Nikon 4.1 Firmware - Initial Field Report

[ajrmd]

The bird is relatively small in the frame and I suspect that when it came in proximity to the more contrasty plant, it jumped to it. I'm a little more surprised that this occurred with 3d focus after it established acquisition.

 

[Dinusaur]

I'm a little more surprised that this occurred with 3d focus after it established acquisition.

Yes, that's my beef. The camera identified the subject (bird) in one frame and jumped to different subject (background) in the next frame and stuck with it, despite when the subject to track is set to new 'bird' mode. I have many more examples like this from this morning's shooting alone and this is how it was from day 1 in Z9.

 

[Hooligans Imagery]

Yes, that's my beef. The camera identified the subject (bird) in one frame and jumped to different subject (background) in the next frame and stuck with it, despite when the subject to track is set to new 'bird' mode. I have many more examples like this from this morning's shooting alone and this is how it was from day 1 in Z9.

Are these screen grabs from NX Studio? Can you check to see what it shows the focus mode on these two frames? That might help sort this out?

 

[John Navitsky]

he noted it was 3d

 

[Hooligans Imagery]

he noted it was 3d

Curious about the shape and size of the red box. I recalled it being smaller in 3d. Just checking.

 

[Steve]

Curious about the shape and size of the red box. I recalled it being smaller in 3d. Just checking.

Keep in mind that once subject detection takes the wheel, it can refine and redefine the size and shape of the AF area as it sees fit.

 

[EricKlees]

I used the new FW this morning in a fair heavy fog shooting Great and Snowy Egrets and a pair of Sandhill Cranes in tall reeds. It had no issues with the reeds or anything. It just grabbed the eye and never jumped anywhere. The Fog was heavy till 10am then it was a bit harsh light. I will be going again tomorrow with no fog forecasted.

There is no question it immediately snaps AF on the bird and doesn't let go with no need to manipulate the AF at all.

 

[Cameron T]

I didn't notice mine being jumpy to the foreground more than it was before, but I use wide area modes for the most part with CSP which can have that happen regardless. More testing is needed, but I need a good day out with lots of birds.

 

[ajrmd]

I used the new FW this morning in a fair heavy fog shooting Great and Snowy Egrets and a pair of Sandhill Cranes in tall reeds. It had no issues with the reeds or anything. It just grabbed the eye and never jumped anywhere. The Fog was heavy till 10am then it was a bit harsh light. I will be going again tomorrow with no fog forecasted.

There is no question it immediately snaps AF on the bird and doesn't let go with no need to manipulate the AF at all.

What AF modes and settings were you using?

 

[NatureBoy]

We took the z9s out this afternoon. Ones my wife’s with the same 500 PF lens. Great Egret hunting in marsh, strong light, before golden.
Im in 3D /Bird and shes in wide L / Bird and BOTH of us had the focus locking onto the body. It would not go to the head despite nice profile views. We both switched to animal instead of bird and the focus immediately picked up the eye And locked hard. There was some marsh grass in the shot but very minimal and not that high, at least to my eye.
I managed to find some Phoebe that it locked onto in DX When they were still. I was trying BIF but it wasnt working out to well but thats on me.
My wife meanwhile said WL/Bird was killing it on flying gulls. solid lock.
Started to loose the light a little but ended up with multiple killdeer in puddles. They were moving around in short running burst and the Bird SD worked good on profile shots. But when the birds were head on, the focus seemed to be on the chest bars.
Just my experience after less than 2 hours out, it doesn’t mean much.
For sure distance and light still matter when talking about auto focus / subject det and how it will Preform. Im looking forward to getting a lot more time in with it.

 

[NatureBoy]

What AF modes and settings were you using?

And Im wondering how close the birds were to you?

 

[Richard_M]

I updated the firmware and took the Z 9 with 800mm PF lens attached to a local gardens. The weather was mostly overcast, with the odd times where sun would shine through a break in the clouds. There was also a strong breeze.

