Consensus on 70-200 with FX?

Discussion in 'Nikon' started by kuryan_thomas, Nov 29, 2008.

  1. After reading past posts on the topic, am I right in concluding the consensus is that the current 70-200 does not work well with FX? If so,
    what is a good alternative?

    Now that the D3x seems to be real, I would like to upgrade from my D2x to a D3x. I know my DX lenses won't carry over, but I was hoping
    the 70-200 would. It just goes to show that you should buy lenses that work well with the camera body you actually have instead of the one
    you hope will be available in 3 years!
     
  2. ShunCheung

    ShunCheung Administrator

    Consensus? No way.

    There are a few people who loudly complain about the 70-200mm/f2.8 AF-S VR on FX DSLRs.

    I have now tested the 70-200 extensively on the D3 and D700.
    For the most part it works fine. However, at and near 200mm, image quality drops drastically at the four extreme corners,
    but it only affects small areas in the four corners. On the D3 and D700, I would say it affects approximately 400x400-pixel
    squares. For any portrait, wedding, news, events, sports ... type images, those areas are typically out of focus without any
    subject anyway so that it is a complete non issue.

    If you shoot landscape and demand corner-to-corner sharpness, it can be a major issue.

    Personally, I am very happy with the 70-200mm/f2.8 AF-S VR on FX bodies, but of course I am well aware of its limitations.
     
  3. It seems to depend on which audience/group you'd like a consensus from. Many seem to find that the 70-200/2.8 behaves wonderfully while doing actual, real-life work on a full-frame body. Certainly it has behaved fine for me on actual 35mm film, and I've seen accounts from plenty of people using it on, say, a D700 doing PJ-ish style work where they're very happy. The other camp, which can find misery in less-sharp far corners of the frame while shooting wide open, may not be shooting a practical style that I understand. It's not a landscape lens, to be sure.

    This may be one of those things that you'll need to see tested on a D3X, since the sensor is a different beast. The way that light falls on its corners from that lens, and the way that sensor behaves is something better seen than guessed at. Lots of variables, there. And at 24.5MP, some minor grievances may turn into true show-stopping complaints.
     
  4. ShunCheung

    ShunCheung Administrator

    Here is a landscape example captured with the 70-200 at 200mm, f8 on a D700. The four red squares are 400x400 pixels each. They indicate the areas where I consider image quality poor.
    00RdsZ-93231584.jpg
     
  5. ShunCheung

    ShunCheung Administrator

    Here is a 500x700-pixel crop from the top left corner. The red square is once again the 400x400 area. In the cropped image, it should be very clear that the left (corner) side is much worse than the right side.
    00Rdsw-93235584.jpg
     
  6. Thank you very much, Shun and Matt. I appreciate you taking the time to respond and post sample shots. Shun, your
    images have helped me: I'd assumed the sharpness issue was all along the edges, not just in the corners.

    I'd definitely be buying the D3x for landscapes. It makes sense to wait and see how that lens performs with that body.
    Maybe at that resolution, I can crop out the problem areas. In any case, it might be August of 2009 before the camera is
    generally available; I have plenty of time to decide and budget accordingly.
     
  7. The D3X can't make more detail in the corners of the 70-200 image - 35mm film doesn't, the Kodak full-frame DSLRs
    don't, and D3 doesn't so the likelyhood of some new development here is practically 0%. The corners of the 35mm
    frame are simply outside of the covering power of the lens.

    If you make a portrait in dark surroundings, true, the vignetting would be hard to see. But if you're
    photographing a person against something white, such as a white wall, white backdrop, snow or ice (in skating,
    hockey), at f/2.8 the dark corner problem is very evident - and you can still see it at many FLs at f/4.5. Shoot
    an architectural detail against the blue sky? The vignetting will greet you. You
    can of course correct it in post-processing if that's what you like to spend time on. Personally this
    kind of performance is appalling in a lens in this price range, but my opinion is partly based on the fact that I
    often photograph people against white backgrounds (snow) and also I like high midtone contrast which I achieve
    with a curves adjustment; this increases the visibility of the vignetting. Luckily I got what I paid for mine
    back so I effectively got an almost free rental of the lens (new) for 2 years (almost, since inflation is nonzero
    and I had to have it repaired for a stuck diaphragm once).

