Am I doing something wrong, or... sharpness/focus issues with D500

[SCoombs]

For the past few weeks I've been digging very deeply into a problem I've had with sharpness on my D500 and Nikkor 200-500 f/5.6. I've always been somewhat disappointed with the results I get from this pair but have had little success improving it with fine tuning. I know the lens isn't going to give the same kinds of results as high end prime, but it also ought to do better than something like this - representative of much of what I've gotten from it (some of the photos below are edited for general color, exposure, etc., while others are not):



After dozens of hours of testing and hundreds of test shots over the past two weeks or so I feel like I've actually gotten it to the point of getting good results, for instance here:



It's like night and day, and they're even similar in terms of crop size. Or this:



It's not great, but it's pretty good. Photos of humans can be particularly sharp to my eye:



Yet I am finding myself often still getting photos that are really lacking in sharpness or even downright blurry, even by all indications they ought to be much better. For instance:



The bird is downright blurry here. Yet, it's 1/2000 of a second, it's f/8 so the depth of field is larger, and the lighting was decent, with a relatively low ISO as a consequence. Now one might say that this is just one photo and maybe the focus point was a bit off or something, but bear in mind that it's just one example I'm sharing to illustrate a large trend.
Similarly, this duckling:



It's not horrible horrible, but it's also definitely below the level of sharpness I'd expect for this combo and far below some of the better examples above. The duck is pretty small in the frame so I'm not expecting perfect detail, but you don't even have to zoom in to 50% to start to see that it's not really sharp.

Some others which turn out okay after a lot of noise removal suffer a lack of detail and much heavier noise than I'd expect given the conditions. However, that's not the primary concern of this post. Remember the slightly blurry gull a couple of photos up? Here's the very next photo I took, just a few seconds later:



This one's still noisier than I think I'd expect for the conditions - maybe I'm wrong? - but it's clearly much sharper than the other one, and in all the same conditions with the same settings. I have another version of this I've done some denoising on and it's a reasonably decent photo, at least as far as IQ is concerned. What's the difference?

About 3 meters - according to the EXIF data, anyways. The first one reports a focus distance of 17.8 meters while the second of 15 meters. Noticing this, I started to go through a bunch of photos and found that as a general rule, anything over ~15 meters or so seems to be fairly lacking in sharpness, whereas photos under that distance generally displays an appropriate sharpness for their conditions. It's also important to note that the size of the subject in the frame doesn't seem to matter - the trend is about the same whether a subject is large in the frame or small, the only difference being how much room there is to see additional detail by zooming with larger subjects.

I can think of at least three causes for this - maybe you can think of something I'm missing.

1) The lens is simply not sharp at those focusing distances.

2) The AF fine tuning is accurate within a certain range and needs to be different for subjects at greater distances.

3) There is something wrong or some limitation with the camera's AF system.

As a first step to trying to figure this out, I have looked at a few photos taken with this lens on my Z7ii mirrorless camera. I have tended to avoid it for wildlife but have used it some, especially as I've been trying to sort my D500 out and because I am also considering selling them both to buy a Z8 and wanted to use the Z7ii to get at least a very rough ballpark idea of the AF and IQ that I might expect on the Z8 (my idea being that while the Z7ii has a hard time getting focus on some quickly moving subjects, when it does that AF seems very accurate and so the Z8 would only be better). I found this one to start. It reports 17.8 meters - the same as the worse of the gull photos I shared from the D500 but clearly much sharper. This has been denoised, but the sharpness was there in the original.



Here's one in which the subject was pretty small in the frame, so it's been cropped quite a bit. I wouldn't say it's perfect and it may be ever so slightly front focused (maybe on the water splashes, actually), but it's hard for me to say whether that's the same issue being discussed or it's just a miss from the Z7ii's AF system in trying to do something that it doesn't really excel at:



Either way, for 30 meters it seems better than any of the >15m D500 shots. I feel like this one at least looks relatively sharp when you aren't zooming (and remember it's already cropped a lot to begin with) whereas so many of the D500 ones are looking fuzzy even before you start getting closer.



Here's one that's cropped quite a lot. There isn't room to zoom in really at all and see more details and for me at least that makes it slightly harder to judge sharpness but I think it's actually pretty good, and this at about 24 meters.



Another, this one at 20 meters and pretty significantly cropped (focus is on the mother; I don't expect the duckling to be in good focus given the DoF):



Finally, at 17.8 meters but what I'd call a pretty extreme, 100% crop:



Unfortunately, I don't think I have quite enough quantity or quality of examples to really say for sure that the mirrorless is getting sharper, more in focus shots at these greater distances, and truth be told I don't feel like I have enough poor ones at greater distances to say with absolute confidence that this is a problem on the D500, though as you can see there is thus far a notable trend here. In particular, I'd like to get some shots of some larger and more contrasty (vs. a white bird, which a lot of these are) subjects at greater distances with both of these cameras. Nevertheless, what do you think about the general issue and the trend I've tried to illustrate with a few selections here? What would you recommend I do by way of testing to try to nail down what's going on with some measure of confidence?



I do feel like I need more examples to say for sure, but I feel at least mostly confident in ruling out possibility #1 here, that the lens isn't sharp at greater focus distances. Again, I'm sure it isn't the sharpest lens on the planet, but some of the Z shots do seem to give reason to think that it is possible to get at least decent looking photos out of this lens from these greater distances. That leaves AF fine tuning or a camera problem. Any thoughts on trying to figure out which it is?



Right now I'm looking to upgrade my setup, most likely selling the Z7ii to hep fund it. One possibility with pretty clear advantages would be to keep the D500 and add some more useful glass. For example, I would like a 70-200 f/2.8. And, I love my D500 and the photos I get out of it when they're sharp. The other option is to sell both and get a Z8, putting off more glass for now. This is the "less bang for the buck" option as it doesn't add any options for lenses, though it does pave the way for focusing on the Z mount in the future. If the problem here is the D500 itself, then unfortunately I think the Z8 is the way I'd have to go. I also think that an AF fine tune issue might mean the same thing. After all, if I need to tune the lens significantly differently depending on whether the subject is within a certain distance or not it's a fairly significant problem. On the other hand, if there's some middle ground that gives me sharp photos at all distances, then I can more seriously consider keeping the D500 - which is what I ideally would prefer.

