D5 auto AF fine-tune: User notes

[Leonard Shepherd]

What would be worth knowing is, have Nikon changed the length of the AF detection lines used in single point AF with the D5/5000? This is easy to check out.

I am speculating that the increased number of AF points in the D5/500, including none selectable, enable shorter single detection lines. If this turns out to be right then several AF test targets may need to be used at a lesser focal length magnification factor for best results compared to recent 51 point AF bodies.

It's unavoidable that the detection lines are a little shorter, as the AF point array has no space between points, and the number of points has increased since the D4.

This is what I anticipate but do not own a D5/500 yet.

In order to present the D5 AF points with the same detail seen by D4 AF points, the D5 must be moved further away from the target, i.e., at a greater focal length multiplication, not lesser.

It depends. It maybe that micro electronics has advanced sufficiently for Nikon to pack more individual AF detection points on a shorter detection line.

The first thing I do when I get a new type of body is to quickly check the length of the single point AF detection lines. If the length is changed a few test shots may be advisable as the new AF system may focus on some things the previous could not, and vice versa.

It is somewhat more complex today. Often more than one AF point is in use and the camera has sole control over AF points in between the user-selectable points. Nikon do not describe the detail of the algorithms which determine how the AF points are used, so it is left to the user community to discover them.

I do not own a D5/500 yet but it may be relatively easy to "discover" by selecting an outer line detector and comparing the inner and outer edge results.

--
Leonard Shepherd
You can buy kit. The rest is mainly down to you.
The more you practice, as with a musical instrument, the better you are likely to become.

 

[Marianne Oelund]

It depends. It maybe that micro electronics has advanced sufficiently for Nikon to pack more individual AF detection points on a shorter detection line.

I would expect a slight increase, in line with the linear resolution increase of the image sensor from D4 to D5. Also, the D5 fine-tune range is about 10% less than the D4.

 

[LynniePad]

It is important to understand that due to the fickle nature of phase-detect AF, the calculated fine-tune value may not be optimal. I recommend repeating the procedure at least half a dozen times (if not 10 or more) to establish the range of values that can result. Then, manually select a value in the center of the range. This means, of course, that the complete process is not so quick, and you may prefer to use manual AF fine tune methods instead.

I'd venture to suggest that the new technology is not worth bothering with if you have to repeat it so many times.

It's much quicker and more accurate to use what I call the "Direct Method" of taking shots at a range of AF-FT values and choosing the best.

Lately, I've been confirming good calibration using a special target that produces interference patterns in the OVF when focus is spot on.

A new method of AF-FT.

Comments?

 

[Leonard Shepherd]

the D5 fine-tune range is about 10% less than the D4.

Another difference (according to the instruction manual) is that several cross type sensors are not available with smaller than f4 combinations.

The manual clarifies the 200-400 f4 and 500 f4 (without a converter) and a few other f4 lenses cannot use some outer cross points

The manual seems unclear as to what cross type sensors might be available with the 200-500 f5.6.

 

[areallygrumpyoldsod]

Thanks I finally got mine to work on D500/400mm

 

[Leonard Shepherd]

The manual clarifies the 200-400 f4 and 500 f4 (without a converter) and a few other f4 lenses cannot use some outer cross points

Clarification - just to make it clear the outer cross type points seem to switch to single direction AF acquisition. The out AF points seem to work but not as the usually better cross type.

 

[Carl16033]

I recently purchased a D500 and went through the fine tuning process with an older Tamron pre-VC 17-50mm f2.8 lens. My results were all over the map from zero to +19. No negative number results, just zero and up. At first I wasn't sure what to think. (I'm still not 100% sure, but after reading this thread I think there is some issue with the way the Tamron lens determines good focus coupled with the info thaqt is given to the camera. Either that or the Tamron and D500 don't play well together.

My first thought, prior to the one mentioned above was that the Tamron was internally worn out (like a loose or worn focus gear that "thinks" it's in focus but overshoots or undershoots due to excessive wear, resulting in a wider margin of error in what the lens thinks is pin point focus.

I don't know.

After testing and re-testing the Tamron, I attached a newer Nikon 18-140mm kit lens that I don't use much and re-tested. Most tests were +1 or +2 with an occasional, sporadic 4 or 5. Zero, +1, and +2 were typical. Definitely not the wild swing as demonstrated with the Tamron.

It might be time to find a replacement for the Tamron.

 

[Marianne Oelund]

I recently purchased a D500 and went through the fine tuning process with an older Tamron pre-VC 17-50mm f2.8 lens. My results were all over the map from zero to +19. No negative number results, just zero and up.

Were you allowing the camera to re-focus each time, or did you set focus manually at the beginning, then leave AF off?

On the basis of your results, a FT value of +9 or +10 would be reasonable.

At first I wasn't sure what to think. (I'm still not 100% sure, but after reading this thread I think there is some issue with the way the Tamron lens determines good focus

The lens doesn't determine focus. It can't do anything except follow the camera's commands.

coupled with the info thaqt is given to the camera. Either that or the Tamron and D500 don't play well together.

