Nikon, better subject detection, please!

[EricKlees]

f/4 vs. f/6.3 lens. Alternatively see the comment regarding predictive af algorithms. Either way, this did not occur on my Sony/Canon gear (with the exception of the R7) for these subjects. Interestingly, Canon was known for having poorer tracking for subjects moving directly towards the camera. That was something that they improved and the MILC R5/R3 were much better than the Canon DSLR's in that regard. Again, all of these AF systems have opportunities for improvement.

Are you certain? I have friends with A1's and R5/R3's that also complain about this issue at least a couple times every time we're out shooting at Conowingo Dam getting the Bald Eagles. One with the R3 that travels and sits with Mark Smith have up on the R3, have it to his wife and went back to the R5 due to at issues he no longer wanted to deal with.

There is no system that is perfect and the Z9 is better in some situations and the A1 is better in some. It really is how well you learn the system you're using. I've seen people just not adapt well to a system ava so wonderfully with another.

 

[John Navitsky]

Are you certain? I have friends with A1's and R5/R3's that also complain about this issue at least a could times every time we're out shooting at Conowingo Dam getting the Bald Eagles.

actually this is a great point. people’s expectations differ, and their evaluations differ

as someone who has become hyper aware and obsessive about critical focus, it’s also clear some people share that perspective snd some do not

this makes it hard to compare different people’s opinions. what might be acceptable to one person is not to another and vice versa

 

[John Navitsky]

The other thoughts I have is perhaps it is an error in the predictive af algorithms as the subject is moving at a relative speed during the soar, swoop, and dive followed by an acute slowing as it moves through the strike. All of the OOF images appear to be FF'ed so that's what leads me to believe that the predictive af algorithms may be involved.

i do think it's helpful to observe these types of direction and velocity changes and note how they interact with the af result.

we mostly think about the ML training in the context of subject detection, but it stands to reason they also feed into the af predictions. it stands to reason that they made a stand-alone bird af for a reason, and one reason might be because birds tend to have their own movement patterns

but generally speaking, i think there is a need for a LOT more ML training for all the brands. it's so hard to understand what all the things in the world are and how they move.

One thing I do need to explore is whether changing the a3 setting from steady to erratic would be helpful or not. It would be interesting to hear from those who shoot large mammals hunting prey whether they use steady or erratic.

for dogs, i'm using erratic, but tbh, it's very hard to understand exactly what that setting is, it's hard to observe the effects of that setting, thus it's hard to know if it's the right setting

 

[JAJohnson]

If I recall correctly, the Z9 used phase detect for AF and contact is only used for confirmation of the AF lock

Nikon Z cameras, in all autofocus modes other than Pinpoint, use a combination of contrast detection and phase detection to achieve focus. Pinpoint mode uses contrast detection only.

Your essential guide to Nikon autofocus

While post-production editing can sort out many issues, an out-of-focus image isn’t one of them. Sharp shots are essential for great photography, so here’s everything you need to know about autofocus on your Nikon DSLR or Z series camera.

nikonschool.co.uk

 

[EricKlees]

Nikon Z cameras, in all autofocus modes other than Pinpoint, use a combination of contrast detection and phase detection to achieve focus. Pinpoint mode uses contrast detection only.

Your essential guide to Nikon autofocus

While post-production editing can sort out many issues, an out-of-focus image isn’t one of them. Sharp shots are essential for great photography, so here’s everything you need to know about autofocus on your Nikon DSLR or Z series camera.

nikonschool.co.uk

Yes but I think Pinpoint is only available in AF-S not AF-C. Contrast Detection is for confirmation. It may only do that with AF modes that use Eye/SD

 

[Cameron T]

Yes, pinpoint is af-s only.

 

[EricKlees]

i do think it's helpful to observe these types of direction and velocity changes and note how they interact with the af result.

we mostly think about the ML training in the context of subject detection, but it stands to reason they also feed into the af predictions. it stands to reason that they made a stand-alone bird af for a reason, and one reason might be because birds tend to have their own movement patterns

but generally speaking, i think there is a need for a LOT more ML training for all the brands. it's so hard to understand what all the things in the world are and how they move.



for dogs, i'm using erratic, but tbh, it's very hard to understand exactly what that setting is, it's hard to observe the effects of that setting, thus it's hard to know if it's the right setting

Erratic setting for subject motion is for when the camera is still such as on a tripod and then the subject suddenly moves like a perched bird suddenly takes off. But if you are actively tracking a subject you are generally keeping the subject "Steady" in the frame so the better setting would be Steady, not Erratic.

I get far more consistent results in Steady shooting Terns, Kingfishers and Swallows than using Erratic. I photo foxes and fox kits running at me and playing but I don't shoot dog sports but I don't think the dogs are faster or more "Erratic" than the birds listed. Have you tried using Steady to see if you get more consistency?

