Nikon Z8 losing focus on birds just prior to landing on water

[klfenton]

I have recently picked up a Z8 and the 180-600 lens for bird photography. One issue I've noticed is the bird subject tracking seems to do quite well maintaining focus when tracking birds coming in for a water landing or strike, but often seems to lose focus just prior to the bird contacting the water and then quickly reacquiring focus. I am using subject detection in bird mode, and have the focus tracking/persistence settings at the standards (3 out of 1-5, and steady rather than erratic).

I'm curious if others have run into this same issue, and if so have you found any adjustments to improve the hit rate on that moment just before birds are contacting water.

 

[Krümelkraft]

Water seems to distract the bird detection. I try to keep the AF area smaller and above water level when I shoot on water.

 

[Cameron T]

There's a topic (I don't remember which right now) where some people report this issue, others don't have it. What AF mode are you using?

 

[Tiny Birder]

Check out Steve’s videos on heat shimmer. That could be part of the problem.

 

[JAJohnson]

There was a thread on this topic two months ago. See below.

Z8/Z9 with diving birds

How does the Z8 or Z9 do with diving birds with the Bird AF? Specifically, Kingfishers or Terns, where you lose sight of the bird for a second then want to capture it exiting the water with a fish? I imagine Kingfishers would be the more challenging of those examples. Any special settings or...

bcgforums.com

There was another thread as well. If I can find it, I’ll post it here.

 

[DRwyoming]

I have recently picked up a Z8 and the 180-600 lens for bird photography. One issue I've noticed is the bird subject tracking seems to do quite well maintaining focus when tracking birds coming in for a water landing or strike, but often seems to lose focus just prior to the bird contacting the water and then quickly reacquiring focus. I am using subject detection in bird mode, and have the focus tracking/persistence settings at the standards (3 out of 1-5, and steady rather than erratic).

I'm curious if others have run into this same issue, and if so have you found any adjustments to improve the hit rate on that moment just before birds are contacting water.

Minimizing the size of the AF area can help quite a bit so personally I avoid full frame tracking modes like 3D and Auto when shooting diving birds and opt for smaller AF areas like Wide Small or Large or even better yet a custom C1/C2 area that doesn't cover as much of the frame.

Reflections off the water can trip up automated AF tracking as the reflection of the subject is physically closer to your position than the subject itself and a lot of modern AF tracking seems to have a closest subject priority. This can be seen with birds swimming on the surface but can happen with diving birds as well as the bird approaches the surface and throws a strong reflection off the water in still conditions.

Splashes, waves and other relatively high contrast textures in the water can also trip up AF tracking which is one reason to use as small an AF Area as you can get away with and still track the subject.

 

[ajrmd]

I have recently picked up a Z8 and the 180-600 lens for bird photography. One issue I've noticed is the bird subject tracking seems to do quite well maintaining focus when tracking birds coming in for a water landing or strike, but often seems to lose focus just prior to the bird contacting the water and then quickly reacquiring focus. I am using subject detection in bird mode, and have the focus tracking/persistence settings at the standards (3 out of 1-5, and steady rather than erratic).

I'm curious if others have run into this same issue, and if so have you found any adjustments to improve the hit rate on that moment just before birds are contacting water.

Yes, I've been one of those who have pointed out the challenges with the Z8's af system expressly in the cases of diving birds who "strike" tangentially (think osprey/eagles) against the water. It seems to work fine for vertically diving birds such as terns and pelicans though for those other subjects, no matter what AF mode/area used, it frequently loses focus just prior to the strike, misses the strike itself, and then reacquires the bird afterwards. I've tried every combination of AF area (AA, Wide S/L, Custom. 3D), SD on/off, changing the blocked response (3->4->5) without noticing any difference. The only thing I haven't experimented with much is changing the tracking from steady to erratic.

The irony in all of this is I think the issues lie in the predictive af algorithms rather than the AF system latching onto something else. The camera has no difficulty tracking the bird up to the point that it is about to strike and then it simply loses AF for a few frames. What leads me to believe that the challenges are with the focus algorithms is that the AF failures are reproducible and predictable. Additionally, the exif records the AF point clearly on the subject, the bird hasn't deviated in the frame, yet the actual af is typically in front of the bird for the few OOF frames. It seems that the AF system eventually recognizes the subject is OOF and it corrects a few frames later. Supposedly, the AF system is operating at 120 Hz vs. the 20 FPS that one is typically employing.

