say these are tight spreads. With the D5 and D500, I typically see spreads of +/- 5 or +/- 6. It seems the D850 system has been improved, which is good news.

There was one difference in the set-up. The first set was collected at f5.6. The second at f8. Should the aperture make a difference?

Typically, yes, due to focus shift. Your numbers indicate that the lens tends to focus further back when stopped down more, which is consistent with the normal sort of focus shift due to under-corrected SA.

If so, is there a preferred aperture to use? Should it be the aperture you're most likely to be using in the field?

Yes, but since you see a deviation of only 2.5 between your average figures, you would probably find a compromise setting such as -7 or -8 to be effective.

 

[kennebunk larry]

Hi Marianne,

Thank you, thank you, thank you!

You're suggestion of getting out of Live View and moving the focal point off-center and re-centering for solving the "Auto AF fine tune is not available at current focus settings" problem worked for me after much frustration.

Also, your later comment that one should not re-focus in the middle of collecting a set of AF fine-tune values. This solved my problem of getting a wide spread of values with which I could have no faith.

I do have a question. Today was my first attempt at fine tuning. Following all your suggestions, I started with my Nikon 200-500mm at 500mm fl on my D850. I first collected 8 values as follows : -6,-6,-6,-9,-6,-9,-5,-5. The average is -6.5. I then thought to start all over and do another set. This time I got: -7,-9,-8,-8,-11,-11,-10,-8. The average is -9.

Your peak variation is only +/- 2, so I would say these are tight spreads. With the D5 and D500, I typically see spreads of +/- 5 or +/- 6. It seems the D850 system has been improved, which is good news.

There was one difference in the set-up. The first set was collected at f5.6. The second at f8. Should the aperture make a difference?

Typically, yes, due to focus shift. Your numbers indicate that the lens tends to focus further back when stopped down more, which is consistent with the normal sort of focus shift due to under-corrected SA.

If so, is there a preferred aperture to use? Should it be the aperture you're most likely to be using in the field?

Yes, but since you see a deviation of only 2.5 between your average figures, you would probably find a compromise setting such as -7 or -8 to be effective.

 

[conger]

I have just tried the AF fine-tune on the D850 + Nik14-24 /2.8. I used a simple B&W drawing as target; I placed the camera apx. 80cm from the target; the ambient natural lighting was murky, but definitely well within the scope of the AF.

Up to now, the results I have been getting from this camera & lens were mediocre; let us face it, crispy focus and a wide angle lens are not a hugely important combination.

I found the numeric results inconclusive.

They depended on the light, the selected focal length, and apparently, even the amount I zoomed in in LV. I would like more time to go back and verify what I saw... but time does not allow this now.

Anyways, this entire process brings me back to a theme I raised on dprev some time ago... NIKON getting into the real world, and opening up a proper 3rd party applications interface for their cameras.

All of this hokus-pokus could be largely automated; calibration scenarios could be described and run using external instances (e.g. PC) to provide a proper focus value contour map for the camera body, together with each lens. This is standard practice for industrial environments, and for this issue it could be implemented so simply.

Instead, we still have to make guesses, estimates, and compromises.

As for the results from today; superficially, an improvement. Field testing will tell more.

 

[Leonard Shepherd]

I have just tried the AF fine-tune on the D850 + Nik14-24 /2.8. I used a simple B&W drawing as target; I placed the camera apx. 80cm from the target; the ambient natural lighting was murky, but definitely well within the scope of the AF.

Whether a "simple" drawing is a good or not so good contrast detect or phase detect AF target is as long as a piece of string, as explained in your D850 camera manual.

 

[conger]

Gosh Leonard, I had missed the explanation of the piece of string in the D850 manual!

And, that remark was as helpful as your response!

I mentioned the 'simple' drawing to indicate that I had used a planar image, without a surface that would offer low reflection phase diversity that may challenge the the focusing mechanism unnecessarily.

I also omitted to tell of whom the drawing was... thanks Leonard!

Now, what about some constructive comments about the subject under discussion. Surely, your vast knowledge base can present with something useful.

KYDIAV, .g.

 

[Leonard Shepherd]

I mentioned the 'simple' drawing to indicate that I had used a planar image, without a surface that would offer low reflection phase diversity that may challenge the the focusing mechanism unnecessarily.

