AF accuracy Z lenses vs F lenses on Z9

Due to the way the phase detect light path works via the mirror relay on DSLRs, lenses faster than f/4 tended to have issues with focus accuracy/consistency, particularly if there was a lot of spherical aberration in the lens. You can see this with a tool like Reikan Focal pretty clearly. The mirrorless cameras see the full context of the lens (and the lenses have less aberrations/coma), and have proven to be more consistent/accurate in lab testing. Look for Marianne's analysis of the DSLR phase detect technology on dpreview; it goes into detail about some of the DSLR issues.

Again, look at Marianne's comments about what portion of the lens image circle the DSLR system sees for fast lenses, and the issues with edge performance.

To answer your specific question, I've found the Z cameras to be more accurate with focus at f/1.4 with DSLR lenses than the DSLRs (don't have any f/1.2 F-mount lenses). Like JimK, I can vouch for the 58mm f/1.4 and 105mm f/1.4, which were two of my favorite DSLR lenses. They just perform more accurately on the Z's.

Are you using continuous focus? AF-C? Keeping the focus button down at all times with the focus target where you want it?

If not using continuous focus, with the very shallow depth of field of very fast lenses, all you have to do is take a breath and the distance from the focal plane to the subject will change. It's almost impossible to maintain the proper focus without using AF-C.

I do not have that issue with my F mount lenses on my Z9. I always use back button continuous focus.

Unfortunately no one answered my actual question - which was a comparison of equivalent Z and F mount lenses on Z9. I wasn't looking for a comparison with DSLRs since I do not use DSLRs anymore.

Of course Z mount does not quite have f1.4 lenses, and F mount doesn't have f1.2 AF lenses, so its not possible to do an exact comparison.

Most focus accuracy tests are for still subjects understandably.

I do use AF-C and I have been on occasion surprised that the image is slightly out of focus - even for still or slow moving subjects, usually people. For privacy reasons I cannot post examples here.

I have not seen a focus accuracy difference between F-mount and Z-mount lenses on a Z9. I have seen a focus speed difference between a Z9 with F-mount lenses and my D6 (the Z9 is generally faster).

Since you mention AF-C and people specifically, I believe you're also using Subject Detection. This becomes one of those "how was the AI trained?" questions very quickly, which is unanswerable. The focus data stream the algorithms are using is coming from the EVF stream on a Z9. Nikon hasn't given any technical information on how that actually works. I'd be curious whether you see a difference based upon the CSM #D20 setting.

I thought I answered your question. To summarize: I have seen no difference in accuracy between F and S lenses on Z cameras.

Yes, AF-C, usually 3D with subject/eye detection.
It is good to know you have not observed an AF accuracy penalty for using F-mount lenses.
I am going to more carefully analyze my images in the future - it is hard to form an opinion as there are many factors, and I have no point of comparison.

Maybe I misunderstood it, as I thought you didn't have a fast f1.2 Z lens, and you had compared at f2.8.

But its good to know that you haven't seen a difference.

You said:

The f/2.8 comparison was an example of two 70-200 zooms against each other. I have made comparisons at f/1.8, too. I have seen no differences relating to f-stop of the lens.

Perhaps you could look at the many comparison reviews on line including on YouTube OR be far more specific about what specific question you are precisely looking to be answered.

"What do you mean by more accurate"? ALL Z-mount and AF-S lenses are accurate on the Z9 if used properly, correctly adjusted (see below) and able to keep up with subject motion.

As a rule the latest high end Z-mount glass uses "better" AF-motors, newer coatings; provide more Z controls (buttons and rings) than the f-mount equivalents and appear to have addressed many focus breathing issues that were seen in the best AF-S glass. BUT where direct comparisons of image sharpness have been made the differences are marginal.

