sk66 wrote:
In my testing, the best results in a demanding situation are to set "blocked shot" to the maximum delay and set "subject motion" to erratic.
This seems to tell the camera to continue tracking for the maximum time before reverting to the originally selected point (blocked shot delay) but it's still not a terribly long time (about 6 seconds max). And it tells the camera to expect to loose the focus area/focus momentarily so it doesn't revert as quickly (subject motion).
+1. On my D500 the blocked shot settings directly affect dynamic behavior. I've come to think of them as dynamic mode fine tuning.
I did some more experiments to further clarify how Nikon determines which AF point to report in the EXIF in dynamic mode. These experiments demonstrate that the reported AF point is not necessarily the most recent AF point the camera used to acquire focus but instead whether the initial AF point is in focus at the time of photo. The camera only reports an alternate AF point within the dynamic group if both that point is in focus and the camera is actively acquiring focus (AFON pressed) when the photo was taken.
I did these experiments on a D7200 but should apply to the D5/D500 as well. This was with D21, with the Center AF point as the primary point.
Experiment #1
Step 1. Focus on subject with center AF point, take photo while AFON still held. EXIF reports "Primary AF Point" and "AF points used" as the Center point. So far so good.
Step 2. Release AFON while holding center AF point over subject, then take photo. EXIF still reports "Primary AF Point" and "AF points used" as the Center point.
Step 3. With AFON still released, move camera so that adjacent dynamic AF point is over subject, then take photo. EXIF reports "Primary AF Point" as Center point and "AF points used" as (none)
So experiment #1 demonstrates that the camera will report the initial AF point as acquired even if its not actively engaged in AF-C (ie, AFON released), provided the initial AF point is still in focus for the subject.
Experiment #2
Step 1. Focus on subject with center AF point. While holding AFON, move camera right so that subject now on adjacent dynamic AF point (and center AF point no longer over target), take photo. EXIF reports "Primary AF Point" and "AF points used" as the adjacent point
Step 2. Release AFON while holding center AF point to the right of subject, with the adjacent dynamic AF point still over subject, take photo. EXIF reports "Primary AF Point" as the Center point and "AF points used" as (none)
Step 3. With AFON still released, move camera so that center point is now back on the subject, take photo. "Primary AF Point" and "AF points used" as the Center point.
Experiment #2 demonstrates that the camera will revert to reporting the Center AF point as the "Primary AF Point" and whether or not its in focus (via "AF points used") if AFON has been released, even if the adjacent AF point is still over the subject and the Center AF point is not. It will not do the same for an adjacent AF point if AFON has been released.
Experiment #3
Step 1. Focus on subject with center AF point, take photo while AFON still held. EXIF reports "Primary AF Point" and "AF points used" as the Center point.
Step 2. Put camera down and swap out subject. Original subject was white text on yellow cereal box - replaced subject with a grey camera box with black lettering. New subject carefully placed at same position as original subject
Step 3. Pick camera back up and without pressing AFON, took photo. EXIF reports "Primary AF Point" and "AF points used" as the Center point.
Experiment #3 demonstrates the same thing as experiments #1 and #2, with the additional information that the camera will report the Center AF point acquired not just if its on the original subject and in focus with AF-C disengaged (ie, AFON released) but also for a completely different subject. This further proves that camera will report the Center point has successfully acquired if its in focus, irrespective of whether the original target is what's in focus. However as experiment #2 demonstrates, the camera wont report the adjacent point as the "Primary AF Point" and "AF points used" when AFON is released even if the adjacent point is still over the subject - it instead will report the Center point as "Primary AF Point" and (none) as the "AF points used"; this makes sense since because AFON is released, the camera reverts to only measuring whether the user-specified point in the dynamic group is in focus.
henry albert wrote:
+1. On my D500 the blocked shot settings directly affect dynamic behavior. I've come to think of them as dynamic mode fine tuning.
Same on the D5, and I have mine set to Quick and in D25 when I loose focus on my subject the AF snaps back quickly when AF point comes back to the subject. I like this a lot better.
I've been following this thread from the beginning and want to say thanks to Steve, Snapsy, Kwilliam8, and the others that helped with this analysis. Great job.
I've read through Steve's description of what he thinks is going on several times now, and I just don't see any way this new behavior can be an improvement. From what I understand, if the selected AF point comes off the target the camera will use another AF point within your Dynamic area. However at some point, the camera will abandon that new AF point and try to return to the originally selected AF point. This process then repeats itself for as long as the focus sequence continues.
