Thom Hogan
Forum Pro
Doesn't really change my answer. Only potentially eliminates a suspect.
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
Strange problem with images from 180-600
- Thread starter SCoombs
- Start date
- Status
Thom Hogan
Forum Pro
Being in a hurry led me to an oversimplification. Should say "not necessarily" instead of "not," but also important is where the optical center of the lens is vis-a-vis the correction being applied, as well as the type of correction. It's complex.
Thom, let me put it this way: the fundamental question I'm trying to sort out here is whether the sort of "double image" I'm seeing with that gull is consistent with some range of factors that can be expected from normal operation of a properly functioning and at least reasonably well calibrated lens or whether there is some indication here of a problem with the calibration (not the AFFT), misalignment of element(s), etc.
Thom Hogan
Forum Pro
I understand that. Unfortunately the answer is indeterminate from an image.
The easiest way to figure out if it is the lens is to test under the same conditions with another sample of the lens. Most people can't do that, however.
This doesn't make a whole lot of sense to me. Of course one can't say for sure what is wrong with a lens from a single image, but surely one can say if a particular effect can be the result of a particular class of defect or not.
If a doctor sees a photo of a broken leg, he isn't going to be able to give an extensive and specific diagnosis, but he's going to be able to say whether or not the injury is the possible result of eating a bad clam.
Thom Hogan
Forum Pro
But we can't say what caused the issue with a single image. Was it lens, camera, handling, intersection of frequency items, etc. Heck, I've see a single image that was distorted by a very temporary atmospheric "burp". Your first image appears to have motion in it. The question is what caused that motion.
Bad analogy. I was trained in wilderness first response. "Seeing" a broken leg isn't always easy to recognize without an X-ray. The pain the respondent indicates can be caused by other things. That's exactly where we are with your image. Like any doctor, I need more information to verify a hypothesis.
I think there's a misunderstanding about what I'm asking here. I'm not asking for anyone to say what caused the issue. I'm not asking for anyone to verify a hypothesis.
The way you're talking about the analogy really suggests that I haven't gotten my question across. Of course someone needs more than a quick look to verify or to be sure what has caused an injury. I'm asking about something much higher level than that. I'm talking about something even more fundamental than a doctor of first responder even forming a hypothesis. I'm talking about the basic knowledge a doctor has to know what to even consider when trying to first approach a problem.
My point is that if a person is complaining of leg pain there are a heck of a lot of things the doctor is going to immediately discount - not even think of, in fact - because they're totally inconsistent with the symptom. When the patient says "leg pain," the doctor is not going to consider a stomach ulcer. The doctor is not going to consider lung cancer. The doctor is not going to consider a broken finger. The symptoms are totally inconsistent with any of these things.
The doctor is instead going to start to think about broken bones in the leg, sprains, torn ligaments, perhaps blood clots, and various other things. There will need to be a more comprehensive examination and more tests to actually nail down which of these is the cause of the leg pain, but the doctor knew before even looking at the leg what sorts of things can be a cause of leg pain and which sorts of things cannot be a cause of leg pain.
I'm very simply asking "Is it possible for atmospheric distortion to cause an image that looks like this?"
"Is it possible for simple missed focus to look like this?"
"Is it possible for a decentered element to look like this?"
Put differently, my assumption when I have been getting some poor results during some recent trips to the location in question has been thermal distortion. This has been based on past experience with thermal distortion which led me to think it's the same phenomenon. However, when I saw that gull photo I really started to question this because I don't recall ever seeing atmospheric distortion result in such an effect before.
So my question is NOT "Did atmospheric distortion cause this?" My question is NOT "Is it likely atmospheric distortion caused this?"
Rather, my question very simply has been "Is this effect consistent with atmospheric distortion?" Or, "Does atmospheric distortion sometimes look this way?" Or, "Is it possible atmospheric distortion caused this, or would that be like saying a stomach ulcer caused foot pain?"
Or, to get even more general, "Is it possible to get an effect like this out of a properly functioning lens, or does an effect like this always indicate a hardware problem?"
Last edited:
Thom Hogan
Forum Pro
I would say no. Too many clues point elsewhere, and atmospheric distortion tends to be variable across the image, not continuous.
Yes.
No. A decentered element is going to cause side to side issues. I don't see any evidence of that here.
I would say no.
Again, no.