My first thought, prior to the one mentioned above was that the Tamron was internally worn out (like a loose or worn focus gear that "thinks" it's in focus but overshoots or undershoots due to excessive wear, resulting in a wider margin of error in what the lens thinks is pin point focus.

If you are following the procedure correctly, all of those mechanical issues are taken out of the equation, at least for the purpose of finding the optimum AF fine-tuning.

 

[Carl16033]

I recently purchased a D500 and went through the fine tuning process with an older Tamron pre-VC 17-50mm f2.8 lens. My results were all over the map from zero to +19. No negative number results, just zero and up.

Were you allowing the camera to re-focus each time, or did you set focus manually at the beginning, then leave AF off?

On the basis of your results, a FT value of +9 or +10 would be reasonable.

At first I wasn't sure what to think. (I'm still not 100% sure, but after reading this thread I think there is some issue with the way the Tamron lens determines good focus

The lens doesn't determine focus. It can't do anything except follow the camera's commands.

coupled with the info thaqt is given to the camera. Either that or the Tamron and D500 don't play well together.

My first thought, prior to the one mentioned above was that the Tamron was internally worn out (like a loose or worn focus gear that "thinks" it's in focus but overshoots or undershoots due to excessive wear, resulting in a wider margin of error in what the lens thinks is pin point focus.

If you are following the procedure correctly, all of those mechanical issues are taken out of the equation, at least for the purpose of finding the optimum AF fine-tuning.

 

[BobWorrell]

I recently purchased a D500 and went through the fine tuning process with an older Tamron pre-VC 17-50mm f2.8 lens. My results were all over the map from zero to +19. No negative number results, just zero and up.

Were you allowing the camera to re-focus each time, or did you set focus manually at the beginning, then leave AF off?

On the basis of your results, a FT value of +9 or +10 would be reasonable.

At first I wasn't sure what to think. (I'm still not 100% sure, but after reading this thread I think there is some issue with the way the Tamron lens determines good focus

The lens doesn't determine focus. It can't do anything except follow the camera's commands.

coupled with the info thaqt is given to the camera. Either that or the Tamron and D500 don't play well together.

My first thought, prior to the one mentioned above was that the Tamron was internally worn out (like a loose or worn focus gear that "thinks" it's in focus but overshoots or undershoots due to excessive wear, resulting in a wider margin of error in what the lens thinks is pin point focus.

If you are following the procedure correctly, all of those mechanical issues are taken out of the equation, at least for the purpose of finding the optimum AF fine-tuning.

-
Source credit: Prov 2:6
- Marianne

Refocus each time using auto focus. I believe the lens has something to do with determining focus, otherwise every lens placed on the camera would either have wild results swings or tightly grouped results. I have one lens with a large swing and another tightly grouped

Your right the lens needs to communicate with the camera, the tamron may not be compatible with the D500.

Marianne knows what she is talking about before you dispute her advice, please do your homework and understand what your talking about.

Bob W

 

[Leonard Shepherd]

The manual seems unclear as to what cross type sensors might be available with the 200-500 f5.6.

The D500 guidance is that only the central 15 AF points work as cross type for "AF-S lenses with a maximum aperture slower than f4" and adds a note "At maximum zoom, in the case of most zoom lenses".

The 200-500 is f5.6 at all zoom settings.

I presume the F5 has the same limitation.

 

[Leonard Shepherd]

There are suggestions any high contrast target is OK - but maybe not






Results with 3 different high contrast detail subjects

With the D500 when I photograph part of this target with the central viewfinder AF point entirely occupying vertical lines on the right - I get nil fine tune

When I fill the central AF point with the horizontal lines I get a minus factor between 2 and 8 using the 200-400 and 200-500 at 400mm and the 300 PF and 300 f2.8 VR at 300mm

Why might "fine tune correction" change depending on the direction of the detail in the target?

My conclusion is the central viewfinder cross type AF point reads horizontal and vertical detail well, but Nikon caution live view may not read horizontal detail well (landscape framing), leading to a Liveview readout error and a spurious fine tune correction.

Depending on where I frame the viewfinder AF point within the circle on the left I get nil, or plus or minus.

Using the very fine detail as the target I usually get a "correction" in the opposite direction to horizontal compared to vertical lines. This may be because phase detect is often not good with fine detail and Liveview is generally much better.

I need to do more testing with different focal lengths and focus distances to be sure but it seems horizontal or very fine detail used as a target can cause a spurious result not present in many everyday subjects

The D5 Nikon suggested target uses the equivalent of 2 lines in the direction Liveview reads well.

--
Leonard Shepherd
You can buy kit. The rest is mainly down to you.
The more you practice, as with a musical instrument, the better you are likely to become.