 

[Richard James]

For the past few weeks I've been digging very deeply into a problem I've had with sharpness on my D500 and Nikkor 200-500 f/5.6. I've always been somewhat disappointed with the results I get from this pair but have had little success improving it with fine tuning. I know the lens isn't going to give the same kinds of results as high end prime, but it also ought to do better than something like this - representative of much of what I've gotten from it (some of the photos below are edited for general color, exposure, etc., while others are not):

I can think of at least three causes for this - maybe you can think of something I'm missing.

1) The lens is simply not sharp at those focusing distances.

2) The AF fine tuning is accurate within a certain range and needs to be different for subjects at greater distances.

3) There is something wrong or some limitation with the camera's AF system.

I would suggest:

1- trying to compare images with different image magnification (crop) is not giving you useful comparisons. They all need to be the same crop.

2 - noise reduction reduces sharpness, and should not be an issue at the low ISO for many of your sample.

3 - focus fine tuning is subject-distance (and possibly aperture) specific. It is always a compromise.

4 - trying to compare sharpness only with the lens wide-open is difficult (shallower DoF confuses the issue). You need to compare with the same subject at different apertures.

5 - to help you, use a focus-testing specific target and do testing on a tripod at various image distances and apertures, concentrate on your "usual range" with some outliers.

Use of a tripod and your focus mode have an impact, especially on repeatability in both testing and real life.

And finally, if you are not "filling the frame' with the subject you will loose image quality. Yes, I know, you don't want to go swimming (or fall off a cliff, or into an alligator-filled swamp) to get closer, but...

Richard

 

[SCoombs]

For the past few weeks I've been digging very deeply into a problem I've had with sharpness on my D500 and Nikkor 200-500 f/5.6. I've always been somewhat disappointed with the results I get from this pair but have had little success improving it with fine tuning. I know the lens isn't going to give the same kinds of results as high end prime, but it also ought to do better than something like this - representative of much of what I've gotten from it (some of the photos below are edited for general color, exposure, etc., while others are not):

I can think of at least three causes for this - maybe you can think of something I'm missing.

1) The lens is simply not sharp at those focusing distances.

2) The AF fine tuning is accurate within a certain range and needs to be different for subjects at greater distances.

3) There is something wrong or some limitation with the camera's AF system.

I appreciate the reply. I can't say that I understand the reasoning behind all of these suggestions and would appreciate some elaboration.

I would suggest:

1- trying to compare images with different image magnification (crop) is not giving you useful comparisons. They all need to be the same crop.

Why should this matter? Cropping doesn't change anything about the sharpness of the image - only what the viewer sees "from a distance." I have cropped the photos above either for artistic composition or just to get rid of the part of the image that is not expected to be in focus anyways, but if we zoom in to any of the photos at100% we'll see an exact one to one comparison, and if we don't zoom in then we won't necessarily be able to see the degree of sharpness anyways. Would you suggest that I had cropped all of them to 100% before posting?

2 - noise reduction reduces sharpness, and should not be an issue at the low ISO for many of your sample.

I have not used noise reduction on any of the images I posted to illustrate poor sharpness (well, actually I probably did for the first woodpecker image since it is quite old and wasn't produced for this analysis. Here it is with no NR and we can see it isn't a whole lot different.) What's more, the flip side of this is that noise also reduces apparent sharpness so there is a balancing act. Many images I have look much sharper after NR. For instance, here's the "good" gull picture with NR applied: it looks much sharper.









This could tie into the other issue I raised of images I am getting from the D500 having far more noise than I would expect for the lighting and ISO, but for now I'll just leave it that my samples meant to illustrate poor sharpness have not had any NR applied beyond perhaps the very minimal default that Lightroom applies.

3 - focus fine tuning is subject-distance (and possibly aperture) specific. It is always a compromise.

I understand this of course. The question is always how much of a difference it makes in any particular case. You might have a lens/camera combo which requires very little tuning at any distance with another which requires drastically different tuning at different distances. I'm trying to determine if I'm simply seeing an example of the latter case or something else - or if I'm seeing an example of that difference where an acceptable compromise can be found. If getting the nice sharpness of the clean woodpecker shot requires the very poor sharpness of the gull, that's not really workable. If there's a better middle ground, that's obviously good, though I've spent hundreds of photos looking for one.

4 - trying to compare sharpness only with the lens wide-open is difficult (shallower DoF confuses the issue). You need to compare with the same subject at different apertures.

Most of these do not have the lens wide open.

5 - to help you, use a focus-testing specific target and do testing on a tripod at various image distances and apertures, concentrate on your "usual range" with some outliers.

I've done lots of tripod testing using many methods with everything from focus targets to different subjects. Up until this recently I had not been able to come away with any consistent results. I've also used FoCal and sometimes it gives AF suggestions within a given range of numbers, while other times it complains that it can't get consistent results, either.

Use of a tripod and your focus mode have an impact, especially on repeatability in both testing and real life.

And finally, if you are not "filling the frame' with the subject you will loose image quality. Yes, I know, you don't want to go swimming (or fall off a cliff, or into an alligator-filled swamp) to get closer, but...

Richard

Several of the examples I posted are about as full a frame as one might expect. For instance, the two gull photos are not cropped. Could the bird be larger in the frame? I suppose a little, but then of course people would ignore my questions about sharpness and tell me the composition is bad (I don't mean to sound cynical, but this is literally what has happened when I have posted "fill the frame" photos in the past with similar questions).

 

[capanikon]

Hrm dunno.