 

[John Navitsky]

Erratic setting for subject motion is for when the camera is still such as on a tripod and then the subject suddenly moves like a perched bird suddenly takes off. But if you are actively tracking a subject you are generally keeping the subject "Steady" in the frame so the better setting would be Steady, not Erratic.

I get far more consistent results in Steady shooting Terns, Kingfishers and Swallows than using Erratic. I photo foxes and fox kits running at me and playing but I don't shoot dog sports but I don't think the dogs are faster or more "Erratic" than the birds listed. Have you tried using Steady to see if you get more consistency?

i'll have to give it a shot. one gotcha is the dogs tend to be running, then suddenly jump into the air, and i try to frame tight, so it can be a lot of chaos in the viewfinder.

 

[SoCalRick]

Nikon Z cameras, in all autofocus modes other than Pinpoint, use a combination of contrast detection and phase detection to achieve focus. Pinpoint mode uses contrast detection only.

Your essential guide to Nikon autofocus

While post-production editing can sort out many issues, an out-of-focus image isn’t one of them. Sharp shots are essential for great photography, so here’s everything you need to know about autofocus on your Nikon DSLR or Z series camera.

nikonschool.co.uk

I'm not so sure about that assertion ^^. Yes, the Z cameras do make use of two AF systems, but I think you have to parse the article very carefully to see where it is actually stating that contrast detection is being used on Mirrorless Z series cameras. I would submit that it is only used for Pinpoint AF and low light conditions when the #d10 (for Z8) or #d11 (for Z9) Starlight view setting is on.

Thom Hogan discusses the two AF systems in his Z9 Guide. Here is an excerpt:
"The phase detection autofocus system used by Nikon in the Z9 is generally accurate, but it is not always backed by a contrast verification step. Indeed, contrast verification is not performed most of the time; it’s only done in low light conditions or with specific focus settings (e.g. Pinpoint AF)."

 

[JAJohnson]

Yes but I think Pinpoint is only available in AF-S not AF-C. Contrast Detection is for confirmation. It may only do that with AF modes that use Eye/SD

I'm not so sure about that assertion ^^. Yes, the Z cameras do make use of two AF systems, but I think you have to parse the article very carefully to see where it is actually stating when contrast detection is being used on Mirrorless Z series cameras. I would submit that it is only used for Pinpoint AF and low light conditions when the #d10 (for Z8) or #d11 (for Z9) Starlight view setting is on.

Thom Hogan discusses the AF systems in his Z9 Guide. Here is an excerpt:
"The phase detection autofocus system used by Nikon in the Z9 is generally accurate, but it is not always backed by a contrast verification step. Indeed, contrast verification is not performed most of the time; it’s only done in low light conditions or with specific focus settings (e.g. Pinpoint AF)."

As explained to me by Ricci Chera (Nikon UK/Europe), the combination of Phase & Contrast Detection is used on all focus modes other than Pinpoint. Phase Detection is used first with Contrast Detection used to confirm (under any circumstance) when PD is unable to lock focus. Subject Detection is a separate but related process. He didn’t provide more detail than that, although I asked.

I think Ricci, Thom, and the Nikon AF explainer are in general agreement. The Nikon AF explainer is circa 2020 or 2021 (prior to the Z9 announcement) and needs a refresh. I’ve written to Nikon UK about that last fall but it’s not been updated.

There may be aspects of subject detection which impact AF lock but Nikon isn’t providing that information. I wish they’d share it with experts like Steve because he’d clarify it for us with examples.

With the importance of this technology for wildlife and sports photographers I’m certain Nikon has teams working on substantial improvements, perhaps made available with the next generation of Expeed processors.

 

[SoCalRick]

As explained to me by Ricci Chera (Nikon UK/Europe), the combination of Phase & Contrast Detection is used on all focus modes other than Pinpoint. Phase Detection is used first with Contrast Detection used to confirm when PD is unable to lock focus. Subject Detection is a separate but related process. He didn’t provide more detail than that, although I asked. I think Ricci, Thom, and the Nikon AF explainer are in general agreement. The Nikon AF explainer is circa 2020 or 2021 (prior to the Z9 announcement) and needs a refresh. I’ve written to Nikon UK about that last fall but it’s not been updated.

There may be aspects of subject detection which impact AF lock but Nikon isn’t providing that information. I wish they’d share it with experts like Steve because he’d clarify it for us with examples.

With the importance of this technology for wildlife and sports photographers I’m certain Nikon has teams working on substantial improvements, perhaps made available with the next generation of Expeed processors.

Hmm..., well I'm not going to argue with what Ricci has to say, as I consider him a trusted source of Nikon information. But I do find it odd that Thom characterizes it differently in his book.