 

[Fog City]

https://www.dpreview.com/forums/post/57358005

Horshack's AF Target

DPReview user Horshack produced an AF target consisting of variable-size crosshatch lines, arranged at random intervals. I printed this target at size 13x16 cm. For the first test, I used it at about 12x focal length, which produced good results (range 6 units), but the distribution of the values tended to be end-weighted, i.e., there were not very many results in the center of the range.

Moving back to a distance of 40x focal length, the results improved dramatically. I was rather amazed to find a distribution range of only 3 units. I have to give this target my highest recommendation, used at this distance.

--
Source credit: Prov 2:6
- Marianne

 

[Fog City]

Thanks for taking the time to post your results and methodology.

I don't have the D5, but the alternate AF tuning method I know, and that I have had good success with, is via Keith Cooper (Northlight Images), a Moire fringe technique

http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html

I wondered if you had any experience with it, in particular how its results and distribution compared.

 

[Marianne Oelund]

Refocus each time using auto focus. I believe the lens has something to do with determining focus, otherwise every lens placed on the camera would either have wild results swings or tightly grouped results. I have one lens with a large swing and another tightly grouped

You're causing that by re-focusing each time.

The auto AF fine-tune concept is to only calibrate the camera's internal AF optics and processing. It has absolutely nothing to do with the focus-motor drive or lens mechanics. To achieve this, the lens focus-group position must not be disturbed after the exact focus has been set. That initial precise focus is the reference that the system uses, and even if you use LV focus to achieve it, you should check that it's accurate.

You need to set focus very precisely before you start, then turn AF off at the lens or be very careful that you don't activate AF at all during the process (running the auto fine-tune doesn't activate AF - it only runs the internal AF calculations).

 

[Carl16033]

Marianne knows what she is talking about before you dispute her advice, please do your homework and understand what your talking about.

Bob W

I believe that is exactly what I am doing by engaging in a discussion on this forum. "Doing my homework" is why I am here asking questions. If someone makes a statement along the lines of "this is the way it is......", and I'm not clear on the concept or don't understand the statement because my thought process is going down a different path, than I believe I'm permitted to ask further questions. I would not describe my follow-up with her as "disputing her advice", but trying to grasp something that doesn't yet make complete sense to me. So thank you, but your comments have not added anything to my understanding. Please allow me the luxury of continuing to "do my homework" by asking more questions on this forum. Thank you in advance.

 

[Marianne Oelund]

There are suggestions any high contrast target is OK - but maybe not


Results with 3 different high contrast detail subjects

With the D500 when I photograph part of this target with the central viewfinder AF point entirely occupying vertical lines on the right - I get nil fine tune

When I fill the central AF point with the horizontal lines I get a minus factor between 2 and 8 using the 200-400 and 200-500 at 400mm and the 300 PF and 300 f2.8 VR at 300mm

That is an interesting disparity.

Why might "fine tune correction" change depending on the direction of the detail in the target?

The PDAF system uses different separator lenses, and different detection lines, for each direction. Because of that, it's certainly conceivable that it could focus differently.

My conclusion is the central viewfinder cross type AF point reads horizontal and vertical detail well, but Nikon caution live view may not read horizontal detail well (landscape framing), leading to a Liveview readout error and a spurious fine tune correction.

Try the procedure by setting initial focus in LV while the center LV point is on the vertical lines, then re-aim the camera to the horizontal lines and test. That way you can avoid using LV focus on the horizontal lines.



--
Source credit: Prov 2:6
- Marianne

 

[BobWorrell]

Marianne knows what she is talking about before you dispute her advice, please do your homework and understand what your talking about.

Bob W

I believe that is exactly what I am doing by engaging in a discussion on this forum. "Doing my homework" is why I am here asking questions. If someone makes a statement along the lines of "this is the way it is......", and I'm not clear on the concept or don't understand the statement because my thought process is going down a different path, than I believe I'm permitted to ask further questions. I would not describe my follow-up with her as "disputing her advice", but trying to grasp something that doesn't yet make complete sense to me. So thank you, but your comments have not added anything to my understanding. Please allow me the luxury of continuing to "do my homework" by asking more questions on this forum. Thank you in advance.

Sorry, the way you phrased the first part of your statement, I misunderstood that it may have been a question.

I apologise for misunderstanding your question.

Bob W

 

[Carl16033]

Refocus each time using auto focus. I believe the lens has something to do with determining focus, otherwise every lens placed on the camera would either have wild results swings or tightly grouped results. I have one lens with a large swing and another tightly grouped

You're causing that by re-focusing each time.

The auto AF fine-tune concept is to only calibrate the camera's internal AF optics and processing. It has absolutely nothing to do with the focus-motor drive or lens mechanics. To achieve this, the lens focus-group position must not be disturbed after the exact focus has been set. That initial precise focus is the reference that the system uses, and even if you use LV focus to achieve it, you should check that it's accurate.

You need to set focus very precisely before you start, then turn AF off at the lens or be very careful that you don't activate AF at all during the process (running the auto fine-tune doesn't activate AF - it only runs the internal AF calculations).

 

[Marianne Oelund]

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