Nikkor 300/4 PF VR firmware fix: sample image comparisons to pre-firmware fix version

[rodeo_joe1]

<p>I think I'll stick with my Tamron SP 70-300mm VC zoom thanks, even though it's only f/5.6 at maximum FL. If the VR is, say, two stops worse on an f/4 lens, then you gain nothing by having that extra stop of aperture - provided the subject is fairly static of course.</p>

<p>Below is a handheld shot at over 50' (according to the lens distance scale) using the Tamron. The shutter speed was 1/160th, and despite the weather being dull and overcast I was able to use an aperture of f/8 at 200 ISO on my D800. The RAW file was processed and re-sized to the equivalent of a D750 before being 100% cropped. The complete frame is inset. All that's been done to the image is some tone-curve and WB adjustment and a small amount of USM applied in NX2 before resizing in IrfanView.</p>

<p>I've noticed that my handholding is getting quite shaky lately as well. Much worse than it used to be. Tempus fugit!</p>

<p>FWIW, I didn't even notice that a fly had crawled onto the top-left of the alarm box until I was viewing the image on my laptop. It was far too small to see in the camera viewfinder anyway.</p><div></div>

[Dieter Schaefer]

<blockquote>

<p>My recommendation is not to judge the 300mm PF based on its VR performance alone, but consider the whole package.<br /> As I said, to me, VR on a 300mm/f4 is not a critical issue.</p>

</blockquote>

<p>Whether VR is a critical issue or not (that's a personal preference and not everyone is concerned with shutter speeds fast enough to freeze motion) - Nikon advertises up to 4.5 stops gain through the use of VR on the 300 - so it damn well better delivers on that promise (and yes, I am well aware that under certain circumstances, even with VR images can be blurry (not talking subject motion here)). After all, the price includes the VR and I haven't heard Nikon giving discounts because a lens features doesn't work as advertised. I haven't even seen the lens yet - so I can't comment based on my own experience - but the images that Alan shows are unacceptable and I would return the lens instantly.</p>

<p>I had considered the 300 mainly because of its compact size and light weight - but as Shun correctly points out (though for other reasons), it is hard to justify the cost for a 300/4 lens. Whether I cannot or will not use a tripod in some situations shouldn't distract from the fact that in a rather broad shutter speed range, the lens' VR does not perform. Somewhat lower optical quality and even slower aperture aside - but in the situations I have in mind, the 70-300 VR might well be a viable alternative.</p>

<p>I tried yesterday with both my 70-200/4 and AF-S 80-400 and had no issues getting sharp images in the shutter speed range that the 300 fails miserably in (and well below as well). In fact, using my non-VR AF-S 300/4, I managed to get better results in the range discussed here than what the OP's images show. Doing so consistently would prove something of a challenge though.</p>

[ilkka_nissila]

<p><em>Nikon advertises up to 4.5 stops gain through the use of VR on the 300 - so it damn well better delivers on that promise</em></p>

<p>The improvement in stops is evaluated according to the CIPA standard which is based on assessing the acceptable sharpness of 4x6 prints at a viewing distance of roughly 2-3 feet. I think most people would be hard pressed to see the kind of unsharpness shown in this thread (in 100% crops) when viewing a small print (made of the entire image, not a subsection) at that distance. The acceptable circle of confusion is set at about 63 micrometers in the sensor plane (for FX). The pixel spacing on the D750 is about 6 micrometers. To contribute to the test result the blur has to be 10 pixels in diameter, or greater, assuming I understood the details right (according to a draft of the CIPA document that I found online). I suspect Nikon is covered where it comes to the accuracy of its claim regarding VR performance of the 300mm PF measured according to the CIPA standard.</p>

[rodeo_joe1]

<p>CIPA need a Kipper round the face!</p>

[ilkka_nissila]

<p>CIPA is being progressive. They understand that now photographs are primarily viewed on mobile phone screens (soon watch displays?) and posted on facebook at 0.2MP size instead of displayed as large prints.</p>

[ShunCheung]

<p>One thing we need to keep in mind is that vibration reduction has always been a percentage game. When I use the 70-200mm/f2.8 AF-S VR shooting a wedding inside a church without a tripod, I may end up using 1/50 sec, and VR is going to help stabilizing my hand holding and will increase my number of keepers. If I capture a static subject at 200mm, 1/50 sec with VR, in 10 consecutive frames, some will have more camera shake than the others. Obviously we have some limitations on this thread and we shouldn't post 10, 20 consecutive samples, but we really cannot judge the effectiveness of the 300mm/f4 PF's VR based on a few samples on this thread, one sample for each shutter speed.</p>

<p>For example, I may hand hold the old 300mm/f4 AF-S, without VR, at 1/50 sec and capture 100 images of a static subject. Maybe 3 out of the 100 could be very sharp, and I can just post one very sharp image and then brag about my ability to hand hold, but I would never show you the 97 very shaky ones. Doing so would lead to some highly misleading arguments.</p>