So, in some regard, this makes sense in that the subject is moving at some speed, "V1" during the dive and then it immediately brakes (or slows down) at some point with a velocity of "V2" just before impact. The camera AF detection/processing algorithms aren't fast enough to compensate, and it simply misses. For nearer strikes - the ones we would want to keep - the distance traveled over time, delta "t" would be greater than those occurring with a strike further away - thus that's why it appears so awful.

Most Z9 users don't report this problem so maybe the larger body has better algorithms or processing speed/power (in spite of what Nikon says or doesn't say). FWIW, I was in the Caribbean recently and the Z8's AF system was great for soaring birds including brown boobies, frigates, terns, etc. as well as shore birds such as sanderling, ruddies, yellow legs, etc. It had no difficulty acquiring any of these subjects in flight and tracked them quite accurately. My hit rate was nearly perfect with perhaps a couple oof frames in the thousands and thousands of images taken. Unfortunately, I did not have any opportunities to shoot any osprey strikes on this trip.

 

[klfenton]

Yes, I've been one of those who have pointed out the challenges with the Z8's af system expressly in the cases of diving birds who "strike" tangentially (think osprey/eagles) against the water. It seems to work fine for vertically diving birds such as terns and pelicans though for those other subjects, no matter what AF mode/area used, it frequently loses focus just prior to the strike, misses the strike itself, and then reacquires the bird afterwards. I've tried every combination of AF area (AA, Wide S/L, Custom. 3D), SD on/off, changing the blocked response (3->4->5) without noticing any difference. The only thing I haven't experimented with much is changing the tracking from steady to erratic.

The irony in all of this is I think the issues lie in the predictive af algorithms rather than the AF system latching onto something else. The camera has no difficulty tracking the bird up to the point that it is about to strike and then it simply loses AF for a few frames. What leads me to believe that the challenges are with the focus algorithms is that the AF failures are reproducible and predictable. Additionally, the exif records the AF point clearly on the subject, the bird hasn't deviated in the frame, yet the actual af is typically in front of the bird for the few OOF frames. It seems that the AF system eventually recognizes the subject is OOF and it corrects a few frames later. Supposedly, the AF system is operating at 120 Hz vs. the 20 FPS that one is typically employing.

So, in some regard, this makes sense in that the subject is moving at some speed, "V1" during the dive and then it immediately brakes (or slows down) at some point with a velocity of "V2" just before impact. The camera AF detection/processing algorithms aren't fast enough to compensate, and it simply misses. For nearer strikes - the ones we would want to keep - the distance traveled over time, delta "t" would be greater than those occurring with a strike further away - thus that's why it appears so awful.

Most Z9 users don't report this problem so maybe the larger body has better algorithms or processing speed/power (in spite of what Nikon says or doesn't say). FWIW, I was in the Caribbean recently and the Z8's AF system was great for soaring birds including brown boobies, frigates, terns, etc. as well as shore birds such as sanderling, ruddies, yellow legs, etc. It had no difficulty acquiring any of these subjects in flight and tracked them quite accurately. My hit rate was nearly perfect with perhaps a couple oof frames in the thousands and thousands of images taken. Unfortunately, I did not have any opportunities to shoot any osprey strikes on this trip.

Thanks, the scenario you describe is essentially the same one I've encountered. Maybe "erratic" would have some benefit given the suspected nature of the issue, so I may try that to see how behavior changes. Good to hear that you've explored quite a few other options without seeing much improvement on this issue, provides a good starting base of evidence on whatever experimentation I end up doing. One of the major challenges is the particular scenario (e.g., eagle plucking a fish out of the water) is not very frequent, so it's hard to explore a range of settings under consistent conditions. I'm also loath to compromise performance for more general settings in order to optimize for the rare chance of a strike, but that's the nature of the game I suppose.

 

[klfenton]

Water seems to distract the bird detection. I try to keep the AF area smaller and above water level when I shoot on water.

I have noticed that water is an appealing focus target, though for my specific case it's just as the bird is about to contact the water so even with a small focus area the water would be within the focus area.

 

[Steve]

The only thing I haven't experimented with much is changing the tracking from steady to erratic.

"Erratic" might help as @klfenton mentioned. Steady and Erratic are not exactly what they sound like - erratic is not for erratic movement but rather sudden stops and starts - and it sounds like that might be where the issue is. Just a thought.