    The 70-200 is otherwise an excellent lens and in some ways better on FX than DX - for example, the center
    sharpness of the images at 200mm f/2.8 is excellent on the D3 whereas they are clearly soft on the D200 (but on
    that camera it would be the FOV equivalent of a 300mm, so it's not exactly the same thing). It's all a question
    of which part of the image you look, you may love it or hate it.

    Personally I think the lens is a pain to lug around but it has many virtues also, but some limitations. I am
    waiting for a new version and use other glass in the meanwhile.
     
  8. Actually, I used it for three years mostly on DX. I am not recommending for or against this lens at this time; if
    you need it you need it, and no other lens may do what this lens does for you. As long as you're aware that you
    may need to do additional post-processing if you shoot at wide apertures, and don't use it in applications
    requiring corner sharpness it should be fine.

    An option is to use the 4:5 crop mode in the D3X and compose your photographs accordingly. This way you might
    avoid the issue altogether.
     
  9. I used the 70-200 on my D3 for six middle school football games this year. I was extremely pleased with the quality,
    and apparently so were the parents who bought quite a few enlargements all the way up to 16x20. My biggest
    problem was the very poor light quality at middle school fields. Luckily I had the D3 and had useful quality all the
    way to ISO 6400. Of course I shot RAW and processed with NX2. I have no complaints at all with the 70-200.

    --Barry Clemmons
     
  10. If you look at the quality you get with some of the zoom lenses of recent years (e.g. the 24-70mm) the image
    quality is amazing. In
    many cases primes (that may be of very old design) are not as good as some of the new zoom lenses. So I can
    understand why some folks get the idea to own zoom lenses to do it all.

    I personally
    still have the "old" concept in mind that a zoom is the lens of choice for action and the prime is the choice of
    lens if you got the time to set up your image. In case of the 70-200 AFS VR lens this follows the old school.
    Great lens if you need the fast action. Amazing crisp detail in the center. BUT: A 180mm AFD f2.8 or a 200mm F2.0
    will give you better corner to corner IQ
    than the 70-200mm.

    The example that Shun posted is really nice because you can look at it and decide if the price tag of say the 200
    F2.0 lens is worth the corner sharpness^^. If you crop an image obtained from the 70-200 to the view of the
    fairly cheap 180mm
    AFD or if you are able to go a bit farther away with your tripod with your 180mm you get rid of the corner
    problem :)
     
  11. By my standards it is very bad. The point of the FX sensor is to use the full sensor not to crop in, otherwise you might as well shoot DX. Now here is what is funny my old AF 80-200 f/2.8 push/pull has far less vignetting. When I first shot the 70-200mm f/2.8 VR I could literary see the vignetting in the viewfinder. I thought it was the lens hood. Later I learned it was a combo of things. The FX sensor is slightly larger than 35mm film and the shot the 70-200mm f/2.8 VR is a compromise between full frame and DX. Here is a full frame image you might notice the vignetting. Camera: Nikon D3 Exposure: 0.001 sec (1/1600) Aperture: f/2.8 Focal Length: 82 mm ISO Speed: 100
    00ReDB-93399584.jpg
     
  12. ShunCheung

    ShunCheung Administrator

    There is vignetting in practically every f2.8 zoom I have used, including both the 80-200mm/f2.8 AF-S and 70-200mm/f2.8 AF-S VR wide open at 200mm on FX. You probably lose about a stop in the corners at 200mm/f2.8. But I have never seen something as serious as what Ralph's example shows, except on the 24-70mm/f2.8 AF-S. The 24-70 has the worst vignetting among Nikon f2.8 zooms.