I've had AF problems before, with my D2h. I sent the camera to Nikon and it came back perfect. The D2h has no AF adjustment ability in the menu system.

Perhaps send the body and lens to Nikon for evaluation/adjustment.

I wonder if VR is affecting IQ adversely?

I've casually observed that multi-point AF doesn't always work as well for me as single-point AF. Thus I typically ignore multi-point AF modes. I usually use the center AF sensor or ones close to the center and typically avoid using the outer AF sensors.

My D3 has had bang-on perfect AF so far. My old D200 had perfect AF.

My D300 was problematic.

My D2h was problematic, but Nikon made it perfect.

My D1, D100 and a different, earlier D2h all had perfect AF.

My F4s, N8008, N8008s, and N90s all had perfect AF.

I like to shoot a book or newspaper page at a 45 degree angle with camera mounted on tripod for AF tests. There's also a special chart for AF accuracy you can download and print. e.g. http://regex.info/blog/photo-tech/focus-chart

--
irc.libera.chat #photogeeks

 

[SCoombs]

Hrm dunno.

I've had AF problems before, with my D2h. I sent the camera to Nikon and it came back perfect. The D2h has no AF adjustment ability in the menu system.

Perhaps send the body and lens to Nikon for evaluation/adjustment.

I wonder if VR is affecting IQ adversely?

I've casually observed that multi-point AF doesn't always work as well for me as single-point AF. Thus I typically ignore multi-point AF modes. I usually use the center AF sensor or ones close to the center and typically avoid using the outer AF sensors.

My D3 has had bang-on perfect AF so far. My old D200 had perfect AF.

My D300 was problematic.

My D2h was problematic, but Nikon made it perfect.

My D1, D100 and a different, earlier D2h all had perfect AF.

My F4s, N8008, N8008s, and N90s all had perfect AF.

I like to shoot a book or newspaper page at a 45 degree angle with camera mounted on tripod for AF tests. There's also a special chart for AF accuracy you can download and print. e.g. http://regex.info/blog/photo-tech/focus-chart

Thanks for the thoughts! I have tried turning the VR off since I know some people think that Nikon cameras suffer a loss of sharpness if you use VR with higher shutter speeds and while I'll admit I haven't tested this very thoroughly, I haven't noticed much difference in what I have done.

I've also generally tried to avoid things like group AF when the subject affords it (for things like BIF I feel like I really don't have much choice if I'm going to actually acquire focus in the first place). Especially once I started testing this, I've tried to use more single point modes at least for some key shots to look at and examine later.

it's funny you mention the photo method and the chart on that page. The last time I posted about this issue and explained focus testing I've done (including a lot with similar charts and methods to that site's) I had several people jump on it as the wrong way to do focus testing and I feel like after that they started assuming I knew very little about photography which made it hard to have much relevant discussion. :/

 

[SCoombs]

Doing some more testing, I have hit something that has me confused. The Exif data for all of these photos contains a depth of field estimate based on the focus distance, aperture, etc. For most of the photos, the exif DoF lines up very well with what a standard DoF calculator tells me. However, for photos I have taken at ~40 meters away as part of the tests, it's telling me the DoF is 9 meters - far more than the 2 meters that the standard formulas/calculators point to. I wonder why this is.

 

[davethebirder]

Doing some more testing, I have hit something that has me confused. The Exif data for all of these photos contains a depth of field estimate based on the focus distance, aperture, etc. For most of the photos, the exif DoF lines up very well with what a standard DoF calculator tells me. However, for photos I have taken at ~40 meters away as part of the tests, it's telling me the DoF is 9 meters - far more than the 2 meters that the standard formulas/calculators point to. I wonder why this is.

All the shots that are not sharp are due to front or back focus ..... overall I'm not seeing any special tendance either way, which to me implies that it could be you choice of focusing modes ... for best results shooting slow moving birds I would choose single point on the eye ....... BIF panning group mode (maybe a dynamic mode 9/21/53) ....... I also get good success when using auto area mode for approaching birds or group mode ......

All AF configured to the various configurable button choices for use on the fly ........ I use single point/ auto area mode and group as a basis ....

I would also go mirror up in the menu and give a blast with a rocket blower under the mirror just in case there a bit of dust on the prisms ....... if you want to check for focus consistency, use just single focus point on almost static subjects at various distances, I find I get 80% perfect focus minimum and up to 100% ..........

I've never looked at the exif data in order to fathom out the whys and wherefores of AF issues btw.

https://www.flickr.com/photos/124690178@N08/

 

[Richard James]

I appreciate the reply. I can't say that I understand the reasoning behind all of these suggestions and would appreciate some elaboration.

I would suggest:

1- trying to compare images with different image magnification (crop) is not giving you useful comparisons. They all need to be the same crop.

Why should this matter? Cropping doesn't change anything about the sharpness of the image - only what the viewer sees "from a distance." ... but if we zoom in to any of the photos at100% we'll see an exact one to one comparison,

It is all about the pixels, if the duck in image one has 2000 px on it and the duck in image two only has 1000 px on it, then it will necessarily have less resolution which impacts apparent sharpness.

Cropping does not change the pixels, it simply reflects what the sensor captures, which is an image that has less detail.

Resizing does not solve the problem, it generates interpolated pixels...

2 - noise reduction reduces sharpness, and should not be an issue at the low ISO for many of your sample.

I have not used noise reduction on any of the images I posted to illustrate poor sharpness

You did mention it, so I commented.

Noise reduction, using a few basic tools only, potentially reduces sharpness, but modern NR tools may use techniques to "restore" apparent sharpness ( ie light, maybe selective, "sharpening").

Many images I have look much sharper after NR.

"Much" to me implies heavier sharpening has been applied in the NR process.

At the low ISO levels in the images you posted, noise should not be readily apparent at normal (print) size/viewing distance with an uncropped image, and even less on a monitor (unless you are looking at a 40" monitor <G>).

