sk66 wrote:
That's the "blocked shot" response... for changes in distance when something comes between you and the subject. It's what I was describing earlier in that the D8xx and other cameras will refocus, but not as consistently/quickly as the D5/D500.
It's why I described the D5/D500 as being hyperactive...
However, I had become fairly convinced (w/ bench testing) that the tracking point had to be lost, and you're now saying that's not the case(?) w/ the moving target scenario... that's what I originally thought.
It's not really the blocked shot response - at least not the way it usually works for the older system. With the scenario I describe, I'm keeping the initial back target well within the Dynamic Area. If you go the opposite way (front target and move the primary to the back target) the system will stay with the first target indefinitely. With the old system, Blocked Shot Response usually happens when the subject completely leaves or is obscured from the AF area. Also, when it does do the switching from back to front target, if those targets are close the switch is nearly instant - regardless of the Blocked AF response setting.
To see what I mean, just place two targets on a table, one about 2 feet behind the other. Using about a 50mm lens, stand 6 feet back or so and lock in the back target. Now, move your primary point over the front target while still keeping the Dynamic area over the back one. The camera will switch to the front. Try the same thing the opposite direction, and the camera holds the front target. Of course, if there's enough separation between the targets it will stick with the original (like if you get within a foot of the front target).
sk66 wrote:
By switching to a dedicated AF sensor it is possible that it significantly increased the speed and sensitivity(range). I.e. the incremental increases in AF performance really were not enough to be noticed as such... let's face it, tracking birds in flight was never entirely reliable and it's easy to discount a change in behavior as being something else.
I had a lens develop focusing issues... it took quite a while and it had to get quite bad before I started looking for equipment issues.
I still don't think it's a speed issue. A slower system should still do the same thing in some circumstances and it really doesn't seem to. Like you say, I suppose it's easy to dismiss an occasional mistake from the AF system here and there. but I still don't think this is a result of an increase in speed.
"With Single point, Group, d25, d72 the same thing applies: they track moving subjects better than ever before. Some of that is the extra autofocus sensors, which give the system more nuance of position in any given area than before (part of the accuracy thing, too). But a larger part has to be whatever’s in that focus CPU Nikon cooked up.
Indeed, as many folk have discovered d25 and d72 can now be problematic for certain shooting situations because they're so fast at tracking. That, coupled with a slightly higher tendency to come back to the initial sensor selected means that people are discovering that Dynamic AF doesn't perform the same on the D500 (or D5) as it did on the previous models (e.g. D810). Most of this is how fast the dedicated focus CPU works, but it also is a bit of misunderstanding as to why the Dynamic AF modes worked for people on older cameras.
Simply put, d25 and d72 are now wicked fast at tracking erratic movement. They appear to track truly erratic movement better than any previous Dynamic AF. But that's not what people were using Dynamic AF for in older cameras. Most were using Dynamic AF to handle modest, slower, and predictable motion, and to compensate for when they couldn't hold a subject steady in the viewfinder. In my sports shooting, I'm noticing that the D500 (and D5) react very quickly to truly erratic motion, faster than I could react.
I will say this: Nikon seems wrong in their suggestions on CSM #A3: shorter tracking times now make the system too jumpy for me. In almost every case where Nikon has suggested a value of 2 for Blocked Shot AF Response, I’m using 3 or 4. The Subject Motion setting seems to be accurate to Nikon’s documentation, though."
So all that I know is that the Dynamic modes don't work the way I would have expected them to and my keeper rate is way down compared to Group.
viczig wrote:
This is what Thom Hogan says about it:
"Nikon has not changed the way that Dynamic AF works. What has changed is the speed at which it works, which makes for a different behavior than people are expecting. Indeed, many people were using the old Dynamic AF modes BECAUSE they were acting slower. They essentially made for a useful tool for when you couldn't quite hold the camera steady, because they wouldn't immediately move off the selected focus point. Now they do because they're incredibly faster."
Yes but the faster behavior has now made it unworkable. That's where Nikon should have tuned the system to react as it used to. It doesn't have to react as fast as it possibly can; you govern the speed to the circumstances.
I don't think it has anything to do with holding the camera "steady" It has to do with keeping the selected focus point on target. There's obvious difficulty to that and part of the reason why we've moved from 11 point ( and probably less ) to 9, 21, 51 up through 25, 72, and 153. To compensate for the fact that every subject is not easily tracked. I can't see how anyone, Thom included, can make excuses for Nikon when the current dynamic system has essentially turned into a bunch of single point modes.
I have a lot of Nikon gear but no fanboyism ever clouds my opinion on when something is working right or working wrong. And the dynamic AF modes is yet another Nikon screwup. Slow it down by firmware if need be.
viczig wrote:
Thom addresses it in his review of the D500:
"With Single point, Group, d25, d72 the same thing applies: they track moving subjects better than ever before. Some of that is the extra autofocus sensors, which give the system more nuance of position in any given area than before (part of the accuracy thing, too). But a larger part has to be whatever’s in that focus CPU Nikon cooked up.
