Technical aspects of the AFS 300mm f/4 E PF lens

[Jakov Minić]

To put the 300 PF issues into perspective, I just mounted my 50-300/4.5 ED AI Nikkor on the D500 and verified it was at least as sharp as the 300 PF, if not sharper. Since the vintage 50-300 is purely manual (but with CPU)  there is neither AF Fine-tune nor VR behaviour to attend to.

Have you compared sharpness with other manual focus 300mm lenses, like for instance with a 300/4.5 AiS?

[Bjørn Rørslett]

No, just got the D500 and am busy trying to learn the camera's darker secrets. The 50-300 just happened to be in the vicinity when I did the 300 PF tests on the D500.

[Bjørn Rørslett]

Fired up the D800 and again, no issues found with 300 PF and VR over the range 1/20 to 1/250 sec. I feel like being on a ghost chase.

[Erik Lund]

You are Bjørn, the 300PF is a wonderful lens!

[Bjørn Rørslett]

Never said against that ... it has a lot of virtues. Still, surprising that the vintage 50-300 can match it in sharpness .... That came unexpected for sure.

As to the VR issue, I'm none the wiser. I don't doubt it exists for some combinations of 300 PF lenses and cameras, but my 300 PF apparently isn't troubled.

[simsurace]

As to the VR issue, I'm none the wiser. I don't doubt it exists for some combinations of 300 PF lenses and cameras, but my 300 PF apparently isn't troubled.

I don't think this is very surprising. It would have been interesting to test with my lens on all your cameras.

Could you please post a sample 100% crop of the D800 at 1/125s? Just to get an idea how sharp it is.
The issue is that the Nikon rep I was talking to was denying that the expectation of getting a sharp result at these speeds on the D750 (which has less resolution than the D800) is realistic at all, and thus my complaints that it was impossible to avoid the strong blurs (for which I posted examples earlier) were misplaced. I then showed them online reviews where people get sharp results, to which he said that they are very surprising and not to be expected.

He told me to use faster than 1/400s speeds at 300mm if I want sharp pictures, a suggestion which renders VR almost superfluous as a feature and as a selling point. The rep told me that he had talked to a supervisor, who had supported this view.

To me, this attitude seems like a joke. Examples such as yours are apparently insufficient evidence to establish that my (and other people's) expectations of what VR needs to accomplish are realistic and justified.

The rep that Chris talked to in the Netherlands seems to have a much stronger belief in the capabilities of Nikon's products.

[Ilkka Nissilä]

I basically agree with the response Simone got from Nikon. Although acceptable (slightly blurry, with sharpness varying from shot to shot, most shots "usable") results can be obtained at slower speeds, sharpness in my 300 PF shots increases towards faster shutter speeds and 1/400s sounds like a reasonable guideline for consistently high level of sharpness. I prefer a bit faster speeds myself, from 1/500s to 1/800s. At these speeds, VR stabilizes the viewfinder image, resulting in more precisely controlled compositions and better autofocusing and probably also reduces the effects of hand shake on the images. With the 300/4D AF-S, I needed to go to 1/800s or faster for acceptable hand held sharpness even with 12 MP FX  (1/500s didn't do it) so there is some advantage to having VR in the lens, in the form of improved sharpness at borderline shutter speeds. To get the advertised number of stops you have to switch to a  lower resolution standard i.e. something that you might see in small print at some distance away, or on a typical laptop display without cropping, or on a mobile phone screen.

I have the same experience basically with other lenses, e.g. I don't normally let the shutter speed below 1/200s with the VR 70-200/2.8II since if I do, sharpness invariably starts to fall. However, at 1/200s VR does help, at 200mm focal length; it helps with composition, autofocusing and reduces the effect of camera shake. I'm perfectly happy having this advantage in my VR lenses and think expecting perfect pixel-level sharpness at slow speeds is just unrealistic and a tripod should be used instead. However, if the usage of the image is a relatively low resolution presentation, then shots taken at very slow speeds can sometimes be considered acceptable.

[Bjørn Rørslett]

As requested, this is 300 PF at 1/125 hand-held with the D800. Test target the curtains in my living room, well suited actually as light is shining through the fabric and thus any blur would be easily spotted.

'Sharp enough' in my book. Perhaps it would be even sharper at a higher shutter speed? Really a moot point. This level of sharpness will print to any realistic size from a D800 file.

100% crop, centre of frame. VR set to 'Normal'. Do observe the subject is not flat so follow a wrinkle downwards once it appears to be in focus. The salient point, however, is the absence of visual blurs into the darker areas.