What also appears clear is the a Nikkor AF-S F-mount lens works very well on the Z9 -- but one MAY see benefits if one uses FoCal to optimise the AF-fine tuning (there are a growing number of us who have completed Focal tests and posted results for both Z and AF-S glass) -- in summary most Z-mount glass "needs" no fine tuning - but there have been some individual lenses which may be towards the edge of manufacturing tolerances and some have seen minor benefits from changing these settings for these lenses as well. Whereas there is a growing body of test results (so not just mine) that show AF-S glass can benefit from minor AF-fine tune adjustments when used on Z bodies (including the Z9).

Personally I love all the Z glass I own (which is why I sold all bar 1 F-mount lens last year) and I particularly like the built in TC in my two longest Z lenses - which vastly outperform the AF-S FL versions I owned before.
The Z50/1.2S is exceptional -- the AF-s 50/1.4 and 58/1.4 were both OK.
As is the Z100-400S (which is a VAST improvement over the AF-S 200-500 and 80-400).
My work horses for domestic (ie when I am forced to be at home and not in africa) work are the Z14-24/2.8S (which is vastly superior to the AF-S version); the Z24-70/2.8 (somewhat improved over the AF-S) AND the 70-200/2.8S (which is also exceptional).
The 400/4.5PF and 800/6.3PF are both exceptional -- but many like the fact the 500/PF and 300/PF are cheaper and smaller -- some folk are using the 500/PF with TC14 at 700mm with success.
The Z105/2.8MC - is exceptional; but then so are the 40/2, 28/2.8 and 24-120/4S -- to be honest all the f/1.8 and f/4 Z lenses appear to be very good and "far" better than their f-mount equivalents.

I wait for DxOMark and other systematic tresters to complete their testing to see the results of very specific tests for my comparison of lenses. AND yes I wait to see these tests completed on the cameras I use -- so that is the Z9.

Folk like @JimKasson also perform their own tests which have proven exceptionally useful to many of us in the past -- not least in the medium format "group" on DPR Forums. So do be a little more respectful when folk are simply trying to answer your question. Just be more specific in what it is you are seeking.

Here is an example of what I am trying to understand.

Eye detection was on, and the camera was saying it focused on the nearer eye, but actually focus is somewhat behind.

I have been investing in F mount lenses since I got into Z system. The only Z mount lens I have is the 28mm f2.8.
The F mount lenses when they are focused correctly are good enough for my purposes.

Your "I'm trying to understand is different to the original question you posed" -- but OK - you consider the Af leads to a "back focus" issue.

I recommend you Buy Recan FoCal and a target and complete their lens calibration test -- as noted in MY previous post some Af-S lenses benefit from AF-fine tuning -- and it looks like your setup may be one of these. These are your LENS and your BODY specific -- so just do it.

Yes there are other tools you can use to identify if your "system" is focusing towards the front or rear -- including a ruler at 45 degrees -- but FoCal is the BEST by far I have used.

MY 58/1.4 on my first Z9 benefited from -2.0 -- YOURS may be different. Go to the Settings menu and look for Af-fine tune and adjust the default for your lens to see when it is sharpest - use single point AF-S to start. AND obviously turn Af-fine tune ON.

It wasn't really - I was curious if using a Z lens of similar speed, I would get better focus accuracy.

I will have a look at that, thank you.

This is trickier than at first appears.
First, the "focus point used" indicator is not always accurate. It is NOT where focus was made, it's generally where focus was acquired, which can be different. The indicator does not seem to be updated after acquisition in most modes.
Second, you appear to have Subject Detection on while using 3D-tracking. There's considerable debate going on among those of us who test thoroughly as to whether that's the right approach or not. I'm now leaning towards, no, it isn't. 3D-tracking works most like DSLR 3D-tracking when Subject Detection is off. When Subject Detection is turned on, you get some interaction effects, including "seeing outside the box." These effects seemed to change with different versions of the firmware, as well.

Hi Thom,
If I am looking for eye focus, would using 3D tracking alone give the desired results?
The image above was taken in July last year, I cannot remember what firmware version was on. I tend to be up to date with firmware.