 

[ajrmd]

"Erratic" might help as @klfenton mentioned. Steady and Erratic are not exactly what they sound like - erratic is not for erratic movement but rather sudden stops and starts - and it sounds like that might be where the issue is. Just a thought.

For "blast off's" I find that turning SD "off" tends to be a better strategy (or I guess one could employ the .jpg pre-capture) for higher yield rates. It's an approach that I use for KF's too where the dive from the perch is too quick for the camera AF system to react. Turning SD off might work for these types of tangential water strikes though the camera might get distracted by something closer in the frame. It's a technique which bears further exploration and when I have the opportunity to go back to Conowingo or Sebastian inlet these are the last two user adjustable parameters (turning SD off or switching to erratic) which have yet to be tested.

The tangential strikes against water (osprey, eagles) are the only two scenarios where I've encountered difficulty with the Z8 for BIF. The best explanation that I can offer up based on the reproducibility and evidence is that it is something related to the predictive AF algorithms. It's not the AF area, AF point stickiness, user technique, or other settings. Otherwise, as a camera, the Z8 performs exceedingly well for BIF. In fact, I was amazed at how well it captured and tracked a sanderling near the surf/rocks/other birds without flinching. Likewise, it followed a single Ruddy among a flock of Ruddies/Sanderlings without hiccupping (as these two sequences illustrate).

Of note, capturing the eagle/osprey strikes were not problematic for me with the Canon R3, R5 or Sony A9II. One of the weaknesses of the Canon AF system was manifest by birds flying towards the camera. Turning off the Canon eye recognition improved the hit rate in those scenarios.

 

[Steve]

For "blast off's" I find that turning SD "off" tends to be a better strategy (or I guess one could employ the .jpg pre-capture) for higher yield rates. It's an approach that I use for KF's too where the dive from the perch is too quick for the camera AF system to react. Turning SD off might work for these types of tangential water strikes though the camera might get distracted by something closer in the frame. It's a technique which bears further exploration and when I have the opportunity to go back to Conowingo or Sebastian inlet these are the last two user adjustable parameters (turning SD off or switching to erratic) which have yet to be tested.

The tangential strikes against water (osprey, eagles) are the only two scenarios where I've encountered difficulty with the Z8 for BIF. The best explanation that I can offer up based on the reproducibility and evidence is that it is something related to the predictive AF algorithms. It's not the AF area, AF point stickiness, user technique, or other settings. Otherwise, as a camera, the Z8 performs exceedingly well for BIF. In fact, I was amazed at how well it captured and tracked a sanderling near the surf/rocks/other birds without flinching. Likewise, it followed a single Ruddy among a flock of Ruddies/Sanderlings without hiccupping (as these two sequences illustrate).

Of note, capturing the eagle/osprey strikes were not problematic for me with the Canon R3, R5 or Sony A9II. One of the weaknesses of the Canon AF system was manifest by birds flying towards the camera. Turning off the Canon eye recognition improved the hit rate in those scenarios. View attachment 103262View attachment 103263

I agree, overall Nikon generally does a good job with most BIF

I suppose it makes sense that Subject Detection off works better with take offs - I too think the camera is faster to react with Subject Detection off. I still think setting to Erratic might help. I usually don't, but I'm not often shooting in those types of scenarios.

One thing I've noticed with Subject Detection is that it ignores tracking settings too. So, if it's on with those dives, it's really not doing anything anyway. So, Erratic might still work, but try it without Subject Detection. As you said before, it might be something best addressed with a FW update

 

[maxxus46]

As a former Sony A1 user, I can tell you the A1 has similar issues to the Z8/Z9 when birds approach the water line, I've seen it with all three cameras. This issue is not specific to the Z8 and requires special attention to not only environmental atmospherics but also water texture, glare etc. which will trip up ANY cameras AF system. You will have to find what works for you. For me, I often use Custom Wide (wide horizontal, I believe 1x5) which helps greatly. I also pre focus near where I expect the bird to land, and also "pump" the back button AF. This method seems to work well for me. YMMV

Yes, I've been one of those who have pointed out the challenges with the Z8's af system expressly in the cases of diving birds who "strike" tangentially (think osprey/eagles) against the water. It seems to work fine for vertically diving birds such as terns and pelicans though for those other subjects, no matter what AF mode/area used, it frequently loses focus just prior to the strike, misses the strike itself, and then reacquires the bird afterwards. I've tried every combination of AF area (AA, Wide S/L, Custom. 3D), SD on/off, changing the blocked response (3->4->5) without noticing any difference. The only thing I haven't experimented with much is changing the tracking from steady to erratic.