I also omitted to tell of whom the drawing was... thanks Leonard!

Now, what about some constructive comments about the subject under discussion. Surely, your vast knowledge base can present with something useful.

My reply was intended to be very constructive.

I am sorry if you seem to give the impression you do not know enough about using AF, despite the explanations in your camera manual, that AF (either contrast detect or phase detect) varies between good, not good and not working at all - depending on what the AF point is aimed at.

A black and white drawing has a plus point in that AF reads in monochrome, needs contrast, and some colours that look very different as colour have very similar contrast when converted to monochrome.

Your camera manual tells you contrast detect AF needs detail parallel to the long dimension of the frame which a black and white drawing should include. If the LiveView AF is aimed at detail parallel to the long dimension of the frame it may not be accurate.

If the drawing contains parallel lines close together they can cause a contrast detect or phase detect AF error.

Turning to contrast detect AF, detail is too small (though big enough for contrast detect) or very fine detail under the AF point, as your camera manual confirms, can cause focus error.

What nobody else knows, because you have not posted the drawing, is whether it is suitable for fine tune and, if it is. whether you aimed the AF at a place on the drawing where both Liveview AF and contrast detect should work well.

Put another a black and white drawing (like some so called fine tune targets) can be good, poor or useless for lens/body calibration.

--
Leonard Shepherd
In lots of ways good photography is much more about how equipment is used rather than the equipment being used.

 

[Deci]

3. Press OK again to confirm test completion.

^I've been running through the fine tune process with a D850 and Sigma ART lense as described and it is going well - repeating a dozen times for different distances. One thing I noted is I don't always get a test completion note saying the data has been written. Sometimes the completion note comes up after starting the test, sometimes not.

Anyone else have this issue? I don't think it has any impact though, because saved values are being recorded if the completion message shows up or not.

 

[conger]

Hi Leonard... I was indeed right....your vast knowledge base can present with something useful.

I will keep on it.

Thanks..... -g-

 

[arbitrage]

I realize this is an older thread, but when reading through it today I was intrigued by the discussion about the "Default" value being a wider spread (ie +10 as Default is more correction than +10 as Lens Specific).

My issue is with my D500, 2xTCII and 500FL I need more than +20...+20 gets it close and in normal use sometimes it is in focus due to the shot to shot variance of normal shooting (I'm shooting small birds). I'm using +20 as a Lens Specific Value. Auto-AF Fine Tune usually fails with this combo which confirms it needs more than +20.

Therefore, my question is if I was to use a Default Value could that be a solution to get the combo into the correct range?

Following that...can you combine a Default Value for the camera with a Lens Specific Value or does it only use one or the other. My D850 has a trend to need negative Fine-tune values compared to my D500. This trend seemed to have "Saved" it with the above 2xTC/500Fl combo as it "only" needed +15. With my other lenses the D850 usually requires a negative value. So I'm wondering if I can tune in a base Default value on the D500 of say negative 5 or negative 10 and combine that with a + lens specific value for my troubled combo? Possible??

Thanks for any help you can provide.

 

[Marianne Oelund]

I realize this is an older thread, but when reading through it today I was intrigued by the discussion about the "Default" value being a wider spread (ie +10 as Default is more correction than +10 as Lens Specific).

My issue is with my D500, 2xTCII and 500FL I need more than +20...+20 gets it close and in normal use sometimes it is in focus due to the shot to shot variance of normal shooting (I'm shooting small birds). I'm using +20 as a Lens Specific Value. Auto-AF Fine Tune usually fails with this combo which confirms it needs more than +20.

Therefore, my question is if I was to use a Default Value could that be a solution to get the combo into the correct range?

Following that...can you combine a Default Value for the camera with a Lens Specific Value or does it only use one or the other. My D850 has a trend to need negative Fine-tune values compared to my D500. This trend seemed to have "Saved" it with the above 2xTC/500Fl combo as it "only" needed +15. With my other lenses the D850 usually requires a negative value. So I'm wondering if I can tune in a base Default value on the D500 of say negative 5 or negative 10 and combine that with a + lens specific value for my troubled combo? Possible??