Indeed, as many folk have discovered d25 and d72 can now be problematic for certain shooting situations because they're so fast at tracking. That, coupled with a slightly higher tendency to come back to the initial sensor selected means that people are discovering that Dynamic AF doesn't perform the same on the D500 (or D5) as it did on the previous models (e.g. D810). Most of this is how fast the dedicated focus CPU works, but it also is a bit of misunderstanding as to why the Dynamic AF modes worked for people on older cameras.
Simply put, d25 and d72 are now wicked fast at tracking erratic movement. They appear to track truly erratic movement better than any previous Dynamic AF. But that's not what people were using Dynamic AF for in older cameras. Most were using Dynamic AF to handle modest, slower, and predictable motion, and to compensate for when they couldn't hold a subject steady in the viewfinder. In my sports shooting, I'm noticing that the D500 (and D5) react very quickly to truly erratic motion, faster than I could react.
I will say this: Nikon seems wrong in their suggestions on CSM #A3: shorter tracking times now make the system too jumpy for me. In almost every case where Nikon has suggested a value of 2 for Blocked Shot AF Response, I’m using 3 or 4. The Subject Motion setting seems to be accurate to Nikon’s documentation, though."
So all that I know is that the Dynamic modes don't work the way I would have expected them to and my keeper rate is way down compared to Group. ...Show more →
It's not clear why he believes faster tracking of erratic subjects is related to the camera choosing to select a new target that falls under the primary AF point. You would think faster tracking would enable the exact opposite - the ability to track an erratic-moving subject longer by not confusing a new subject for the tracked subject, whereas a slower implementation might mistake the two subjects due to a slower sampling and focusing rate. Even my lowly D7200 is capable of instantly switching to a new target when it wants - luckily it does so in a smarter and more predictable manner.
I still think that the D5 AF is far better than it was in earlier cameras and that one simply has to adjust one's camera-handling to take full advantage of the new system instead of bewailing the fact that old habits no longer cut the mustard nor catch the bird.
T-O Shooter wrote:
Yes but the faster behavior has now made it unworkable. That's where Nikon should have tuned the system to react as it used to. It doesn't have to react as fast as it possibly can; you govern the speed to the circumstances.
I don't think it has anything to do with holding the camera "steady" It has to do with keeping the selected focus point on target. There's obvious difficulty to that and part of the reason why we've moved from 11 point ( and probably less ) to 9, 21, 51 up through 25, 72, and 153. To compensate for the fact that every subject is not easily tracked. I can't see how anyone, Thom included, can make excuses for Nikon when the current dynamic system has essentially turned into a bunch of single point modes.
I have a lot of Nikon gear but no fanboyism ever clouds my opinion on when something is working right or working wrong. And the dynamic AF modes are yet another Nikon screwup. Slow it down by firmware if need be. ...Show more →
TO Shooter has got it quite right here, and Thom Hogan has got it wrong!
This is nothing to do with faster processing or expectations, it has to do with Nikon's implementation of the previous excellent dynamic mode into the D500, the net result is we've gone from patting Nikon on the back for their previous system to suffering from it in it's present form. We now have three dynamic modes that serve at best to track birds across a blank sky, and even that is a maybe!
Hogan is telling you how AF is actually implemented in the D5 and the D500 which is quite different from the way that many people seem to think that it works.
Once you understand the new dynamics and learn how to work with them, they can work extremely well — they certainly do for me.
Something can overall be better but still not be right. Obviously the new AF overall is much better. I think from following this thread they likely have a coding error that they picked up adapting to the new processor and the non-selectable/displayable points. Will be interesting to see what Nikon says. It appears they have started testing it now at least in the subsidiaries from some of the recent posts.
AnnJS wrote:
I still think that the D5 AF is far better than it was in earlier cameras and that one simply has to adjust one's camera-handling to take full advantage of the new system instead of bewailing the fact that old habits no longer cut the mustard nor catch the bird.
p.17 #10 · UPDATE - Nikon D5/D500 Dynamic AF Issue
Steve Perry wrote:
Also, when it does do the switching from back to front target, if those targets are close the switch is nearly instant - regardless of the Blocked AF response setting.
I had some time today while I was watching a Screech Owl sleep and I ran the test with targets about 2ft separated in depth (and shallow DOF). In my test w/ the D810 the blocked shot setting definitely made a difference in how quickly it would switch to the near target.
I also had an opportunity to try Harriers with busy backgrounds... they stayed too far out for good pictures. What I found was that the smaller/tighter sensors made it much easier to miss the target during initial focus lock. When I obtained a good lock it tracked, unless I lost the subject from the AF zone. This is based upon back of the camera review... I haven't had time to download the images yet but I don't see it changing the results.
p.17 #11 · UPDATE - Nikon D5/D500 Dynamic AF Issue
AnnJS wrote:
I still think that the D5 AF is far better than it was in earlier cameras and that one simply has to adjust one's camera-handling to take full advantage of the new system instead of bewailing the fact that old habits no longer cut the mustard nor catch the bird.