I set the focus to 3D and was amazed at how quickly the camera focused on the eye. What I did like was if the lens was focused past where the subject was, the lens would refocus. Prior to this update I would get focus freeze more often then not when the lens was focused past the subject.

I was photographing another bird probably 12~15mtrs away from me when I heard a bird land to the right of me. The bird must've been very close to minimum focus distance. The new firmware focused on the eye. Prior to the firmware update I would've needed to refocus manually.


Little Wattlebird very close to minimum focus distance

Although I had the body set to bird subject detection, the camera had no problems focusing on the eye this Bandicoot


Southern Brown Bandicoot

I managed to photograph 10 species of birds on my small walk this morning. I also photographed a few flowers while I was at the gardens even though I had bird subject detection active. So far, I'm really pleased with the new firmware. I'll head to one of the local wetlands in the coming days to try some birds in flight, and hopefully some Kangaroo's and/or Wallabies to see if the new firmware works okay for them when set to birds.

 

[BarryRic]

Thanks for the video Steve. It provides a peak into the what additional AF capabilities will be added to the Z8.

 

[fcotterill]

If you zoom in you should see that part of bird's tail is covering the tree in focus - the bird is indeed in the foreground. The focus didn't recover in next 18 frames. During the bird's next pass I switched to Wide L, the same thing happened, focus was lost to a background tree as soon as the bird came in front of it but recovered quickly (after 5 frames). 3d seems to have hard time resetting focus compared to Wide L against bright and contrasty background (reddish tree leaves in the photos above). I'll keep your suggestion in mind next time I go out to see if 3d tracking improves.

My 0.02c is a tighter AF search area works better for subjects - typified by raptors - flying against/within clutter. Obviously Subject Recognition remains active.

I prefer the Custom Area modes, usually a thin rectangle that doesn't extend too much around the subject. Once this has grabbed the subject, a hand over to full 3D often works well. Depending on the subject's behaviour, not too fast/erratic, one can toggle between the AF modes when 3D loses the plot.

Nikon engineers originally developed C1 and C1 working with AFP (Agence France Press) in the D6; these Custom Area modes definitely rate as one of Nikon's better innovations, especially now in MILC deep learning Subject Recognition.

 

[JoelKlein]

Thank you @Steve for the Video. Nice rainy background.
The sand piper shows an amazing achievement with the box glued on the eye while the bird rapidly moves its head. I have never seen this before.

Question: Regarding the issue of jumping from eye to body or grass, What is the reason you didn’t change to a smaller AF box to exclude the grass?

 

[Joel N]

Tested a short bit. To me definitely a significant improvement in these area.
Initial acquisition - for “common” bird at least, it will grab focus without much hesitation from out of focus to focus.

Initial acquisition bif - same as above. It helps a great deal to reduce or eliminate the lag the hesitation for the initial grab.

Af Stickiness - for common birds, the af is glued to the head or eye even when subject is hoping all over the place. Much stickier if that is a thing.

Focus more tune to eye and head of subject - referring to higher chance of getting subject eye or head in focus even in flight.

One downside or probably note worthy aspect is that I feel af seems more aggressive and if there are two subject it will very often alternate between either. Can probably use specific af mode to counter this, so depending is might not necessary a downside.

 

[fcotterill]

More testing earlier today.... Here follow a few screen grabs from NX Studio to show the S-R of the updated AF in v4.10. Thanks to RSF [Hold] I toggled between Animal mode and this new Bird mode . The latter demonstrated a clear advantage on all these species. Each bird was very active, head bobbing up and down. The eye was often in and out of view.
Z9 v4.10, 800 PF using 3D Tracking only (I did not use any tighter AF Custom Area mode).