The irony in all of this is I think the issues lie in the predictive af algorithms rather than the AF system latching onto something else. The camera has no difficulty tracking the bird up to the point that it is about to strike and then it simply loses AF for a few frames. What leads me to believe that the challenges are with the focus algorithms is that the AF failures are reproducible and predictable. Additionally, the exif records the AF point clearly on the subject, the bird hasn't deviated in the frame, yet the actual af is typically in front of the bird for the few OOF frames. It seems that the AF system eventually recognizes the subject is OOF and it corrects a few frames later. Supposedly, the AF system is operating at 120 Hz vs. the 20 FPS that one is typically employing.

So, in some regard, this makes sense in that the subject is moving at some speed, "V1" during the dive and then it immediately brakes (or slows down) at some point with a velocity of "V2" just before impact. The camera AF detection/processing algorithms aren't fast enough to compensate, and it simply misses. For nearer strikes - the ones we would want to keep - the distance traveled over time, delta "t" would be greater than those occurring with a strike further away - thus that's why it appears so awful.

Most Z9 users don't report this problem so maybe the larger body has better algorithms or processing speed/power (in spite of what Nikon says or doesn't say). FWIW, I was in the Caribbean recently and the Z8's AF system was great for soaring birds including brown boobies, frigates, terns, etc. as well as shore birds such as sanderling, ruddies, yellow legs, etc. It had no difficulty acquiring any of these subjects in flight and tracked them quite accurately. My hit rate was nearly perfect with perhaps a couple oof frames in the thousands and thousands of images taken. Unfortunately, I did not have any opportunities to shoot any osprey strikes on this trip.

 

[ajrmd]

As a former Sony A1 user, I can tell you the A1 has similar issues to the Z8/Z9 when birds approach the water line, I've seen it with all three cameras. This issue is not specific to the Z8 and requires special attention to not only environmental atmospherics but also water texture, glare etc. which will trip up ANY cameras AF system. You will have to find what works for you. For me, I often use Custom Wide (wide horizontal, I believe 1x5) which helps greatly. I also pre focus near where I expect the bird to land, and also "pump" the back button AF. This method seems to work well for me. YMMV

I haven't used the A1 so unfortunately, I cannot comment on its abilities/challenges though I can say that the issues I am seeing with the Z8 are unique to it. It's not an issue of atmospherics, glare, distracting elements, etc. As I illustrated, the Z8 can detect and follow a small white sanderling against a busy and distracting background full of highlights. Also, it can track and capture dozens of frames accurately of an eagle/osprey in the dive, only to mysteriously and consistently lose focus a frame or two before the strike, capture ~6-8 oof frames through the strike (even though the exif demonstrates the af point still on the bird) and then recover focus a short time later. This doesn't occur with vertically diving birds which splash down such as pelicans/terns with similar settings so it's not as simple as an issue of distracting elements such as waves, splashes, specular highlights, etc. As I indicated, it seems that the loss of AF seems to occur as the bird is braking (slowing) and it never back focuses. That suggests it's related to AF prediction algorithms.

 

[maxxus46]

You state it's unique to the Z8, But then why do I see similar behavior with other bodies (such as I noted with my A1? and have also seen same behavior pattern with Canons R5). I shoot ospreys , eagles diving and have been getting more consistent results then what your experiencing. But then again, I have my own AF techniques that I've learned and work for me . Birds Approaching moving water at high speeds has always been challenging for any camera , so I think you will need to figure out a method that works for you, to help get more consistent results around moving water. Could the AF algorithms use improvement? No doubt they could, but same can be said for any current camera model from any brand. Every one of them suffers from different AF inconsistencies at this time. There is no perfect AF system or camera, it doesn't exist (yet)
Good luck

I haven't used the A1 so unfortunately, I cannot comment on its abilities/challenges though I can say that the issues I am seeing with the Z8 are unique to it. It's not an issue of atmospherics, glare, distracting elements, etc. As I illustrated, the Z8 can detect and follow a small white sanderling against a busy and distracting background full of highlights. Also, it can track and capture dozens of frames accurately of an eagle/osprey in the dive, only to mysteriously and consistently lose focus a frame or two before the strike, capture ~6-8 oof frames through the strike (even though the exif demonstrates the af point still on the bird) and then recover focus a short time later. This doesn't occur with vertically diving birds which splash down such as pelicans/terns with similar settings so it's not as simple as an issue of distracting elements such as waves, splashes, specular highlights, etc. As I indicated, it seems that the loss of AF seems to occur as the bird is braking (slowing) and it never back focuses. That suggests it's related to AF prediction algorithms.

