I've heard that some lenses have 2 focus motors while others have only one. Or the type of motor varies stepping vs. other types. Perhaps the lens can't keep up with the frame rate and rapid refocusing. Might be worth an experiment to reduce the frame rate and see if it eliminates the issue.
Help Identifying a Particular Sort of Image Defect
- Thread starter SCoombs
- Start date
If you would like to post, you'll need to register. Note that if you have a BCG store account, you'll need a new, separate account here (we keep the two sites separate for security purposes).
As Cameron said, atmospherics absolutely do change in fractions of a second. Of course, not the overall condition of "there's an atmospheric condition" vs "not an atmospheric condition" - but with the turbulence of the air, you can get brief pockets where you get a clean/clear image in the middle of a bunch of blurry ones, or vice-versa depending on the specific conditions at the time. I can show shots with my 800f5.6 f-mount and 600f4 f-mount on D850 bodies, as well as those same lenses on Z9 bodies, as well as the 600TC and 400TC on my Z9 bodies all showing similar results.
Where I live and shoot I deal with atmospheric distortion nearly every day I'm out shooting, and I have taken thousands of images that look exactly like your gull shot including actual gull shots. And in fact when I first saw the photo, even without looking at the text/question/etc my first thought was "atmospherics does it again". I'm not nearly as experienced as folks like Steve and others who have said they're leaning toward atmospherics too, so I'm not sure my opinion really helps one way or the other, but I'd strongly lean toward that being the problem.
One thing I *have* noticed though, FWIW, is that my Z9 seems to get a bit more tripped-up with AF in the presence of atmospheric distortions than my D850 did, and the Z9 seems to do a lot more minute hunt-and-pecking in the presence of atmospherics than the D850.
Cameron T
Well-known member
Out of curiosity, are you using ai focus modes, or non ai focus modes when it hunts more?
I'd expect more hunting with ai modes as the shapes it's trained to recognize (bodies, heads, eyes in that order) shift more constantly and come in and out of focus (due to distortion) more and more, while the old method should stay relatively more stable.
I could be way off base there though, that's just been my experience
I also thought this was atmospherics because overall the look of a lot of the problem images looks like it to me, BUT what prompted the question was the double image/ghost effect on that gull, which is something I'd not previously noticed with atmospherics. When you say it looks just like it, do you also mean that double image effect?
Interesting!
I got results like that one time when I left my VR on and shooting at 1/2000 to 1/4000 with my Z9 and 600 f6.3. Had to toss most of the BIF and moving shots. I won't make that mistake again for a while.
I was 99% sure my VR was off, but after your comment I did check the file and verify. It was indeed off.
I suspect you're right, and yes I use subject-detect most of the time. In the presence of atmospherics, it's definitely more likely to bounce between different sizes and locations of boxes, which is usually my signal to turn off SD. Even with SD off, it still seems to have more issues than my D850 did, but that's based loosely on memory/feeling - I've never done any rigorous tests.
Oh yes, I've definitely seen that ghosting effect. If I have time, I might try and dig up some examples from my library. But just generally, I would actually *expect* that type of ghosting to occur. The light is getting bent in different ways as it moves through the mixed density air, so I would expect to see some parts of the image slightly "out of place".
Hi Shane - Good posts.
The double ghosting seems to me like a focus error. As focus moves away from the desired plane, you get an out of focus area. If it was caused by motion, it would most likely be only on one side of the subject at the speeds you are using. I'm not saying this is a user caused focus error. LensREntals.com did a test a few years ago that showed all lenses have a distribution of focus variation - even using contrast detect AF only.
I know the AF is constantly being refined on the Z cameras. Normally that's a good thing. But I wonder if that constant refinement with high frame rates is catching some images that are part of the camera's adjust and refine routine? Perhaps some testing at slower frame rates would provide some insight?
Don't overweight the AF boxes being green or on your target. While it uses the same data stream for the EVF and focus, it can be a backward looking view of the image just taken or focused, not the image to be captured. While it is very accurate most of the time, it's not captured concurrently with the image. The Green box is also an overlay rather than an output from actual focusing by the camera.