Cropping then enlarging may make "acceptable" noise intrusive. In which case you are stuck. But this does not mean that there is a problem.

This could tie into the other issue I raised of images I am getting from the D500 having far more noise than I would expect for the lighting and ISO,

"Far more" compared to what (equivalent sensor) camera?

A subject 20mm wide on a D500 sensor should look very similar to the same image occupying 20mm on a D850 as the sensors are quite similar (px density anyway).

OTOH, if shot on my D3x, I would expect a big difference, different px density, different noise char. etc.

3 - focus fine tuning is subject-distance (and possibly aperture) specific. It is always a compromise.

I understand this of course. The question is always how much of a difference it makes in any particular case. You might have a lens/camera combo which requires very little tuning at any distance with another which requires drastically different tuning at different distances.

Yes, every combo is different. Each body and each lens sits somewhere within their separate design tolerances. If your combo is at opposite ends, you may have an issue. However, you may have to be persistent with Nikon if they simply say "in spec"...

4 - trying to compare sharpness only with the lens wide-open is difficult (shallower DoF confuses the issue). You need to compare with the same subject at different apertures.

Most of these do not have the lens wide open.

Good... But you need same subject/same aperture/same distance etc. comparisons.

5 - to help you, use a focus-testing specific target and do testing on a tripod at various image distances and apertures, concentrate on your "usual range" with some outliers.

I've done lots of tripod testing using many methods with everything from focus targets to different subjects. Up until this recently I had not been able to come away with any consistent results. I've also used FoCal and sometimes it gives AF suggestions within a given range of numbers, while other times it complains that it can't get consistent results, either.

Yes, I have had inconsistent results as well...

Use of a tripod and your focus mode have an impact, especially on repeatability in both testing and real life.

As another poster here suggested, make sure VR is off if on a tripod (unless your lens has a "tripod VR" mode). Also off if the shutter speed is over about 1/500 as this apparently conflicts with the sampling rate for the VR system

And finally, if you are not "filling the frame' with the subject you will loose image quality. Yes, I know, you don't want to go swimming (or fall off a cliff, or into an alligator-filled swamp) to get closer, but...

Several of the examples I posted are about as full a frame as one might expect.

See my comment above about pixels under the subject... that is critical.

For any realistic comparison all variables need to be fixed except the one at a time you are changing. In this case px under the subject is critical.

For example, use base ISO for all images (eliminates Noise issue), use the same aperture (removes DoF as an issue).

You can always do another series at a different ISO/Aperture (etc) to see what impact it has.

For instance, the two gull photos are not cropped..... and tell me the composition is bad (I don't mean to sound cynical, but this is literally what has happened when I have posted "fill the frame" photos in the past with similar questions).

We are talking about testing to identify if the problem is real, so composition is not relevant.

Hope this helps.

Richard

 

[SCoombs]

Doing some more testing, I have hit something that has me confused. The Exif data for all of these photos contains a depth of field estimate based on the focus distance, aperture, etc. For most of the photos, the exif DoF lines up very well with what a standard DoF calculator tells me. However, for photos I have taken at ~40 meters away as part of the tests, it's telling me the DoF is 9 meters - far more than the 2 meters that the standard formulas/calculators point to. I wonder why this is.

All the shots that are not sharp are due to front or back focus ..... overall I'm not seeing any special tendance either way, which to me implies that it could be you choice of focusing modes ... for best results shooting slow moving birds I would choose single point on the eye ....... BIF panning group mode (maybe a dynamic mode 9/21/53) ....... I also get good success when using auto area mode for approaching birds or group mode ......

I agree that there does not seem to be a consistent front or back tendency, however I do not believe this is due to the focus mode selection. For one thing, in almost all of these cases (other than BiF, for which I feel like I must use group AF) I am using single point or dynamic (which on a still subject basically is single point) with a very clear LOS to the subjects and never having the point drift off of the subject. Since starting to investigate this issue, I have especially been precise about not using multi-point focus modes to be sure to rule out as many variables as possible.

Please also note that I have extensively worked at tuning the focus and done literally hundreds of tests in controlled conditions with tripods and so forth and found no pattern as to whether it front or back focuses. Sometimes it's one, sometimes the other. I have also used FoCal which at times works but other times says the AF is behaving erratically and does not give consistent results.

Just tonight I did yet another test - a less "scientific" one but one with striking results nonetheless. I had a woodpecker which sat in the same place on a feeder for several minutes so I took a bunch of shots - some with the AF fine tuning on to the most effective value I've found so far (+8) and some with it off. I then loaded them in Lightroom and without knowing which was which I picked the most in focus ones. Afterwards I looked at the EXIF data to see which were which and there was no real pattern. Some I had picked as sharpest were with the fine tuning while others were without, but there's really no discernable difference between the two groups. This is all with a pretty shallow depth of field of about .6 inches.

All AF configured to the various configurable button choices for use on the fly ........ I use single point/ auto area mode and group as a basis ....

I would also go mirror up in the menu and give a blast with a rocket blower under the mirror just in case there a bit of dust on the prisms ....... if you want to check for focus consistency, use just single focus point on almost static subjects at various distances, I find I get 80% perfect focus minimum and up to 100% ..........

As noted above, I've done many, many such tests. I've also tried cleaning as you suggest.

I've never looked at the exif data in order to fathom out the whys and wherefores of AF issues btw.

https://www.flickr.com/photos/124690178@N08/

 

[SCoombs]

I appreciate the reply. I can't say that I understand the reasoning behind all of these suggestions and would appreciate some elaboration.

I would suggest:

1- trying to compare images with different image magnification (crop) is not giving you useful comparisons. They all need to be the same crop.

Why should this matter? Cropping doesn't change anything about the sharpness of the image - only what the viewer sees "from a distance." ... but if we zoom in to any of the photos at100% we'll see an exact one to one comparison,

It is all about the pixels, if the duck in image one has 2000 px on it and the duck in image two only has 1000 px on it, then it will necessarily have less resolution which impacts apparent sharpness.