<p>Another thing we need to keep in mind is that any piece of equipment is going to have its limitations. As far as I can tell, the 300mm/f4 PF's VR issue seems to be related to the fact that it is small and light but with the magnification of 300mm. Sometimes you cannot have it both way. As we know:</p>

<ul>

<li>It is well known that the 70-200mm/f2.8 AF-S VR version 1 has terrible corners at 200mm on FX.</li>

<li>The 70-200mm/f2.8 AF-S VR version 2 has focus breathing issues when focused to 2, 3 meters.</li>

<li>The D800E and D810 (and D7100, D7200) are more prone to moire issues due to the lack of the AA filter.</li>

</ul>

<p>And the list goes on and on. The fact of the matter is that in real-life photography, I probably wouldn't use the 70-200mm/f2.8 AF-S VR version 1 for landscape photography at 200mm, otherwise, I wouldn't worry about any one of these "issues" at all. I have used the D800E, D7100 and now D7200 extensively, and I have used the D810 a bit also. Moire is extremely rare. I had a hard time getting it when I was trying to induce it on purpose with the D800E.</p>

<p>As advanced photographers, we need to be familiar with our equipment and know their limitations. The problem with web forums and bloggers is that frequently people only focus on the limitations to show their expertise, even though in real life some of those are rare and minor issues.</p>

[Dieter Schaefer]

<blockquote>

<p>I suspect Nikon is covered where it comes to the accuracy of its claim regarding VR performance of the 300mm PF measured according to the CIPA standard.</p>

</blockquote>

<p>I am sure they are - the lawyers will see to that. But we know that there are a lot of VR lenses that can do better than the CIPA standard demands - so there should not be any need for Nikon to fall back on the cop-out that the lens performs within the CIPA specs.</p>

<p>Also, we aren't talking about 4.5 stops in the images above - we are talking about around 1-2 stop(s) down from the 1/focal length guideline for shutter speed selection when hand-holding. If one accepts that one can hand hold a 300mm lens at 1/300s shutter speed and expect "decent" results - then one can surely expect that VR that claims as much as 4.5 stops improvement at least can manage 1 stop or even 2? Without blurring the image over a 10-pixel diameter?</p>

<p>Shun is correct - using VR is a numbers game - everyone needs to find out what to expect when using ever slower shutter speeds and relying on VR to save the bacon. And the slower the speed, the higher the percentage of "unacceptable" shots even wit VR. Shun mentioned testing the 80-400's VR capability at 1/20s - I can make that work too but certainly not 100% of the time (more like 5-10%). At 1/160-1/200s, I expect to do a lot better though - with not more than one out of ten failure rate. Without VR, that ratio stays at 1 out of 10 - success rate in that case.</p>

<blockquote>

<p>As far as I can tell, the 300mm/f4 PF's VR issue seems to be related to the fact that it is small and light but with the magnification of 300mm</p>

</blockquote>

<p>Then why does the 70-300 VR not have the same issue: it's also 300mm and their weights differ by 10 grams? And according to Nikon, the D750 shouldn't even show the VR issue...</p>

 

[ShunCheung]

<p>Dieter, I don't think the 70-300mm AF-S VR @ 300mm is as short as the 300mm/f4 PF. I haven't even used the 300mm/f4 PF, but according to some detailed discussions, adding the grip (i.e. weight) to the D800/D810 will improve VR performance. Holding the camera horizontally vs. vertically can also make a difference. Different bodies (heavier vs. lighter) will make a difference ....</p>

<p>Eventually it boils down to a lot of personal factors. If I am forced to depend on VR at slower shutter speeds, I would shoot more samples to improve my odds for getting a few good images, especially when I am concerned that VR doesn't work particularly well or there is a lot of subject movement.</p>

<p>While I really appreciate <a href="/photodb/user?user_id=731572">Alan Wilder</a> taking the trouble to post some samples, it is a major mistake to assume that every image captured with the 300mm/f4 PF at the same shutter speed as Alan's will yield the same amount of camera/lens vibration. The fact of the matter is that each image could be better or worse. Fortunately, most of us are no longer bounded by film and processing cost so that we are free to shoot more samples to improve our odds.</p>

[alan_wilder1]

<p>For perspective of the crop size to "see the forest from the trees", I'm attaching the full frame image.</p><div></div>

[alan_wilder1]

<p>Lastly, the same shot at 1/20 sec with VR switched on. As stated earlier, VR works great at these slow speeds.</p><div></div>

[ilkka_nissila]

<p><em>At 1/160-1/200s, I expect to do a lot better though - with not more than one out of ten failure rate. Without VR, that ratio stays at 1 out of 10 - success rate in that case.</em></p>