I expect the AF motors in a 180-600 could be different from AF motors in the 800mm PF or 400mm f/4.5. It might be possible that the AF motor is not driving the movement with enough precision, so the camera is constantly making adjustments and correcting errors.
Testing at close distances is a start, but probably not close enough to the issue you are having. I'd probably try more testing with progressively more and more distance from the target. I've found that AF at close distances is usually pretty good and with distance any error is magnified.
For Nikon to have any serious discussion with you, you probably need repeatable methodology that shows the issue and isolates other potential issues. Even if it's not how you use the lens, I'd test on a tripod.
The double ghosting seems to me like a focus error. As focus moves away from the desired plane, you get an out of focus area. If it was caused by motion, it would most likely be only on one side of the subject at the speeds you are using. I'm not saying this is a user caused focus error. LensREntals.com did a test a few years ago that showed all lenses have a distribution of focus variation - even using contrast detect AF only.
I know the AF is constantly being refined on the Z cameras. Normally that's a good thing. But I wonder if that constant refinement with high frame rates is catching some images that are part of the camera's adjust and refine routine? Perhaps some testing at slower frame rates would provide some insight?
Don't overweight the AF boxes being green or on your target. While it uses the same data stream for the EVF and focus, it can be a backward looking view of the image just taken or focused, not the image to be captured. While it is very accurate most of the time, it's not captured concurrently with the image. The Green box is also an overlay rather than an output from actual focusing by the camera.
I expect the AF motors in a 180-600 could be different from AF motors in the 800mm PF or 400mm f/4.5. It might be possible that the AF motor is not driving the movement with enough precision, so the camera is constantly making adjustments and correcting errors.
Testing at close distances is a start, but probably not close enough to the issue you are having. I'd probably try more testing with progressively more and more distance from the target. I've found that AF at close distances is usually pretty good and with distance any error is magnified.
For Nikon to have any serious discussion with you, you probably need repeatable methodology that shows the issue and isolates other potential issues. Even if it's not how you use the lens, I'd test on a tripod.
I've tested this on a tripod with the standard target sort of thing and it seems to produce reasonably good results.
You can only see EXIF info for this image if you are logged in.
However, this is for a distance of only around 10 meters. Until the weather improves significantly, I don't really have a way to do this sort of thing at longer ranges.
I have also used FoCal, which has given the lens a rating of around 78. It could be better (my 200-500 was rated at 94), but it at least indicates it's an above average copy in terms of the optics.
FoCal also complains about inconsistent focus, but for FoCal this means something like 98% consistency.
In real world usage, I've found that as long as I keep the VR either on or off as is appropriate for the shutter speed, the lens will give good results at ~10 meters or closer.
It's longer range usage where I have questions and where things have been a lot harder to parse out. The main location I've been able to test things out at longer ranges is one where I have I have been getting fairly disappointing or even just bad results. The question is whether this is due to the lens itself or to some mitigating factor like atmospherics.
I had believed a lot of what I was seeing was atmospheric distortion, but this gull photo sent me down a path which now has me wondering about everything from a lens problem to the focus just being more inconsistent on this lens than on others, possibly exacerbated by atmospherics.
The difficulty is that testing longer range makes it more difficult to distinguish between atmospheric factors and factors more particular to the lens.
Patty
Member
As someone who takes a lot of shots of running dogs, especially running toward me, I have discovered that VR played a factor in keeper rates. So, for fast action I either turn off VR or set it to sport mode. When I do that I get more keepers. Just an observation and same from other pet photographers. I know dogs don't always run as fast as the wild critters but they can move really fast. Give it a try and see what happens.
O
Well-known member
Agree, in my case by default i turn the VR off on my 200-500, 300 2.8 VR II, 600 F4 FL or most other lenses anything over 1600- 2000 ss.