Cropping does not change the pixels, it simply reflects what the sensor captures, which is an image that has less detail.

Resizing does not solve the problem, it generates interpolated pixels...

As far as I know I haven't resized the images. I have cropped them, so it's as if I took a physical photo and cut the edges off to produce something with a smaller surface area, but without enlarging or shrinking the remaining image.

2 - noise reduction reduces sharpness, and should not be an issue at the low ISO for many of your sample.

I have not used noise reduction on any of the images I posted to illustrate poor sharpness

You did mention it, so I commented.

Noise reduction, using a few basic tools only, potentially reduces sharpness, but modern NR tools may use techniques to "restore" apparent sharpness ( ie light, maybe selective, "sharpening").

Many images I have look much sharper after NR.

"Much" to me implies heavier sharpening has been applied in the NR process.

At the low ISO levels in the images you posted, noise should not be readily apparent at normal (print) size/viewing distance with an uncropped image, and even less on a monitor (unless you are looking at a 40" monitor <G>).

Cropping then enlarging may make "acceptable" noise intrusive. In which case you are stuck. But this does not mean that there is a problem.

I'd agree at the low ISO levels the noise should not be readily apparent, but it certainly has been, which is the other issue I have noted but not really dug into much. For instance, looking at that original out of focus gull, the noise is obvious to me in a way that I don't expect for such a low ISO - at least not one where the lighting was also good. You can still get noise at a low ISO in parts of an image which did not receive enough light during exposure, but that's not the case in the noisy parts of the image here.

This could tie into the other issue I raised of images I am getting from the D500 having far more noise than I would expect for the lighting and ISO,

"Far more" compared to what (equivalent sensor) camera?

A subject 20mm wide on a D500 sensor should look very similar to the same image occupying 20mm on a D850 as the sensors are quite similar (px density anyway).

OTOH, if shot on my D3x, I would expect a big difference, different px density, different noise char. etc.

3 - focus fine tuning is subject-distance (and possibly aperture) specific. It is always a compromise.

I understand this of course. The question is always how much of a difference it makes in any particular case. You might have a lens/camera combo which requires very little tuning at any distance with another which requires drastically different tuning at different distances.

Yes, every combo is different. Each body and each lens sits somewhere within their separate design tolerances. If your combo is at opposite ends, you may have an issue. However, you may have to be persistent with Nikon if they simply say "in spec"...

4 - trying to compare sharpness only with the lens wide-open is difficult (shallower DoF confuses the issue). You need to compare with the same subject at different apertures.

Most of these do not have the lens wide open.

Good... But you need same subject/same aperture/same distance etc. comparisons.

5 - to help you, use a focus-testing specific target and do testing on a tripod at various image distances and apertures, concentrate on your "usual range" with some outliers.

I've done lots of tripod testing using many methods with everything from focus targets to different subjects. Up until this recently I had not been able to come away with any consistent results. I've also used FoCal and sometimes it gives AF suggestions within a given range of numbers, while other times it complains that it can't get consistent results, either.

Yes, I have had inconsistent results as well...

Use of a tripod and your focus mode have an impact, especially on repeatability in both testing and real life.

As another poster here suggested, make sure VR is off if on a tripod (unless your lens has a "tripod VR" mode). Also off if the shutter speed is over about 1/500 as this apparently conflicts with the sampling rate for the VR system

Yes, I have tried shutting the VR off at higher shutter speeds but haven't noticed much difference. Unfortunately another issue - unrelated here since we're talking about speeds of 1/2000+, but still something I have to contend with - is that my hands are apparently so shaky that sometimes I need VR to eliminate motion blur at relatively low shutter speeds (e.g., 1/600 for a 250mm lens), even using good technique, arms tucked in, all that.

And finally, if you are not "filling the frame' with the subject you will loose image quality. Yes, I know, you don't want to go swimming (or fall off a cliff, or into an alligator-filled swamp) to get closer, but...

Several of the examples I posted are about as full a frame as one might expect.

See my comment above about pixels under the subject... that is critical.

For any realistic comparison all variables need to be fixed except the one at a time you are changing. In this case px under the subject is critical.

For example, use base ISO for all images (eliminates Noise issue), use the same aperture (removes DoF as an issue).

You can always do another series at a different ISO/Aperture (etc) to see what impact it has.

For instance, the two gull photos are not cropped..... and tell me the composition is bad (I don't mean to sound cynical, but this is literally what has happened when I have posted "fill the frame" photos in the past with similar questions).

We are talking about testing to identify if the problem is real, so composition is not relevant.

Hope this helps.

Richard

Oh, I agree composition is irrelevant to the problem I'm trying to sort out! You're absolutely right. Unfortunately, I've learned that on this forum (and I suppose photography forums in general) I often need to worry about that anyways since too many users will hone in on whatever deficiency they can call out in a person's photos and somehow make the discussion about that rather than the question at hand.

 

[davethebirder]

Doing some more testing, I have hit something that has me confused. The Exif data for all of these photos contains a depth of field estimate based on the focus distance, aperture, etc. For most of the photos, the exif DoF lines up very well with what a standard DoF calculator tells me. However, for photos I have taken at ~40 meters away as part of the tests, it's telling me the DoF is 9 meters - far more than the 2 meters that the standard formulas/calculators point to. I wonder why this is.

All the shots that are not sharp are due to front or back focus ..... overall I'm not seeing any special tendance either way, which to me implies that it could be you choice of focusing modes ... for best results shooting slow moving birds I would choose single point on the eye ....... BIF panning group mode (maybe a dynamic mode 9/21/53) ....... I also get good success when using auto area mode for approaching birds or group mode ......