Funny thing is, there are connections that don't make sense, so I'd avoid that analogy. For instance, a single unexplained knee pain can be a predictor of prostate cancer. No one seems to know the connection, but the correlation is strong.
This is not a result I'd expect. However, we haven't eradicated all the other possibilities. The apparent motion on the pupil (not the highlight) suggests something is moving. The odd highlight on the eye shows something slightly different. The overall lack of detail and no obvious focus plane is another issue.
But getting to your "properly functioning lens" comment, the only ways that can be determined short of a test bench is:
- If you can do A/B comparisons with a known good lens. Same camera, same handling, same situation, same settings should deliver the same results.
- If you can eliminate all the other possibilities one by one (e.g. VR, focus motor, handling, settings, etc.).
Alright, well thanks for the comments.
I'm not sure what to think of it as of now. Based on comments on other forums I do wonder if this was missed focus, which opens up other questions for me about just how inconsistent the focus is behaving on this lens. On the other hand, the overall effect I've seen (not just from this photo but from the other few thousand I took) does look a lot like atmospherics to me, and I do have people in other forums saying they've experienced the effect seen here with atmospherics, but who knows?
For what it's worth, the lens does seem decent optically - not the best copy out there I'd imagine, but decent, and has focused perfectly fine in many cases. It's just leaving me unsure what's going on in some instances and especially when I have tried to take it out to test at distances >10m or so where there are questions of whether the lens is just that much worse at greater distances or if it's been distortion issues, which are also going to pop up more at longer distances.
Leonard Shepherd
Forum Pro
If the problems include sometimes missed focus, focus takes place at the sensor, part affected by camera settings, and part affected by the nature of the subject.
A lens is unlikely to have much of an effect on what goes on in the camera body, other than faster apertures when near the limits of AF detection provide more light for the AF system.
Longer focal lengths generally benefit from faster shutter speeds as they magnify subject movement relative to shorter focal lengths
Hmm, the idea that different lenses do no autofocus with different degrees of accuracy and precision contradicts the view of essentially everyone I've ever seen comment on the subject, including experts, including experts in this thread.
Leonard Shepherd
Forum Pro
Is it an idea - or is it largely reality?
AF takes place at the sensor with ML.
While the degree of image sharpness projected on the sensor can vary from lens to lens - the principle of AF on the sensor is the body focuses until it detects the point of highest contrast - which is also the point of highest resolution.
There was more scope for consistent front or back focus with AF-S lenses on the rare occasions the in lens AF motor did consistently stop at the signalled right focus distance.
There has always been scope for in camera AF (where focus starts) to wrongly calculate a subject distance, especially when the camera is aimed at a subject where Nikon caution AF may not be consistently reliable.
My view, shared with several others, is that on the one hand well over 90% of perceived AF issues are due to less than ideal testing methodology and on the other hand less than 10% are something else.
So far I have seen no images in this thread that for me clarify an equipment issue although I acknowledged early on that there could be.
Thom Hogan
Forum Pro
We need to correct some language for clarity, correct other.
In the mid-teens, it seems that Nikon got serious about AF-S consistency. New lenses were suddenly showing Canon-like standard deviations and zero centers. The AF-P lenses were even "tighter," as have been the Z System lenses.
The data used for AF is derived from all of the light at the image sensor for mirrorless cameras. It is obtained by a separate sensor for DSLRs from only some of the light that was passed to it via a mirror system.
That might be true for contrast detect AF, but for phase detect the determination is done differently. For cameras using ML-taught subject detection, there's yet another different layer of focus determination being done, which actually isn't very dependent upon maximum resolution.
As the folks at FoCal will tell you, this has varied considerably over the history of AF-S (and AF-P and Z) Nikkors. Back in the early D800 days a lot of us were trying to figure out how consistent the lenses were, because about a third of the original D800's had some of their focus sensors adjusted wrong at the factory. This led us to discover that the standard deviation for Nikkors was way higher than for Canon lenses, and also tilted away from zero (perfect focus) for many lenses. This got worse with the "fix" Nikon applied to D800 bodies, which started dialing in a -10 AF Fine Tune to the body itself.
In the mid-teens, it seems that Nikon got serious about AF-S consistency. New lenses were suddenly showing Canon-like standard deviations and zero centers. The AF-P lenses were even "tighter," as have been the Z System lenses.