Thanks for any help you can provide.

Sorry to say, you cannot combine them. As soon as you create an entry for a specific lens, the default setting no longer applies to it.

 

[arbitrage]

I realize this is an older thread, but when reading through it today I was intrigued by the discussion about the "Default" value being a wider spread (ie +10 as Default is more correction than +10 as Lens Specific).

My issue is with my D500, 2xTCII and 500FL I need more than +20...+20 gets it close and in normal use sometimes it is in focus due to the shot to shot variance of normal shooting (I'm shooting small birds). I'm using +20 as a Lens Specific Value. Auto-AF Fine Tune usually fails with this combo which confirms it needs more than +20.

Therefore, my question is if I was to use a Default Value could that be a solution to get the combo into the correct range?

Following that...can you combine a Default Value for the camera with a Lens Specific Value or does it only use one or the other. My D850 has a trend to need negative Fine-tune values compared to my D500. This trend seemed to have "Saved" it with the above 2xTC/500Fl combo as it "only" needed +15. With my other lenses the D850 usually requires a negative value. So I'm wondering if I can tune in a base Default value on the D500 of say negative 5 or negative 10 and combine that with a + lens specific value for my troubled combo? Possible??

Thanks for any help you can provide.

Sorry to say, you cannot combine them. As soon as you create an entry for a specific lens, the default setting no longer applies to it.

 

[xPhoenix]

3. Press OK again to confirm test completion.

^I've been running through the fine tune process with a D850 and Sigma ART lense as described and it is going well - repeating a dozen times for different distances. One thing I noted is I don't always get a test completion note saying the data has been written. Sometimes the completion note comes up after starting the test, sometimes not.

Anyone else have this issue? I don't think it has any impact though, because saved values are being recorded if the completion message shows up or not.

Yep, happens on my D850, too. It appears the values are being saved, so I guess it's no biggie, but still weird.

 

[Marianne Oelund]

Thanks Marianne,

I was told by someone that the Default value is a 2.5x greater value than the Lens specific value and that if no Lens specific value is saved then it will use the Default value. Is this still true for the newest cameras (D500 and D850)?

Yesterday I tried this to see if I could get my troublesome combo into spec. I deleted my current saved value for the 500f/4 and 2xTC so that it just showed a dash and no value. I then tried to go to +10 for Default which based on what I was told would have been equivalent to +25 or so as a Lens value. However, this didn't seem to work and a +10 for Default looked to give a very similar result as a +10 set for the lens. I had to go all the way up to +20 Default to get the combo on par with a +20 Lens value. And again +20 was still front-focusing slightly in most shots. Due to normal variance some shots were in focus where I wanted them but I can tell it still needs more + correction.

So for me I'm not seeing this 2.5x difference between a Default value and a Lens value. Am I doing anything wrong or has this difference been removed in the newer cameras?

Thanks for any insight you can provide.

You get the 2.5x range with faster lenses - at least f/2.8. With an f/4 lens, the ratio is around 1.5x, and with an f/5.6 lens or combination, the two ranges will be about the same.

 

[Van Griff]

Thanks Marianne,

I was told by someone that the Default value is a 2.5x greater value than the Lens specific value and that if no Lens specific value is saved then it will use the Default value. Is this still true for the newest cameras (D500 and D850)?

Yesterday I tried this to see if I could get my troublesome combo into spec. I deleted my current saved value for the 500f/4 and 2xTC so that it just showed a dash and no value. I then tried to go to +10 for Default which based on what I was told would have been equivalent to +25 or so as a Lens value. However, this didn't seem to work and a +10 for Default looked to give a very similar result as a +10 set for the lens. I had to go all the way up to +20 Default to get the combo on par with a +20 Lens value. And again +20 was still front-focusing slightly in most shots. Due to normal variance some shots were in focus where I wanted them but I can tell it still needs more + correction.

So for me I'm not seeing this 2.5x difference between a Default value and a Lens value. Am I doing anything wrong or has this difference been removed in the newer cameras?

Thanks for any insight you can provide.

You get the 2.5x range with faster lenses - at least f/2.8. With an f/4 lens, the ratio is around 1.5x, and with an f/5.6 lens or combination, the two ranges will be about the same.