Southern Red Bishop male

Lesser Masked Weaver male

Helmeted Guineafowl - typcial of galliforms, the head is very rarely still. I focused on the body to see if the feather spots might break the AI, but the AF grabbed the eye consistently

Yellow-Backed Bishop - male. Eye recognition jumped around depending on orientation of his head

Nuptial Display - both birds are moving too and fro within the twigs as the male hovers. Bird AF mode demonstrated a surprising stickiness of the eye-recognition on the female - vacillating between the head and eye

....but then the AF grabbed on to the male

 

[EricBowles]

Perhaps it's differences in how good you're getting subjects in the frame vs what I have, but I've always found with the 1x1 or other area boxes that the subject detection will consider anything in the frame fair game. If I put the 1x1 on the rear end of a mammal, it will still focus on the eye if it sees it. This is why I have a button set to turn off subject detection- because sometimes I'll put that 1x1 on an animal's face but the camera sees a rock on the ground halfway across the frame as a subject and focus on it instead.

Pat of that is by design. With subject detection it uses priority - Body/Head/Eye - but if it does not find the subject, it looks for another high contrast AF target. That can be anywhere in the frame, but it's simply telling you it could not find the Subject and is looking for something else. Usually this is happening with unconventional head position, coloring, posture, or blocking from foliage or grasses but there are many reasons.

 

[Steve]

Thank you @Steve for the Video. Nice rainy background.
The sand piper shows an amazing achievement with the box glued on the eye while the bird rapidly moves its head. I have never seen this before.

Question: Regarding the issue of jumping from eye to body or grass, What is the reason you didn’t change to a smaller AF box to exclude the grass?

It doesn't work. That's normally how I'd approach a problem like that - limit the watch AF area and the watch area. However, with Bird Detection, it doesn't work that way anymore. I posted the specifics a page or so ago, but the short take is that if the AF area is anywhere on the bird, subject detection will go anywhere it likes. If it decides it only sees the body, it'll go for that rather than giving up and using the AF area normally.

Here's my original response:

#41

 

[JoelKlein]

It doesn't work. That's normally how I'd approach a problem like that - limit the watch AF area and the watch area. However, with Bird Detection, it doesn't work that way anymore. I posted the specifics a page or so ago, but the short take is that if the AF area is anywhere on the bird, subject detection will go anywhere it likes. If it decides it only sees the body, it'll go for that rather than giving up and using the AF area normally.

Here's my original response:

#41

Thank you for linking, I missed it.
I feel you should fwd this to Nikon. What’s the point of using a box if it’s ignoring it? Maybe their NEW philosophy is to use the box only for initial focus, then its free to use the entire frame. Just like the Canon EVF eye system.

 

[SCoombs]

Thank you for linking, I missed it.
I feel you should fwd this to Nikon. What’s the point of using a box if it’s ignoring it? Maybe their NEW philosophy is to use the box only for initial focus, then its free to use the entire frame. Just like the Canon EVF eye system.

I'll test it out when I get the chance, but I still say on my Z8 the way Steve describes it is how it's always worked. I've come to think of the AF boxes when subject detection is active as more like the user's way to give the camera "confirmation" that it should focus on a subject it's identified. I've found it works well if for instance, the camera had identified two subjects in the frame and you need to tell it which to focus on.

 

[John Navitsky]

yes, it's worked this way since one of the updates (3.0?). there was a change at one point about if it worked outside the box and that did change, but not "recently".

so Steve's comment suggests just a tweaking of "by how much".

 

[Patrick M]

One of the most interesting threads I’ve read in a while. I can’t wait for the Z8 to get the Bird.

 

[Hooligans Imagery]

It doesn't work. That's normally how I'd approach a problem like that - limit the watch AF area and the watch area. However, with Bird Detection, it doesn't work that way anymore. I posted the specifics a page or so ago, but the short take is that if the AF area is anywhere on the bird, subject detection will go anywhere it likes. If it decides it only sees the body, it'll go for that rather than giving up and using the AF area normally.

Here's my original response:

#41

But remember my decoy ducks in the pool. At least with 1x1 I think there is a limit on how far it will go. It's possible there are limits for each size of AF area but maybe that is overridden by other factors. However, for my decoy ducks, the eyes are pretty clear. Maybe I'll go try with WA-L if they will let me catch them.

 

[John Navitsky]

But remember my decoy ducks in the pool. At least with 1x1 I think there is a limit on how far it will go.

yes, exactly. it's only a ways outside the box.