I agree that there does not seem to be a consistent front or back tendency, however I do not believe this is due to the focus mode selection. For one thing, in almost all of these cases (other than BiF, for which I feel like I must use group AF) I am using single point or dynamic (which on a still subject basically is single point) with a very clear LOS to the subjects and never having the point drift off of the subject. Since starting to investigate this issue, I have especially been precise about not using multi-point focus modes to be sure to rule out as many variables as possible.

Please also note that I have extensively worked at tuning the focus and done literally hundreds of tests in controlled conditions with tripods and so forth and found no pattern as to whether it front or back focuses. Sometimes it's one, sometimes the other. I have also used FoCal which at times works but other times says the AF is behaving erratically and does not give consistent results.

Just tonight I did yet another test - a less "scientific" one but one with striking results nonetheless. I had a woodpecker which sat in the same place on a feeder for several minutes so I took a bunch of shots - some with the AF fine tuning on to the most effective value I've found so far (+8) and some with it off. I then loaded them in Lightroom and without knowing which was which I picked the most in focus ones. Afterwards I looked at the EXIF data to see which were which and there was no real pattern. Some I had picked as sharpest were with the fine tuning while others were without, but there's really no discernable difference between the two groups. This is all with a pretty shallow depth of field of about .6 inches.

All AF configured to the various configurable button choices for use on the fly ........ I use single point/ auto area mode and group as a basis ....

I would also go mirror up in the menu and give a blast with a rocket blower under the mirror just in case there a bit of dust on the prisms ....... if you want to check for focus consistency, use just single focus point on almost static subjects at various distances, I find I get 80% perfect focus minimum and up to 100% ..........

As noted above, I've done many, many such tests. I've also tried cleaning as you suggest.

I've never looked at the exif data in order to fathom out the whys and wherefores of AF issues btw.

https://www.flickr.com/photos/124690178@N08/

I have two D500s and have rejected another due to AF issues (returned exchanged) ..... I had the AF'S 500mm F4D and have the Tamron G2 plus the AF-S 600F4E VR FL plus the TC 14E III all of which require no AF fine tune including on the D750 ......... (even a Pentax and an older Sigma 150-500mm screw drive lens needed no AFFT) .... yet most of my shorter FL lenses require some AFFT or another ...... telephotos usually require little to none if they are no issues ..

so I would be inclined to consider sending both as a pair to Nikon because after all you have written, I see nothing that faults either your testing methods/ determination or your shooting technique ...... I'm not saying I get no misses by any means, but apart from erratic BIF images and the odd situation, the D500 turns in reliably sharp images!

https://www.flickr.com/photos/124690178@N08/

 

[SCoombs]

Doing some more testing, I have hit something that has me confused. The Exif data for all of these photos contains a depth of field estimate based on the focus distance, aperture, etc. For most of the photos, the exif DoF lines up very well with what a standard DoF calculator tells me. However, for photos I have taken at ~40 meters away as part of the tests, it's telling me the DoF is 9 meters - far more than the 2 meters that the standard formulas/calculators point to. I wonder why this is.

All the shots that are not sharp are due to front or back focus ..... overall I'm not seeing any special tendance either way, which to me implies that it could be you choice of focusing modes ... for best results shooting slow moving birds I would choose single point on the eye ....... BIF panning group mode (maybe a dynamic mode 9/21/53) ....... I also get good success when using auto area mode for approaching birds or group mode ......

I agree that there does not seem to be a consistent front or back tendency, however I do not believe this is due to the focus mode selection. For one thing, in almost all of these cases (other than BiF, for which I feel like I must use group AF) I am using single point or dynamic (which on a still subject basically is single point) with a very clear LOS to the subjects and never having the point drift off of the subject. Since starting to investigate this issue, I have especially been precise about not using multi-point focus modes to be sure to rule out as many variables as possible.

Please also note that I have extensively worked at tuning the focus and done literally hundreds of tests in controlled conditions with tripods and so forth and found no pattern as to whether it front or back focuses. Sometimes it's one, sometimes the other. I have also used FoCal which at times works but other times says the AF is behaving erratically and does not give consistent results.

Just tonight I did yet another test - a less "scientific" one but one with striking results nonetheless. I had a woodpecker which sat in the same place on a feeder for several minutes so I took a bunch of shots - some with the AF fine tuning on to the most effective value I've found so far (+8) and some with it off. I then loaded them in Lightroom and without knowing which was which I picked the most in focus ones. Afterwards I looked at the EXIF data to see which were which and there was no real pattern. Some I had picked as sharpest were with the fine tuning while others were without, but there's really no discernable difference between the two groups. This is all with a pretty shallow depth of field of about .6 inches.

All AF configured to the various configurable button choices for use on the fly ........ I use single point/ auto area mode and group as a basis ....

I would also go mirror up in the menu and give a blast with a rocket blower under the mirror just in case there a bit of dust on the prisms ....... if you want to check for focus consistency, use just single focus point on almost static subjects at various distances, I find I get 80% perfect focus minimum and up to 100% ..........

As noted above, I've done many, many such tests. I've also tried cleaning as you suggest.

I've never looked at the exif data in order to fathom out the whys and wherefores of AF issues btw.

https://www.flickr.com/photos/124690178@N08/

I have two D500s and have rejected another due to AF issues (returned exchanged) ..... I had the AF'S 500mm F4D and have the Tamron G2 plus the AF-S 600F4E VR FL plus the TC 14E III all of which require no AF fine tune including on the D750 ......... (even a Pentax and an older Sigma 150-500mm screw drive lens needed no AFFT) .... yet most of my shorter FL lenses require some AFFT or another ...... telephotos usually require little to none if they are no issues ..

so I would be inclined to consider sending both as a pair to Nikon because after all you have written, I see nothing that faults either your testing methods/ determination or your shooting technique ...... I'm not saying I get no misses by any means, but apart from erratic BIF images and the odd situation, the D500 turns in reliably sharp images!

https://www.flickr.com/photos/124690178@N08/

Considering that I am able to get noticeably sharper photos with the same lens on my Z camera, I am inclined to think that the lens itself is okay. Would that be reasonable, or is there reason to think otherwise?