[Andrea B.]

I'm quite eager to get the 300/4E VR regardless of potential AF problems at certain speeds. So it is great to see this lens tested & discussed so thoroughly here.

One thing I've gotten from this thread is that I should get out there and actually test the VR at various speeds on any of my lenses which have it - something I have never done - although I am aware that at certain speeds VR on any lens is best left turned off.

[Andrea B.]

Examples with D750 + 300/4E VR

Photos followed by unresized excerpts.

The final unresized excerpt shows that if speed is moved to 1/400", then there is no "doubling".

[Bjørn Rørslett]

Had I obtained such results with my lens, I would have demanded my money back ....

[simsurace]

I basically agree with the response Simone got from Nikon. Although acceptable (slightly blurry, with sharpness varying from shot to shot, most shots "usable") results can be obtained at slower speeds, sharpness in my 300 PF shots increases towards faster shutter speeds and 1/400s sounds like a reasonable guideline for consistently high level of sharpness.

I really appreciate your thoughts on this, but I fear we are really talking about different things here.

There is an abysmal difference between certain 300PF combos (camera+lens) and others. You may not agree with the evidence presented, but I think it is pretty well-documented by now.

The good combos show consistently higher sharpness than what your post suggests is possible, and certainly much better than the bad combos (just look at Bjørn’s example for a ‘good’ example, even though I believe I have seen even better).

While I agree with your statements and Nikon’s as a conservative ’best practise for the highest consistency of sharpness, I was not asking for that since I already know those practises and I will try to use them as much as possible. However, and this is my main point, those statements are not an acceptable rebuttal (to me anyway, and I don’t think I’m alone in this) to the evidence presented, since they do not address the evidence at all. When has it become acceptable to ignore evidence? At least they should explain why the evidence is deemed invalid. I think the response I got from a rep (which I quoted in one of the previous posts) very badly represents Nikon and the efforts of their engineers to design an excellent product. Hiding behind the safety of worst-case scenarios shows a lack of commitment to their products. (As an analogy, it would be like a producer of off-road vehicles recommending they only be used on paved roads to avoid any damages like broken axles etc. So while that statement is technically true and every sane person knows that any car can break when Murphy strikes and bumpy ground increases the probability of breakage, it is a strange (to use the mildest of all adjectives) thing to tell a customer who has just presented some evidence for a design flaw in the axle which makes damages more likely. The minimum courtesy is to explain why the evidence is deemed invalid). This sort of response is very reminiscent of the case where Nikon told D600 customers that (let me paraphrase) “sensors just get dirty” when that camera’s propensity to accumulate dirt/oil at a much faster rate than any other DSLR had already been documented. I’m not getting bitter over it nor am I blaming it on individual employees, but I will never say that this an acceptable thing to tell a customer.

But I will no longer harp on the issue now. According to Chris, at least one Nikon technician under this sun has already acknowledged that there are unresolved issues with VR on the 300PF and stated that somebody at Nikon is working on them, so there is nothing left for me to prove I guess.

[simsurace]

Examples with D750 + 300/4E VR

Photos followed by unresized excerpts.

The final unresized excerpt shows that if speed is moved to 1/400", then there is no "doubling".

Thanks for posting these examples. They were the basis for the table of working/non-working combos I presented earlier.

[simsurace]

As requested, this is 300 PF at 1/125 hand-held with the D800. Test target the curtains in my living room, well suited actually as light is shining through the fabric and thus any blur would be easily spotted.

'Sharp enough' in my book. Perhaps it would be even sharper at a higher shutter speed? Really a moot point. This level of sharpness will print to any realistic size from a D800 file.

100% crop, centre of frame. VR set to 'Normal'. Do observe the subject is not flat so follow a wrinkle downwards once it appears to be in focus. The salient point, however, is the absence of visual blurs into the darker areas.

Thank you!

[simsurace]

Had I obtained such results with my lens, I would have demanded my money back ....

This is one possibility that I have pondered. I may still demand my money back from the dealer if I don't hear back from Nikon.
I thought that talking to Nikon Support first would be more productive for them. If they are willing to listen...

Giving the lens back and getting another one would give me a reasonable chance to end up with another faulty sample. I haven't decided whether I want to go through this experience. The VR issues no-withstanding, my lens sample is exquisitely well-centered and it could be that I get lucky with regards to VR, but unlucky with regards to centering . How funny is life's randomness...