D500 auto af fine tune

[ArthurDent]

I’ve been reading Secrets of the Nikon Autofocus System by Steve Perry. He suggests that if one desires to fine tune autofocus for a lens, the auto af fine tune procedure should be performed 12 times. After each iteration, the value of the adjustment determined by the camera is recorded. The high and low values are eliminated from the data set and the remaining ten values are averaged to determine the final adjustment. He states that the values thus determined compare quite favorably with values determined by use of his Lens Align chart and software.

Has anyone tried this? If so, how did it work for you? Thanks for any comments.

[Frode]

I’ve been reading Secrets of the Nikon Autofocus System by Steve Perry. He suggests that if one desires to fine tune autofocus for a lens, the auto af fine tune procedure should be performed 12 times. After each iteration, the value of the adjustment determined by the camera is recorded. The high and low values are eliminated from the data set and the remaining ten values are averaged to determine the final adjustment. He states that the values thus determined compare quite favorably with values determined by use of his Lens Align chart and software.

Has anyone tried this? If so, how did it work for you? Thanks for any comments.

Tried the procedure mentioned above with D850 and AFS 85 1.8. Worked OK.

However, with the AFS 20 1.8 and AFS 400 2.8 VR I get different values every time I try.

Very satisfied with my D850, but very dissapointed with the AAFT. Maybe I had too high expectations?

[Ilkka Nissilä]

However, with the AFS 20 1.8 and AFS 400 2.8 VR I get different values every time I try.

What was your distance to target, approximately, for each lens? How bright was the lighting? Did you have a clean background behind the target?

How different were the values from the different attempts?

Thanks. I am trying to get more information about the circumstances of auto fine tune failures.

[ArthurDent]

Ilkka- Have you tried using the auto af fine tune? If so, how did you go about it and how did it work?
Thanks.

[Frode]

What was your distance to target, approximately, for each lens? How bright was the lighting? Did you have a clean background behind the target?

How different were the values from the different attempts?

Thanks. I am trying to get more information about the circumstances of auto fine tune failures.

Thank you, Ilkka.

Distance: «filled» the viewfinder with a full body portrait in vertical «mode» (1,88m high), turned the camera to horizontal position before AAFT.
Light: outside, daylight -stabil.
Background: grey (wall)
Target: white A4 paper with a vertical, black line, about 2 cm Wide.
Followed the procedure as explained in the manual.
Don’t remember the exact values, but varied with about 4- 10 after calculating the average (Steve Perry).
 
Same AFFT in D500 and D850...?

[Pistnbroke]

There can only be one value for the Fine focus adjust as all of us with long lenses know ..being out by 2 just wrecks the picture ..
Average value No NO .
My only other comment is that lenses run in like a car and need adjustment after about 500 shots and every 6 months

[Frode]

There can only be one value for the Fine focus adjust as all of us with long lenses know ..being out by 2 just wrecks the picture ..
Average value No NO .
My only other comment is that lenses run in like a car and need adjustment after about 500 shots and every 6 months

Yes, but when the combinstion gives different values almost every time one give it a try.....

[Pistnbroke]

Do it the old fashioned manual way ...length of  4x2  propped up against the fence at 30 deg and 60m  away , run a taped down it  and make a horizontal line or vertical target to focus on  and off you go .  If the picture is fuzzy due to a poor lens then stop down.

[Ilkka Nissilä]

Ilkka- Have you tried using the auto af fine tune? If so, how did you go about it and how did it work?
Thanks.

I have, with the D5 and D850.

For some curious reason the auto FT is off by 4-5 points with the 105:1.4 compared to what I need in real world photos with both cameras but once corrected, I've been happy.

With the 70-200/2.8 FL + 2X auto fine tune refused to work at all. Conventional means to fine tuning did work but the combination didn't focus approaching subjects well. 70-200FL + 1.4X was much better.

With the 58/1,4 it requires a distance of at least 2-3 m (don't remember exactly) to do auto fine tune. I guess at shorter distances the image has too much aberration but it sharpens up when a bit longer distance is set.

With the 300 PF I recently adjusted from 8 to 7 on the D850 as I noticed at distance the focus seemed a bit off.

All other prime lenses that I can remember auto fine tuned well. I do it in my studio with window lighting and gray seamless paper background using the lensalign target. I do maybe 4-5 repetitions typically and see if there is consistency. Usually the average has been fit to practical use. With most lenses I see about 1-2 points standard deviation (to my memory).

With zooms I test long, short and middle focal lengths and unfortunately get different results for each zoom setting so then I usually set the final value as average over the range.

I have generally not had to do long term changes to the setting but occasionally 1 point.

[Frode]

I’ve been reading Secrets of the Nikon Autofocus System by Steve Perry. He suggests that if one desires to fine tune autofocus for a lens, the auto af fine tune procedure should be performed 12 times. After each iteration, the value of the adjustment determined by the camera is recorded. The high and low values are eliminated from the data set and the remaining ten values are averaged to determine the final adjustment. He states that the values thus determined compare quite favorably with values determined by use of his Lens Align chart and software.

Has anyone tried this? If so, how did it work for you? Thanks for any comments.

If you decide to listen to Nikon, you might find some practical advice:

D850
https://nps.nikonimaging.com/technical_solutions/d850_tips/useful/auto_af_fine-tuning/

D500
https://nps.nikonimaging.com/technical_solutions/d500_tips/af/auto_af_fine-tuning/

D4s
https://nps.nikonimaging.com/technical_solutions/d4s_tips/af_fine-tuning/

Brad Hill:
http://www.naturalart.ca/voice/blog_2016_all.html#D5_AFtune

[antonoat]

Hi Guys,
In the olden days I read(many times) and was told that when focussed correctly on a target, due to optical laws or something, the depth of focus would extend(no matter how shallow) one third in front and two thirds behind the actual focus point. I wonder does auto fine focus adjust still try and replicate this characteristic or is focus split 50/50? Also when folks manually adjust focus fine tune are they sticking to this formula. If the answer to the above is negative would it matter or make much difference in the real world?  I imagine for landscapes or any scenario where extensive depth of field is desirable having the focus area correctly placed is important.
Thanks for any thoughts.
cheers tony.

[Bjørn Rørslett]

The 1/3, 2/3 "rule" is approximate for medium to far distances and entirely wrong for close-ups. Actually, once you get to 1:1 life-size and beyond, depth of field in front is greater than in the rear of the focused plane.

As the default diopter of the finder is -1 (1m distance), I think a 50:50 split is a valid assumption.

[Ilkka Nissilä]

I doubt that they use any assumption about depth of field. The goal of autofocus is to get the sharpest focus where the AF sensor is. To use depth of field in focusing I suppose you'd need to mark two points in the image between which the sharpness would be optimized.

[Erik Lund]

Nikon AF cameras always priorities to get the sharpest possible result and that is when the AF sensors see best peak sharpness.


Regardless of : Depth of Focus and Depth of Field.


In the software/firmware there are of course several rules build in for how the camera and lens software interpret the data and how the mechanical and electronics behave.


Also lenses have very different rendering in front of the focus plane and behind the focus plane so really tricky for the camera to guess where to compensate,,,


Better use your own vision and make an educated guess and take a look at the display.


Using Live view is also an option, but often I am shocked at how narrow Depth of Focus is for the high mega pixel cameras!

[antonoat]

Thanks guys, appreciate your thoughts very much.