What this tells me is that the camera AF will be both unpredictable and unreliable since you don't know exactly when this behavior will begin and how many times it will repeat.
It also complicates things for someone like myself with an older style AF camera and this new AF type equipped camera (a D4 and a D500). I won't be able to shoot both cameras quite the same way when using dynamic mode - or if I do the results will be less consistant on one of the bodies.
I hope Nikon sees fit to fix this on the D5/D500 series.
I'm going to do some experimenting this weekend with D25 and the AF settings. I'm thinking for my use case setting the delay to maximum 5 but setting the behaviour to erratic may be best or maybe not??
When tracking a BIF (only reason I'd use D25) I'd like to have a little time to get my focus point back on the head or body.
However, for most BIF cases other than large birds flying fairly close (which have the risk of focusing on the near wing), I think Group AF is the best mode anyways. If the bird is flying against a busy background then Group is designed to focus on the nearest object which will be the bird. Setting the Delay to maximum will also help in this mode if you do move your Group off the bird for a second you have more time to get back on the bird before it grabs the busy background.
The more I think about it, I don't have a lot of use cases for Dynamic as it is implemented in the D5/D500. I think I will use single point for perched birds and Group AF for BIF. I will only go to Dynamic for large birds (in my case eagles) where I often can shoot them rather close and don't want a near wing AF grab.
AnnJS wrote:
Is it possible that the difference between the way that Dynamic AF modes work on the D5 and D500 and the way that they appeared to work on earlier DSLRs is due to the faster-responding AF in the newer two models?
In other words, if a photographer is not sufficiently skilled to hold the prime focus point over a moving target, they will need to use a Delay in the D5/D500 while the slower reaction to loss of contact between the prime focus point and the object being tracked on a D810 acted as a "Delay" by its nature without the need to program a Delay as such?...Show more →
I would not agree with that. Referring strictly to the D500, I found initial response quicker than my D4 but after that first "jump" my D4 ( now replaced with D4s ) quickly caught up and was more accurate and precise. Only thing that came to mind when I noticed this was when car manufacturers downgraded horsepower due to fuel cost, but to give that performance feel, they went with a higher gear number. Quick at first, but that was it. Anyway............
I could maybe see your second point on not sufficently skilled photographers if it was the D5 only. A very slim maybe. The D500 is not a pro body, and it would be foolish for Nikon to think that it is being mainly used by highly skilled photographers. On top of that, isn't having 9 pts or 25 pts or whatever, an advancement in technology to further AF performance?? To make life a little easier?? Why build a prosumer body that everyone would have to practice their tracking technique to track as if it's single point??
Sorry, but I'm just not buying into the "born this way" thing. It's a flaw. And, the D500 / D5 should be good enough without a delay. I set my D4 / D4s to A3=off
I agree that the old way is better, and again, I want to emphasize that I'm still not entirely convinced that this isn't a flaw in the design. It really looked that way at first due to the different way it was reporting the AF point in View NX-i. However, it's clear that the system is sticking with the subject now, if only briefly before it tries the main point again. Is this intentional or an accident? That's the real question.
Either way, I think the old system is better, even if I know how to use the new one. Just the real world evidence / results from shooters in this thread is enough to convince anyone that this is a step back. I wonder how many are using Group AF now that would actually do better with the old Dynamic system.
In any event, it'll be interesting to see if the behaviour stays the same on the next MultiCam 20K body. If it does, it's there by design, if it doesn't it's a flaw.
(Wouldn't it be funny if this AF glitch is why we haven't seen a new FX body since the D5?)
arbitrage wrote:
I'm going to do some experimenting this weekend with D25 and the AF settings. I'm thinking for my use case setting the delay to maximum 5 but setting the behaviour to erratic may be best or maybe not??
Well, that works pretty good for hockey banzai rushes to the net, but I can't speak for birds. That's the settings I've settled on, along with a renewed effort to keep the primary focus point where it should be in the first place. I have to admit that all this technological assistance might have caused me to be a little lax in my focus point discipline.
p.10 #10 · UPDATE - Nikon D5/D500 Dynamic AF Issue
Well Steve, I am really disappointed in your observations. You have clearly identified that I am a rubbish photographer. I was hoping to blame all my issues on the new system, but oh well.
Seriously, I have a 7100 also and it is really difficult to understand the logic if indeed it is a new method. So much for my safety net
p.10 #11 · UPDATE - Nikon D5/D500 Dynamic AF Issue
boblug wrote:
Well Steve, I am really disappointed in your observations. You have clearly identified that I am a rubbish photographer. I was hoping to blame all my issues on the new system, but oh well.