This had to do with the larger phase detect cells on DSLRs as much as anything. The mirrorless cameras have way smaller detect areas, so many of the conditions Nikon used to warn constantly about just don't happen on a Z System camera. Yes, it's technically still possible to be less consistent in focus with really fine detail, but that now comes from the fact that Nikon is having to figure out which of the several hundred values it obtained for even the minimal size focus area is the correct one, not because the resolution isn't enough.
For a more comprehensive explanation of how the lens can impact AF accuracy, I can suggest Reikan's post on the topic here: https://blog.reikanfocal.com/2023/05/mirrorless-af-calibration-part-4-the-lens/
More simply, though, the general consensus is that lower grade lenses may use less precise motors and this can give a larger margin of error for the final focus position. I've not comprehensively tested this, but my impressions have been that I have observed more consistently accurate focus when photographing with my S line 85mm vs a lower grade lens of similar focal length.
Nikon also offers lens calibration for its Z series lenses, indicating that they think this matters, and various users report improved focus accuracy after having the service done on a lens. Reikan's piece explains what this actually means and how it is another way focus accuracy can vary from lens to lens.
Thom Hogan
Forum Pro
I would suggest it might have to do with speed more than precision, as even the lower cost steppers are pretty precise. The issue comes when any kind of tracking or reinterpretation (subject detection) starts to get involved and the camera starts to change its mind about where focus should be. The camera is evaluating at 120Hz, but is the stepper fast enough to do full on change of direction and reposition as fast as the camera? I don't know for sure, but I'd bet speed is part of the issue.
Hmm, interesting. I had frankly always thought this but had moved away from thinking so after being told - almost scolded - repeatedly that any lens (e.g., the 200-500, about which I heard this a lot) is fast enough to keep up with the new focus position if e.g. a BIF once it has initial acquisition.
ghostfox_1
Senior Member
- Messages
- 3,134
- Solutions
- 5
- Reaction score
- 6,168
I think on a short movement distance (and especially at longer distances) this is still true. Now, expecting something to keep up at mfd or from mfd to infinity quickly isn't something I'd do.
But unless I'm misinterpreting Thom here he seems to be suggesting that even for something like a stationary subject moving its head around or even just with changing light patterns causing the camera's AF to make an adjustment it may be that in some cases a lens' AF motor may not be keeping up and resulting in the effect of more focus misses than a higher end lens with a faster motor.
PHXAZCRAIG
Forum Pro
This sounds very much like what I think was happening to me last summer in Alaska. At one point an eagle glided overhead while I was holding my Z9 + 100-400. Naturally I took a quick sequence of shots hoping to 'get a good one'. To my surprise, about every 3rd or 4th image in the sequence is softer than those before and after. Do this look like a VR issue, or a focus-tracking issue, or a combo or the two or ...?
(Click on an image to see a large version.)
These random misses are consistent with my experience of using both the Z8 and Z9 with different lenses (180-200, 500PF, 800PF) - seems particularly susceptible to focus hunting (which is what I put it down to) when there are bright highlights in the scene. See others posts in various places relating to this same issue.I received my 180-600 about a month ago. Initially I was getting very poor shots until I discovered that like some other Nikon lenses (but unlike my 200-500) using the VR with higher shutter speeds seems to cause problems. I have now been getting pretty sharp shots with backyard birds where everything is close. Static test shots indoors have also been pretty good.
I have NOT been getting good shots when shooting at distances >10 meters or so out at the local park/pond. I had suspected this is due to atmospherics/thermal distorting until I got this shot. I immediately noticed the catch light is basically doubled up (you can see it more clearly when comparing to an almost identical shot without this problem, which I'll share below). The pupil also has a ghost/double image on it. The beak does as well. The left side of the head does, as does the top, though its harder to see.
This shot was from a burst. Two shots - so 0.20 seconds - before was this shot, which is more or less sharp:
I don't think of atmospherics as causing ghosting like this. What would cause ghosting like this? A badly misaligned element I suppose, but none of the elements moved in between these two shots. I am now wondering if what I have really been seeing at >10m is atmospheric distortion or if it's some other problem.
Any thoughts here?
Perhaps these highlights confuses the focussing or subject detection a little.
Do you experience this issue when the scene is more evenly lit - i.e. no bright reflections from the sun anywhere in the scene.
Last edited:
- Status
Similar threads
Keyboard shortcuts
- f
- Forum