I did read the linked article very carefully, and I don't think it contradicts what Thom wrote, but I also don't think it confirms the information that Ricci told you above. But again, I'm not going to argue with anything Nikon-related that Ricci has to say.

Nikon knows how it all plays together, but they may not be releasing enough of the pertinent details for us to put it all together.

 

[JAJohnson]

Hmm..., well I'm not going to argue with what Ricci has to say, as I consider him a trusted source of Nikon information. But I do find it odd that Thom characterizes it differently in his book.

I did read the linked article very carefully, and I don't think it contradicts what Thom wrote, but I also don't think it confirms the information that Ricci told you above. But again, I'm not going to argue with anything Nikon-related that Ricci has to say.

Nikon knows how it all plays together, but they may not be releasing enough of the pertinent details for us to put it all together.

I have Thom’s guides and I’ve heard him the describe Nikon Z AF on webinars more than once. I’m not seeing any substantive disconnect between Ricci and Thom, but if you do, that’s fine.

Also, they may be informed by different sources at Nikon HQ. It’s just impossible for us to know.

 

[fcotterill]

fyi, lame article compared to the exhaustive explanations by a prominent author of ebooks
And also compared against the detailed content shared in these technical SR/Handover-AF threads

Nikon Z Cameras: What to Do When Subject Detection Fails

photographylife.com

 

[arbitrage]

actually this is a great point. people’s expectations differ, and their evaluations differ

as someone who has become hyper aware and obsessive about critical focus, it’s also clear some people share that perspective snd some do not

this makes it hard to compare different people’s opinions. what might be acceptable to one person is not to another and vice versa

100%
There is no way I'd ever claim 90-95% success in focus for BIF with any system I've used (and yes I've used and owned them all from the big three brands)....now yes, I'd say 90-95% have focus on the bird (somewhere) and not on the background or foreground BUT I wouldn't call all of those a successful AF result. DOF is small with these big, fast lenses, DOF rarely cover the entire bird with the lenses I use and distances I shoot at. Therefore just the smallest AF shifts will take the eye out of the focus plane and those wouldn't count towards my final % of focused shots (perfectly focused/keeper shots).
I've also travelled with other shooters to places like Ecuador, CostaRica and Antarctica and realize that my scrutiny of a sharp shot is a lot more demanding than what others call sharp. I call it sharp at 100%, others seem to call it sharp when it looks good zoomed all the way back to full image.

 

[arbitrage]

Are you certain? I have friends with A1's and R5/R3's that also complain about this issue at least a couple times every time we're out shooting at Conowingo Dam getting the Bald Eagles. One with the R3 that travels and sits with Mark Smith have up on the R3, have it to his wife and went back to the R5 due to at issues he no longer wanted to deal with.

There is no system that is perfect and the Z9 is better in some situations and the A1 is better in some. It really is how well you learn the system you're using. I've seen people just not adapt well to a system ava so wonderfully with another.

I also have friends shooting A1 that complain about AF losing it to the water as osprey strike. I rarely have an issue with it on the A1.
Now I shoot mostly from kayak in calm water for my osprey dives and therefore I have good OOF backgrounds and not shooting down on the water so it is mostly blurred out. No issues there with the osprey strikes. When I shoot from shore (as most of my friends do), standing up, then I'm seeing more of the water which will always show the camera more distracting contrasty lines but even then I'm not losing the bird at the strike very often. I'm not sure why some of my friends constantly complain. They seem to be using the same AF modes as me....my guess is that I may be a bit better at keeping the bird centred as I pan down with the dive and at least with Sony A1 when not using the real-time tracking (which I don't for BIF) the camera puts a high priority on a moving subject centre of the frame. Therefore if you aren't panning well, the diving bird may move way off centre and the camera has a much higher risk of switching to the water. Also with Sony if you are using Real-time Tracking AF modes it has a much higher risk IMO of jumping to the water and not recovering as quickly.

 

[ajrmd]

I also have friends shooting A1 that complain about AF losing it to the water as osprey strike. I rarely have an issue with it on the A1.
Now I shoot mostly from kayak in calm water for my osprey dives and therefore I have good OOF backgrounds and not shooting down on the water so it is mostly blurred out. No issues there with the osprey strikes. When I shoot from shore (as most of my friends do), standing up, then I'm seeing more of the water which will always show the camera more distracting contrasty lines but even then I'm not losing the bird at the strike very often. I'm not sure why some of my friends constantly complain. They seem to be using the same AF modes as me....my guess is that I may be a bit better at keeping the bird centred as I pan down with the dive and at least with Sony A1 when not using the real-time tracking (which I don't for BIF) the camera puts a high priority on a moving subject centre of the frame. Therefore if you aren't panning well, the diving bird may move way off centre and the camera has a much higher risk of switching to the water. Also with Sony if you are using Real-time Tracking AF modes it has a much higher risk IMO of jumping to the water and not recovering as quickly.