For instance,



















Of course, I can still get photos like this out of the D500 combo. I just can't quite sort out why too often, when everything looks right, I don't.

 

[davethebirder]

Doing some more testing, I have hit something that has me confused. The Exif data for all of these photos contains a depth of field estimate based on the focus distance, aperture, etc. For most of the photos, the exif DoF lines up very well with what a standard DoF calculator tells me. However, for photos I have taken at ~40 meters away as part of the tests, it's telling me the DoF is 9 meters - far more than the 2 meters that the standard formulas/calculators point to. I wonder why this is.

All the shots that are not sharp are due to front or back focus ..... overall I'm not seeing any special tendance either way, which to me implies that it could be you choice of focusing modes ... for best results shooting slow moving birds I would choose single point on the eye ....... BIF panning group mode (maybe a dynamic mode 9/21/53) ....... I also get good success when using auto area mode for approaching birds or group mode ......

I agree that there does not seem to be a consistent front or back tendency, however I do not believe this is due to the focus mode selection. For one thing, in almost all of these cases (other than BiF, for which I feel like I must use group AF) I am using single point or dynamic (which on a still subject basically is single point) with a very clear LOS to the subjects and never having the point drift off of the subject. Since starting to investigate this issue, I have especially been precise about not using multi-point focus modes to be sure to rule out as many variables as possible.

Please also note that I have extensively worked at tuning the focus and done literally hundreds of tests in controlled conditions with tripods and so forth and found no pattern as to whether it front or back focuses. Sometimes it's one, sometimes the other. I have also used FoCal which at times works but other times says the AF is behaving erratically and does not give consistent results.

Just tonight I did yet another test - a less "scientific" one but one with striking results nonetheless. I had a woodpecker which sat in the same place on a feeder for several minutes so I took a bunch of shots - some with the AF fine tuning on to the most effective value I've found so far (+8) and some with it off. I then loaded them in Lightroom and without knowing which was which I picked the most in focus ones. Afterwards I looked at the EXIF data to see which were which and there was no real pattern. Some I had picked as sharpest were with the fine tuning while others were without, but there's really no discernable difference between the two groups. This is all with a pretty shallow depth of field of about .6 inches.

All AF configured to the various configurable button choices for use on the fly ........ I use single point/ auto area mode and group as a basis ....

I would also go mirror up in the menu and give a blast with a rocket blower under the mirror just in case there a bit of dust on the prisms ....... if you want to check for focus consistency, use just single focus point on almost static subjects at various distances, I find I get 80% perfect focus minimum and up to 100% ..........

As noted above, I've done many, many such tests. I've also tried cleaning as you suggest.

I've never looked at the exif data in order to fathom out the whys and wherefores of AF issues btw.

https://www.flickr.com/photos/124690178@N08/

I have two D500s and have rejected another due to AF issues (returned exchanged) ..... I had the AF'S 500mm F4D and have the Tamron G2 plus the AF-S 600F4E VR FL plus the TC 14E III all of which require no AF fine tune including on the D750 ......... (even a Pentax and an older Sigma 150-500mm screw drive lens needed no AFFT) .... yet most of my shorter FL lenses require some AFFT or another ...... telephotos usually require little to none if they are no issues ..

so I would be inclined to consider sending both as a pair to Nikon because after all you have written, I see nothing that faults either your testing methods/ determination or your shooting technique ...... I'm not saying I get no misses by any means, but apart from erratic BIF images and the odd situation, the D500 turns in reliably sharp images!

https://www.flickr.com/photos/124690178@N08/

Considering that I am able to get noticeably sharper photos with the same lens on my Z camera, I am inclined to think that the lens itself is okay. Would that be reasonable, or is there reason to think otherwise?
For instance,









Of course, I can still get photos like this out of the D500 combo. I just can't quite sort out why too often, when everything looks right, I don't.

The last D500 shot is at F8 ..... I feel it's still a tad front focused .....

 

[SCoombs]

Doing some more testing, I have hit something that has me confused. The Exif data for all of these photos contains a depth of field estimate based on the focus distance, aperture, etc. For most of the photos, the exif DoF lines up very well with what a standard DoF calculator tells me. However, for photos I have taken at ~40 meters away as part of the tests, it's telling me the DoF is 9 meters - far more than the 2 meters that the standard formulas/calculators point to. I wonder why this is.

All the shots that are not sharp are due to front or back focus ..... overall I'm not seeing any special tendance either way, which to me implies that it could be you choice of focusing modes ... for best results shooting slow moving birds I would choose single point on the eye ....... BIF panning group mode (maybe a dynamic mode 9/21/53) ....... I also get good success when using auto area mode for approaching birds or group mode ......

I agree that there does not seem to be a consistent front or back tendency, however I do not believe this is due to the focus mode selection. For one thing, in almost all of these cases (other than BiF, for which I feel like I must use group AF) I am using single point or dynamic (which on a still subject basically is single point) with a very clear LOS to the subjects and never having the point drift off of the subject. Since starting to investigate this issue, I have especially been precise about not using multi-point focus modes to be sure to rule out as many variables as possible.

Please also note that I have extensively worked at tuning the focus and done literally hundreds of tests in controlled conditions with tripods and so forth and found no pattern as to whether it front or back focuses. Sometimes it's one, sometimes the other. I have also used FoCal which at times works but other times says the AF is behaving erratically and does not give consistent results.

Just tonight I did yet another test - a less "scientific" one but one with striking results nonetheless. I had a woodpecker which sat in the same place on a feeder for several minutes so I took a bunch of shots - some with the AF fine tuning on to the most effective value I've found so far (+8) and some with it off. I then loaded them in Lightroom and without knowing which was which I picked the most in focus ones. Afterwards I looked at the EXIF data to see which were which and there was no real pattern. Some I had picked as sharpest were with the fine tuning while others were without, but there's really no discernable difference between the two groups. This is all with a pretty shallow depth of field of about .6 inches.