You know where I've seen something very similar before? My 1DIII. Complete with some owners telling me there was nothing wrong with it and I was just a bad camera operator.
p.17 #12 · UPDATE - Nikon D5/D500 Dynamic AF Issue
Ok, so at the event I photographed last weekend (dog agility) I went between group, 25pt dynamic, and 72pt dynamic. I was using my D5 with a 70-200 VR1 and TC14EII @ F4 1/800SS. My keeper rate with dynamic 72 was much higher than in group. I also found the 25pt was almost as good as 72, but not quite. I had blocked AF set to 4 and subject as erratic. This being said, group AF worked better when the background was busy. 72 and 25pt would jump to the background sometimes in those situations. All in all, as I get used to this new behavior, I am actually liking it. Whereas my D810 is much different, I really don't use it for action. It stays at home in the studio. We'll see if Nikon makes any changes in the future in firmware updates, but now that I know what's going on, it's all good.
p.17 #13 · UPDATE - Nikon D5/D500 Dynamic AF Issue
Henry. When I read Steve's first post the first thing to come to mind was the 1D MK III fiasco but I was afraid to say it out loud. Funny thing is I still have my 1D III.
Mar 23, 2017 at 05:43 PM
T-O Shooter Offline • • [X]
p.17 #14 · UPDATE - Nikon D5/D500 Dynamic AF Issue
henry albert wrote:
You know where I've seen something very similar before? My 1DIII. Complete with some owners telling me there was nothing wrong with it and I was just a bad camera operator.
Always the case. There's always someone who has one what works for them, or are delusional about having one that works for them, who'll try to convince you that it's "poor technique" or "user error" on your part. Same old, same old.
p.17 #15 · UPDATE - Nikon D5/D500 Dynamic AF Issue
Steve Perry wrote:
I still don't think it's a speed issue. A slower system should still do the same thing in some circumstances and it really doesn't seem to. Like you say, I suppose it's easy to dismiss an occasional mistake from the AF system here and there. but I still don't think this is a result of an increase in speed.
Hi Steve..
It's clear to me and even clearer to you "exactly" what the D500 is doing in the three dynamic modes.
I've spent some hours checking the system's behavour.
Unfortunately the three modes serve very little to the sports or BIF. Emailing Nikon is very likely to achieve very little as there are too many "buffers" in front of the chiefs!
I ask you again........... "please do an AF addendum video".......A visual explanation albeit in english will communicate the problem to the world!
Please at least reply to this post, I noticed my previous one went unanswered!
p.17 #16 · UPDATE - Nikon D5/D500 Dynamic AF Issue
daves cliches wrote:
Hi Steve..
It's clear to me and even clearer to you "exactly" what the D500 is doing in the three dynamic modes.
I've spent some hours checking the system's behavour.
Unfortunately the three modes serve very little to the sports or BIF. Emailing Nikon is very likely to achieve very little as there are too many "buffers" in front of the chiefs!
I ask you again........... "please do an AF addendum video".......A visual explanation albeit in english will communicate the problem to the world!
Please at least reply to this post, I noticed my previous one went unanswered!
Thanks in advance for your help in this matter!...Show more →
Hadn't responded because I was still thinking about it
I'm in another project right now, but a video is a good idea. If for no other reason than just to explain how the new system works.
p.17 #17 · UPDATE - Nikon D5/D500 Dynamic AF Issue
Steve Perry wrote:
Hadn't responded because I was still thinking about it
I'm in another project right now, but a video is a good idea. If for no other reason than just to explain how the new system works.
-----------------------------------------------------------------------------------------------------------------------------------------------------
Thanks for your prompt response Steve!!
Just to say I have had a little experience when finding a problem with another company's DSLR to whom I sent several emails, even via an executive. Needless to say when pursuing the matter after some time the executive had heard not a word from "the top". To many intercepting buffers en route!
Time has gone by since the release of the camera and overall it's reputation has only been rising, and surprising as it now seems no-one unearthed this "hiccup" even in spite of all the top notch photog's reviews which abounded, so kudos to you there! I had only a few days of head-scratching which was when I de-niched your thread.
I think our only hope in reality is if the subject is brought out into the open in video form, which would in some part go towards breaching the language barrier that's is probably limiting it's propagation now.
p.17 #18 · UPDATE - Nikon D5/D500 Dynamic AF Issue
i became aware of this most interesting thread on NikonGear [Multi-CAM 20k AF system behavior (D5/D500) ]. There I drew attention to a test report of the D5 in DSLR Photography late 2016 [by leading automotive photographer Dean Smith]. He found that the D5, compared to the D4, under-performed on tracking fast-moving cars. While this likely tested an early released D5 perhaps, a following poster noted a firmware of the D5 had including the 9-pt dynamic area mode, and improved auto-area AF and 3D Tracking.