As much as I would like to attribute it to technique or operator error, I just never experienced this with the R5/R3, A9II, etc. R7, yes but the other cameras no.

 

[EricKlees]

I also have friends shooting A1 that complain about AF losing it to the water as osprey strike. I rarely have an issue with it on the A1.
Now I shoot mostly from kayak in calm water for my osprey dives and therefore I have good OOF backgrounds and not shooting down on the water so it is mostly blurred out. No issues there with the osprey strikes. When I shoot from shore (as most of my friends do), standing up, then I'm seeing more of the water which will always show the camera more distracting contrasty lines but even then I'm not losing the bird at the strike very often. I'm not sure why some of my friends constantly complain. They seem to be using the same AF modes as me....my guess is that I may be a bit better at keeping the bird centred as I pan down with the dive and at least with Sony A1 when not using the real-time tracking (which I don't for BIF) the camera puts a high priority on a moving subject centre of the frame. Therefore if you aren't panning well, the diving bird may move way off centre and the camera has a much higher risk of switching to the water. Also with Sony if you are using Real-time Tracking AF modes it has a much higher risk IMO of jumping to the water and not recovering as quickly.

I completely agree the quality and skill of ones panning ability to keep the subject stable in the EVF makes a big difference.

 

[Wes Peterson]

Erratic setting for subject motion is for when the camera is still such as on a tripod and then the subject suddenly moves like a perched bird suddenly takes off. But if you are actively tracking a subject you are generally keeping the subject "Steady" in the frame so the better setting would be Steady, not Erratic.

I get far more consistent results in Steady shooting Terns, Kingfishers and Swallows than using Erratic. I photo foxes and fox kits running at me and playing but I don't shoot dog sports but I don't think the dogs are faster or more "Erratic" than the birds listed. Have you tried using Steady to see if you get more consistency?

You have that completely wrong. The steady/erratic setting is for when the subject DISTANCE is changing in an erratic or steady way. It has nothing to do with where the subject is in the frame. Its to allow the camera to continue to pull focus at the same rate it was before the subject was lost by the AF system. In the OPs case using the steady setting would tell the camera to continue pulling focus after the bird stopped suddenly which would almost always result in a front focus condition when the bird struck the water.

 

[EricKlees]

You have that completely wrong. The steady/erratic setting is for when the subject DISTANCE is changing in an erratic or steady way. It has nothing to do with where the subject is in the frame. Its to allow the camera to continue to pull focus at the same rate it was before the subject was lost by the AF system. In the OPs case using the steady setting would tell the camera to continue pulling focus after the bird stopped suddenly which would almost always result in a front focus condition when the bird struck the water.

No, I do not have it wrong. This is from a Nikon guy who's last project before retiring was the Z9. This EXACTLY how the Stready and Erratic are different. So I will take his knowledge on this over anyone else's.

Plus, I have tested Erratic with Terns, Kingfishers and swallows in flight and Steady has been far more consistent for everyone I know that uses the 2 settings as I have described

 

[Wes Peterson]

No, I do not have it wrong. This is from a Nikon guy who's last project before retiring was the Z9. This EXACTLY how the Stready and Erratic are different. So I will take his knowledge on this over anyone else's.

Plus, I have tested Erratic with Terns, Kingfishers and swallows in flight and Steady has been far more consistent for everyone I know that uses the 2 settings as I have described

The link is currently down for maintenance but all of this says youre wrong

You can only see EXIF info for this image if you are logged in.

 

[John Navitsky]

I think this all suggests it’s not adequately documented

 

[Wes Peterson]

I think this all suggests it’s not adequately documented

I think it suggests something else lol

 

[EricKlees]

The link is currently down for maintenance but all of this says youre wrong

View attachment 95785

I understand that. Though he said that is not the full story on the the 2 settings for subject motion. I will go with what works far more consistently for me and many others I know and what the Nikon guy has said with far more insight.

But that is also the Sports guide. It's really saying the same thing I said. Subjects steady in the frame (meaning if you are tracking you are keeping it steady in the EVF) but if the camera is still and subject is suddenly erratic, use erratic (like a perched bird suddenly taking off)

I also never pay attention to the manuals, I test everything myself. I have never read the manuals to learn. It's like looking at MTF charts compared to looking at real world images. Nothing else matters but the results

 

[Wes Peterson]

You can only see EXIF info for this image if you are logged in.

 

[Wes Peterson]

I understand that. Though he said that is not the full story on the the 2 settings for subject motion. I will go with what works far more consistently for me and many others I know and what the Nikon guy has said with far more insight.

I also never pay attention to the manuals, I test everything myself. I have never read the manuals to learn. It's like looking at MTF charts compared to looking at real world images. Nothing else matters but the results

Name you "Nikon guy" I'd like to have a chat with him.