 

[ajrmd]

You state it's unique to the Z8, But then why do I see similar behavior with other bodies (such as I noted with my A1? and have also seen same behavior pattern with Canons R5). I shoot ospreys , eagles diving and have been getting more consistent results then what your experiencing. But then again, I have my own AF techniques that I've learned and work for me . Birds Approaching moving water at high speeds has always been challenging for any camera , so I think you will need to figure out a method that works for you, to help get more consistent results around moving water. Could the AF algorithms use improvement? No doubt they could, but same can be said for any current camera model from any brand. Every one of them suffers from different AF inconsistencies at this time. There is no perfect AF system or camera, it doesn't exist (yet)
Good luck

Interesting that you experienced this with the R5 - I never did. Nor with the R3. If either of these cameras locked on and started tracking accurately, they wouldn't lose focus immediately preceding and through the strike. The R7, yes but that was a different beast entirely. In my experience the Z8 generally performs well around water and has, as I've illustrated with a few examples, little difficulty capturing subjects diving directly into the water or flying by it (think GBH's, waterfowl, shorebirds, etc.). I can easily capture an osprey or eagle skimming across the water with a catch or stick, or a pelican or a tern diving straight in with a Z8. KF's diving with the technique that I described are easy to record providing one can keep them in the frame, but for whatever reason, the Z8 loses AF capture at the key moment every, single time with eagles/osprey. I'm glad it's working well for you and not only would I like to learn about your settings and techniques but see some of your strike sequences from start to finish.

BTW, here's a couple of images of a brown booby near the deck. The Z8 had no problem acquiring or tracking across long sequences. https://bcgforums.com/threads/bbif-brown-booby-in-flight.41207/

 

[fcotterill]

{edited} These challenges are quite possibly the reappearing predicament of the lack of cross-type AF sensors in these mirrorless cameras. Presumably, this is because the deeper level algorithms operating the CD PDAF system depend primarily on vertical patterns for contrast - even if a camera may be using pattern information from subject detection. [In any case, unusual shapes like water droplets, and long necked birds can discombobulate subject detection, so it grabs the wrong object ]

It's hypothetical.... however, the rapid changes in the vertical topology of the patterns presenting - in the current split-second - could cause the AF to switch between suddenly appearing objects around/over the subject.

This could explain why it helps to turn off SD, and use tighter Custom Area AF modes - effectively forcing the AF to search only within/around this rectangle. The Custom Area modes are a big advantage in modern Nikon AF systems (one of their best innovations) such that I don't bother with any other of the group Area modes.

The Erratic AF setting should also help with these challenges, particularly if the subject changes speed suddenly.

The D6 still has advantages in tricky situations like these, which I think is logically explained by the cross-type sensor architecture of the AF module

z9 z8 AF won't stick to bird head.

Improvements to the Auto AF area since I first had a Z9 way back have been nothing but remarkable. My hit rate on kingfishers, in the wild, in flight with Auto AF has been remarkably good even with quite busy backgrounds. Almost as good as an A1 where I would mainly be using zone AF. Given how...

bcgforums.com

 

[fcotterill]

A couple more recent discussions FYI

Tips for shooting diving Osprey

I've had two sessions in an Osprey hide recently and while I can get tack sharp shots before the dive and when the bird comes out of the water, sharp focus of the actual dive has eluded me 9 times out of 10. The birds are anything from 35 to 80 feet away and I'm shooting with a D850 and 300mm...

bcgforums.com

Z8/9 Auto Focus (C) w/Sub Detection 🐦+👁️=🧐

Hope you see this one Steve… I ran into an annoying issue with auto focus and subject detection. It finds the eye and nails the focus….in the reflection that is closer to the camera. This “yellowleg” is an example (mock-up) of what I occasionally encounter. The calmer the water the more...

bcgforums.com

Interesting AF tips from Jan Wegener (not brand specific)

Some interesting tips for AF on birds in flight. A couple of strange ones make sense to me, like holding the camera at an angle. Focus pumping is another.

bcgforums.com