<p>I think you're much more optimistic than I am. I don't expect any shot made with VR ON to match the results obtained with VR OFF using a tripod (assuming the tripod, head, and tripod collar are good), and I expect any image that I make to be as good as it can be, which means that I shoot almost all my static subjects using a tripod, and generally stick to 1/(3*FL) or faster as a guideline for minimum shutter speed for hand-held photography. With a 300mm lens (sans VR) I use 1/1000s as a general starting point; with the VR PF I lower that to maybe 1/500s (typical) and in situations of need, 1/250s. This worked well when I tested the PF. 1/200s didn't work at all; so no question about it that there was a problem, however the firmware/PCB swap appears to address this adequately. Even if I buy the 300/4 PF, I expect my shutter speed to be in the 1/500s to 1/1250s range most of the time, and when I shoot a landscape subject I mount the camera on tripod.</p>

<p>The primary benefits of the PF version to me are the smooth, silent, jitter-free AF tracking, and its compactness which leads to many opportunities to use it when traveling. I can just throw in a 24-70, 70-200/4, and 300/4 PF in a small shoulder bag and go walk about whereas with a conventional 300mm prime, I would definitely need a significantly larger bag. Also at some sports events, the size of the gear that regular ticketholders can bring and use is regulated and I think the 300/4 PF has a good chance of being let in because it doesn't obstruct other spectators' views in the same way as a 300/2.8 would. And yet it tracks moving subjects excellently. However, I still want to see the behaviour of the lens under stage lights before committing to a purchase.</p>

<p>I think the 1/100s to 1/200s shutter speed range is probably showing reduced VR effectiveness using a lot of lenses. I tested 1/160s with the 70-200/2.8G II and hand-held shots with VR ON were clearly less sharp than those made with a tripod with VR OFF. Thus even with a 200mm lens I don't expect miracles from VR.</p>

[rodeo_joe1]

<p>But who, in real life, shoots a 300mm lens handheld at 1/20th of a second? Any live subject is going to have moved in that time - even if it's only wind stirring leaves. The "problem" shutter speeds shown are in the very region where VR should be coming into its own, and where I suspect this lens will mostly be used handheld.</p>

<p>I'll certainly be in no hurry to purchase this lens until Nikon have completely re-engineered the VR system.</p>

<blockquote>

<p>"CIPA is being progressive. They understand that now photographs are primarily viewed on mobile phone screens (soon watch displays?) and posted on facebook at 0.2MP size instead of displayed as large prints."</p>

</blockquote>

<p>I hope that was meant ironically Ilkka. Because this lens won't fit on a phone camera. And I get much better image stabilisation from my little Coolpix 100 at the equivalent to a 300mm zoom setting than that shown from this expensive lens.</p>

[ShunCheung]

<p>A 300mm is a super telephoto lens. Of course it makes no sense to hand hold it at 1/20 sec, regardless of how great VR may be at such shutter speed. That is exactly what I pointed out when I reviewed the 70-200mm/f4 VR and 80-400mm VR. Most people use super teles for sports and wildlife, where 1/250 to 1/500 is not at all fast enough to freeze any action. This is a lens that is normally used wide open @ f4 and 1/1000 to 1/2000 sec, which can now be achieved easily even under overcast situations due to the much improved high ISO results from current DSLRs. While I am glad that Nikon's fix to the 300mm/f4 PF has improved things (although most likely not completely), unfortunately, the subsequent discussion is about essentially nothing in real-life photography.</p>

<p>People pay $2000 for a 300mm/f4 PF mainly for its reduced size and weight. 20 years ago when carbon fiber tripods first appeared, a small Gitzo 1228 cost $550, while a bigger, studier and much heavier mid size Gitzo was around $300. You pay a lot extra to get rid of weight. The problem is that even a traditional 300mm/f4 AF-S is not that big such that it is difficult to justify the cost.</p>

<p>I consider the 300mm/f4 PF is more like a trial balloon. With experience from it, if Nikon can introduce a 400mm/f4 PF or 500mm/f4 PF, the weight and size reduction will be significant because those traditional superteles (with matching tripod and head) are big and difficult to carry for long distances. Another important development is E, which will likely be the future for all F-mount lenses.</p>

<p>On the other hand, if you need to take pictures of some street signs @ 300mm, f8 as some of the samples here, by all means get a 70-300mm/f5.6 zoom and stop it down. (The Tamron version is surprisingly great optically, although marginal mechanically.) You'll save a lot of money. Reducing the weight of your wallet is not the objective here. :-)</p>

[ilkka_nissila]