All AF configured to the various configurable button choices for use on the fly ........ I use single point/ auto area mode and group as a basis ....

I would also go mirror up in the menu and give a blast with a rocket blower under the mirror just in case there a bit of dust on the prisms ....... if you want to check for focus consistency, use just single focus point on almost static subjects at various distances, I find I get 80% perfect focus minimum and up to 100% ..........

As noted above, I've done many, many such tests. I've also tried cleaning as you suggest.

I've never looked at the exif data in order to fathom out the whys and wherefores of AF issues btw.

https://www.flickr.com/photos/124690178@N08/

I have two D500s and have rejected another due to AF issues (returned exchanged) ..... I had the AF'S 500mm F4D and have the Tamron G2 plus the AF-S 600F4E VR FL plus the TC 14E III all of which require no AF fine tune including on the D750 ......... (even a Pentax and an older Sigma 150-500mm screw drive lens needed no AFFT) .... yet most of my shorter FL lenses require some AFFT or another ...... telephotos usually require little to none if they are no issues ..

so I would be inclined to consider sending both as a pair to Nikon because after all you have written, I see nothing that faults either your testing methods/ determination or your shooting technique ...... I'm not saying I get no misses by any means, but apart from erratic BIF images and the odd situation, the D500 turns in reliably sharp images!

https://www.flickr.com/photos/124690178@N08/

Considering that I am able to get noticeably sharper photos with the same lens on my Z camera, I am inclined to think that the lens itself is okay. Would that be reasonable, or is there reason to think otherwise?
For instance,









Of course, I can still get photos like this out of the D500 combo. I just can't quite sort out why too often, when everything looks right, I don't.

The last D500 shot is at F8 ..... I feel it's still a tad front focused .....

Interesting. As part of a "test" of the AF fine tuning at close distance I did a sequence where I photographed this woodpecker with my AF fine tune at +8 (where it seemed best) for around 20 shots then quickly turned it off for another 20 shots. I then looked at them all (not knowing which were on and which were off) and picked the ones I thought were sharpest and then looked to see whether the AF fine tune was +8 or 0. This was one of those with no fine tuning - so IF the fine tuning is correct, at least at this distance, it would be front focused.



I will share one of the ones with it on when I am at the PC again.

 

[Richard James]

It is all about the pixels, if the duck in image one has 2000 px on it and the duck in image two only has 1000 px on it, then it will necessarily have less resolution which impacts apparent sharpness.

Cropping does not change the pixels, it simply reflects what the sensor captures, which is an image that has less detail.

Resizing does not solve the problem, it generates interpolated pixels...

As far as I know I haven't resized the images. I have cropped them,

OK, I can not tell, I just threw that in as another possibly confounding issue.

IMO, the keys to comparison are the px under the subject and only change 1 parameter at a time..

2 - noise reduction reduces sharpness, and should not be an issue at the low ISO for many of your sample.

At the low ISO levels in the images you posted, noise should not be readily apparent at normal (print) size/viewing distance with an uncropped image,

You can still get noise at a low ISO in parts of an image which did not receive enough light during exposure,

Correct, all the time, any system. No light = low S/N ratio = noise appears.

but that's not the case in the noisy parts of the image here.

So, are you saying that there is "obvious" noise in mid-high tones, not underexposed shadows?

"Obvious" compared to what body? Is that an "equivalent" sensor?

As another poster here suggested, make sure VR is off if on a tripod (unless your lens has a "tripod VR" mode). Also off if the shutter speed is over about 1/500 as this apparently conflicts with the sampling rate for the VR system

Yes, I have tried shutting the VR off at higher shutter speeds but haven't noticed much difference.

AFAIK, it is dependent on the camera/lens combination and support conditions (how the support vibrates given an initial "push"). The vibration mode has to match that programmed into the camera to work optimally.

Unfortunately another issue - unrelated here since we're talking about speeds of 1/2000+, but still something I have to contend with - is that my hands are apparently so shaky that sometimes I need VR to eliminate motion blur at relatively low shutter speeds (e.g., 1/600 for a 250mm lens), even using good technique, arms tucked in, all that.

IC... In that case what you are seeing may be an improvement, but may not be "optimal". Again, my understanding of VR design is that the sampling frequency may have a negative impact at speeds higher than the frequency.

If it works for you, I guess that is OK.

We are talking about testing to identify if the problem is real, so composition is not relevant.

You're absolutely right. Unfortunately, I've learned that on this forum (and I suppose photography forums in general) I often need to worry about that anyways since too many users will hone in on whatever deficiency they can call out in a person's photos and somehow make the discussion about that rather than the question at hand.

That is a PITA...

Richard

 

[SCoombs]

Here is one of the shots of that woodpecker with the AF fine tuning to +8.





Here is one from today with +8:

 

[SCoombs]

I took some shots today (in unfortunately poor lighting necessitating f/5.6) which look very slightly off. I feel like they're front or back focused by a tiny amount but can't quite tell which.

 

[davethebirder]

I took some shots today (in unfortunately poor lighting necessitating f/5.6) which look very slightly off. I feel like they're front or back focused by a tiny amount but can't quite tell which.

Well they are all very good anyway ....... remember that there is more DOF field behind the focus point than there is in front as DOF increases with distance ..... so you don't divide equal front and back blurryness by two to find the focus point .......

You say that unfortunately you had to shoot at F5.6 because of low light ....... for testing AF that is only an advantage ......

maybe it's time to take pictures and let the mind occupy itself with other things .....

 

[Bob A L]

I'm not sure what your issue is. I took the one photo of bird that you noted was really unsharp, and just a simple sharpening made it as sharp as one could wish for. I don't know how you prepare your photos or view your photos, but different sizes require different amounts of sharpening depending on what you are doing with the photos. I see nothing that throwing money at for new equipment would help. Little to nothing wrong with your photos.