Seriously, I have a 7100 also and it is really difficult to understand the logic if indeed it is a new method. So much for my safety net
It's still a safety net, there's just more holes in it now
p.10 #12 · UPDATE - Nikon D5/D500 Dynamic AF Issue
I'm glad I ran into this post. I just sent my D500 in for what seems like the same issue. I was so used to the dynamic hanging on to whatever I focused on and staying there. Always seemed that when I was tracking it would lose focus mid sequence and jump back sometimes. Othertimes it wouldn't track. I am hoping to hear good news from Nikon, but who knows.
p.10 #13 · UPDATE - Nikon D5/D500 Dynamic AF Issue
For whatever it is worth… I would prefer that the D5/D500 dynamic area AF systems have the ability to stick with the initial focus subject beyond one second (or so), especially if the subject is still within the specified area (e.g., D72). Ideally this delay time would be fully controlled by the user. Unless the firmware is updated, the use of Dynamic Area AF on the D5/D500 will be challenging for many D5/D500 users. I am not sure that Nikon fully appreciates this.
Like others here, I suspect that many D5/D500 users have switched from Dynamic Area AF to either Group-area AF or 3D-tracking AF (although those AF area mode systems have their own advantages and disadvantages). Who knows, maybe the Nikon dynamic area AF systems will disappear in the future, since there are only a few idiots like me still using them and trying to make them work!
p.10 #14 · UPDATE - Nikon D5/D500 Dynamic AF Issue
Steve Perry wrote:
However, it's clear that the system is sticking with the subject now, if only briefly before it tries the main point again. Is this intentional or an accident? That's the real question.
That leads again to what I posted about the D5 being a more sports oriented camera. The release timing and the Olympics, the DR/ISO tradeoff and this AF system. I can't imagine not wanting the camera to use the primary AF point when it can regain focus. Some other point taking over and tracking? That seems more like video. Reading here, the Dynamic behavior is mostly a problem for wildlife photographers. I haven't heard any sports shooter complain about the D5's Dynamic AF on sports specific forums. Viewing Nikon's marketing material on their site, it's clearly leaning towards the sports shooter - view the sample images. On the same token, the D500 has a more wildlife angle - view the materials and the sample images. Plus the bundling of the D500 with the 200-500 is really a shot at wildlife photographers. That's not really a sports lens.
The inclusion of the AF system on the D500 was probably for marketing horsepower, especially because the AF gets complaints on the Canon 7Dnk2. "Bursting with new technology"... are you going to market old technology? Plus it has the AF modes that seem to be working for wildlife.
Interestingly, Nikon Messaging Center is ticker taping as I write this, that there's a new update for ViewNX-i to 1.2.5. No idea what it encompasses, I haven't downloaded it as I never use it.
p.10 #17 · UPDATE - Nikon D5/D500 Dynamic AF Issue
Steve, big thanks to you and your helpers on delving through all this.
For me the old Dynamic works better.
Nikon bring out new AF, shout about it, more AF points, more frame per second etc, then make it harder to use??
I have trouble getting my head around that one!
To make it more difficult, those particular photographs were shot with a TC-20iii on a 300mm PF so I wasn't even able use the D5's brilliant 3D Tracking AF in those cases.
I find it rather interesting that "3D Tracking AF" is the only AF mode to which Nikon actually attaches the word "Tracking"?!
To make it more difficult, those particular photographs were shot with a TC-20iii on a 300mm PF so I wasn't even able use the D5's brilliant 3D Tracking AF in those cases.
I find it rather interesting that "3D Tracking AF" is the only AF mode to which Nikon actually attaches the word "Tracking"?!
Ann, beautiful images! Which dynamic area modes do you use for BIF (e.g., D72)? Thanks for any insight that you can provide.
p.10 #20 · UPDATE - Nikon D5/D500 Dynamic AF Issue
I use 3D Tracking more and more of the time (because it makes tracking so easy!) and I also have my joy-stick set as a second AF-ON button — usually for AF-C Single Point.
Unfortunately, 3D Tracking is not available when you have a TC mounted so I was probably using Dynamic 25 or 72 for the shots in that Thread. I don't remember exactly which and, unfortunately, the AF Mode doesn't seem to be recorded in the EXIF.
That's the settings I've settled on, along with a renewed effort to keep the primary focus point where it should be in the first place. I have to admit that all this technological assistance might have caused me to be a little lax in my focus point discipline.