John Navitsky
Well-known member
i’d bet on Steve knowing better to me but it looks to me like focus wandered a bit, then corrected itself
John Navitsky
Well-known member
i suspect it depends on shutter speed. i find myself gravitating to faster and faster shutter speeds and i think results continue to improve
i find vr can impact me TRACKING dogs in sometimes negative ways, but in general it’s a framing problem not a focus/sharpness issue. as a result, i too run in sport mode, not “normal” when shooting dog sports
that said, i often find images that are slightly out of focus in bursts, but i think this is mostly dogs tend to move in ways (esp during sports or play) that has changes in velocity which i suspect interferes with the ability of the camera to predict accurately where it should focus
John Navitsky
Well-known member
imo, because the other images were sharp, logic suggests it’s not a problem with the lens optics
doesn’t tell us what it is, or how to fix it, but i don’t think i’d fret about the lens
John Navitsky
Well-known member
i think there are cases where the camera has difficulty focusing despite looking like it is doing ok. you note heavily cropped and i think it’s much more common for the af to struggle when the subject is a bit small in the frame
John Navitsky
Well-known member
another thing, we have different ideas: af miss, vr hiccup, motion blur, atmospherics;
take a hypothesis, and make a plan to test it. for example vr is easy, toggle it, test, evaluate and then (if needed) make a new hypothesis, rinse, repeat
take a hypothesis, and make a plan to test it. for example vr is easy, toggle it, test, evaluate and then (if needed) make a new hypothesis, rinse, repeat
I used to think this was the case but I've since realized that the smaller something is (i.e., the further it is from the camera) the harder it is for focus to miss since the depth of field starts to become greater.
John Navitsky
Well-known member
i think the other way to look at it is the more likely it’s not focusing on what you think it’s focusing on
O
Well-known member
As a test only, Bump the SS to 4000 float the iso to 12800, TURN the VR off on the lens completely, i think you may be surprised it fixes the issue, i know it does with the 200-500 and 28-300.
Also i believe default VR setting in the camera is sport mode, you haven't set it back to normal have you??
Also i remember having hits and misses ages ago and i was asking the camera to nail focus using 600 ml with single point on a relatively larger white environment and low contrast surfer on a white foaming wave, i set the point to 12 points, my hit rate went to 100%, that was a DSLR, ie: i am eluding to a settings issue to be looked at as a wild card suggestion.
Also i believe default VR setting in the camera is sport mode, you haven't set it back to normal have you??
Also i remember having hits and misses ages ago and i was asking the camera to nail focus using 600 ml with single point on a relatively larger white environment and low contrast surfer on a white foaming wave, i set the point to 12 points, my hit rate went to 100%, that was a DSLR, ie: i am eluding to a settings issue to be looked at as a wild card suggestion.
Nailing focus is a skill. Even with the same camera and technology, someone who practices, uses good technique, and concentrates will have a much higher hit rate. I'm not saying that is the issue here - just in a general sense. At 100-200%, I can see a difference over a set of images and pick better or worse images - all of which achieve focus.
O
Well-known member
Agree, in fact i love Group focus in the DSLRS, such a handy tool.
I don't think I have a max ISO set. If I do it's 25,600. I'm generally not one of the people who gets concerned about ISO and lets it affect other settings. I also shoot in manual mode almost 100% of the time and have messed around with very high SS like 1/4000 or 1/5000, VR off, as you suggest.
I don't know, Eric. On a DSLR this would make sense to me, but with the way mirrorless focus systems work and the way subject tracking works and all of that I have a hard time seeing someone's skills being an issue for focus unless we're talking about a very drastic or extreme issue they have with keeping the camera steady. There's a lot of skill in getting a focus point or even a more forgiving option like a group focus array on the right place on a target keeping it on there, etc., but when autofocus amounts to pressing a button and the camera puts the focus point on the perfect spot... eh. Sure, using the wrong settings overall could result in unsharp images, but I'd consider stuff like this - motion blur, etc. - not to be a question of focus so much as general sharpness.
O
Well-known member
I guess its a process of elimination looking for the key cause, the specs i suggested is just the way i shoot also in manual its not so much about iso.
Have you had a different lens produce the same symptoms, if so your settings is most likely the issue,
solution reformat the camera and lens completely to factory specs, if that removes the issue your done,
if not you have your answer, take it to Nikon.
Only an opinion
John Navitsky
Well-known member
you might check out Steve’s video on youtube about high iso nr. in there you can see:
1) he has focus misses like this
2) even on static subjects like yours
1) he has focus misses like this
2) even on static subjects like yours