<p><em>I hope that was meant ironically Ilkka.</em></p>

<p>Yes, in part. The reality is that (popular) applications of photographs are moving towards smaller, lower resolution displays and at the same time cameras become higher and higher resolution, in a time when fewer and fewer people even consider making a print. I find that quite ironic. Personally I use an A2 printer but most of the time the prints are smaller (A5 size is typical); for framed prints on walls I do make large ones, but there is only so much wall space. However, it seems I'm an exception in that I like to see images in print before I even decide if I like them and what needs to be done to them in terms of improvements and post work. I lay a number of printed images on a table and decide which images are essential to the story and fit together visually, as a part of the process. Since I consider the print the final result, my criteria for what is acceptable image quality are based on that.</p>

<p>I don't think there will be a second recall of the 300/4 PF. There may be a completely new 300/4 after some 13 years. I think there are a lot of options in Nikon's telephoto lineup at the moment; lenses that I'm interested in but don't exist yet include a 135/2 AF-S. Without doubt some additional fluorine coated and PF or fluoride element based versions of the other supertelephoto primes are in the works, but who will be able to afford them is a big question. If the price of e.g. FL 300/2.8, FL 600/4 increases as much as it did with the FL 400/2.8, I think a lot of people will be gasping their breath. At least with the PF 300/4, many people can afford it. So in terms of return of investment, I think starting with a PF version of the 300/4 was reasonable. For me, 400/4 would be useful as well, but being able to afford one is a different matter entirely.</p>

[rodeo_joe1]

<p>Leaving aside all the discussion about potential use and what shutter speed people ought to be using with a 300mm VR lens. The underlying question for me is: Does Nikon actually <strong>test</strong> any of their newly-developed equipment before releasing it for sale nowadays? Do they have any real photographers on their staff? I mean technically competent people who know what to look for in a piece of photo equipment and are capable of beta testing it properly? Because a computer simulation or design spec is simply proving inadequate.</p>

<p>We've now had about half-a-dozen recalls on Nikon gear within the last 2 years. That's not a good track record, and I would be very wary of buying anything new from Nikon within 6 months of its introduction.</p>

<p>Surely there's no excuse for this sort of thing, now that prototyping should be as easy as generating a 3D printer sample and actually <em>using </em> it. Early adopters shouldn't be unpaid beta testers or QC inspectors.</p>

<p> </p>

[alan_wilder1]

<p>The cropped shot at 1/20 second was simply to demonstrate the VR effectiveness at around 4 stops. Though ridiculously slow for a 300 if any subject movement occurs, I can see it's usefulness for non moving subject or landscapes shot under poor lighting when a tripod isn't handy or permitted.</p>

[ShunCheung]

<p>My local camera store has a sale this weekend, with many representatives from various camera brands showing off their latest gear. I got to handle a 300mm/f4 PF. That lens is surprisingly small and light. I took about 18 test shots to check how well VR works at various shutter speeds from 1/20 sec to 1/200 sec.</p>

<p>Both of the images below were captured at 1/80 sec, wide open at f4, ISO 800 (or ISO 1600) on a D800E. I was a customer inside a store, so it was not a carefully set up test situation. I hand held the camera, focused to the center of the frame, and used the various shutter speeds that are supposed to cause VR issues. As far as I can tell, there is no VR problem at all on that particular lens.</p>

<p>I asked the rep about the firmware on the lens, and he had no idea. But since it is a Nikon test sample, I assume any issues from the early samples have already been fixed. Overall, I see no VR issue at all. Optically the lens seems excellent (as it should be at that kind of price), but it could do better if I were able to line up totally parallel to the sales sign. The 300mm/f4 PF AF-S VR is surprisingly light and small for a 300mm/f4 lens. The $2000 price tag is, of course, not cheap for this type of lens. The older 300mm/f4 AF-S, which is optically sound as well, only costs about half as much.</p><div></div>

[ShunCheung]

<p>Second image</p><div></div>

[alan_wilder1]

<p>Your sample shots at 1/80 demonstrates results similar to my own experiences where VR does a very good job in suppressing blur from shake with possibly only a trace amount evident in the vertical direction on your first example. I found it quite helpful for speeds of 1/20-1/80 with close to tripod supported sharpness. Speeds of 1/100-1/200 were not quite as vibration free but on average better than no VR.</p>

[ShunCheung]

<p>Alan, while I only managed to use the 300mm PF for a few minutes inside a camera store, I found VR works very well. Among 17, 18 images, I have none that shows the kind of vibration your earlier "Security Associate" signs are showing.</p>

<p>Obviously a brief test like that is not totally conclusive, but I think anyone who has a 300mm/f4 PF AF-S VR that doesn't show great VR capability at all shutter speeds should consider sending it back to Nikon for (warranty) repair/adjustment. I would expect VR not to work well in some occasional images, but most of them should be excellent. The 300mm PF should show great VR capability that we/I have accustomed to with the 80-400mm AF-S VR and 70-200mm/f4 AF-S VR.</p>