    This is all moot for Obi Thomas who shoots landscape as you typically stop down to f5.6 to f11 anyway. In fact, the D3 family's frame rate, AF speed as well as an f2.8 zoom all seem to be an overkill for landscape. A lighter body with more pixels and some 70-200mm/f4 would be ideal.

    Unfortunately, the corner problem on the 70-200 at 200mm does not improve much when you stop down.
    You are still stuck with it at f8 as I showed above.
     
  13. I'm with Shun and others here. My 70-200 vignettes on D700 to the same degree at lengths close and up to 200mm despite stopping light down to f/8 - f/11. There is also some distinct loss of sharpness in the extreme corners. Having observed all of this I still rate the functionality and beautiful bokeh produced from this lens as being more than adequate compensation for me. To loosely quote a wise contibutor here at p.net - "in photography as in much of life there is compromise around every corner".
     
  14. "The FX sensor is slightly larger than 35mm film..."

    Ralf I hate bean counters and would not have commented but in this context size matters :)

    If we want to talk about a (probably not relevant) difference in size then FX is a tiny
    bit smaller
    than 24x36mm film frame. Of course film frame size in real live depends on your individual camera but I have
    many frames on film
    that are pretty much 24mm x 36mm.

    In reference to the topic I would say the discussed vignetting is not at all related to FX being larger than 24mm
    x 36mm
    film size. It is like Matthew points out the result of compromises in the design by Nikon.

    Just in case I am in error please give your reasoning. (I took the FX frame size 36.0 mm x 23.9 mm from Nikon.)
     
  15. This is all moot for Obi Thomas who shoots landscape as you typically stop down to f5.6 to f11 anyway. In fact, the D3 family's frame rate, AF speed as well as an f2.8 zoom all seem to be an overkill for landscape. A lighter body with more pixels and some 70-200mm/f4 would be ideal.
    Shun, you're not suggesting I switch to a Canon 5D II with their 70-200, are you? (Just kidding...) I prefer the ergonomics of Nikon bodies and specifically the large form factor of the pro systems.
    You're absolutely right that I don't make use of many of the features in even my D2x. I only ever shoot raw and almost always with mirror lockup on a tripod. However, the weight is mostly OK with me, as I personally prefer the large form factor. I realize everyone has their own preference in this regard. This seems to be a factor in many items I buy. I want a few features that are somehow considered "high end" and thus end up dragging in hundreds of features I'll never use.
    As for the lenses, I don't often use the 2.8 end of the aperture range, but the same thing happens: when I look for IQ, versatility, and other "quality" features, the 2.8 gets dragged in.
    Thank you, everyone, for weighing in.
     
  16. Dpreview.com measured the vignetting on the 70-200 and in the lower right corner it was >2.3 stops at
    200mm f/2.8. Interestingly it is slightly less in the other corners, as it was on my 70-200; this asymmetry meant that I had to fix
    the lower right corner separately and not just apply a generic vignetting correction.
     
  17. I have a question and would apprecieate an answer. If you use the 5:4 mode of the D3 does it eliminate enough area to crop out the circle of fuzziness?
     
  18. ShunCheung

    ShunCheung Administrator

    If the amount of vignetting is so uneven at the four corners, you clearly have a defective sample and therefore it is completely meaningless in this discussion anyway. Fortunately, I have a unifrom gray sky outside. I just went outside and took this quick snapshot, without any post-processing other than scaling it down to fit the browser window here.
    00ReNc-93505684.jpg
     
  19. ShunCheung

    ShunCheung Administrator

    Here is a more real-life shot with the exact same lens at 200mm, f2.8.
    00ReO3-93509584.jpg
     
  20. I just happened to be working on an long lens comparison when I saw this post. One quick thought. The 200 f/2 vignettes badly as well, so don't be in a hurry to sell your 70-200 for it (the prime is sharper of course). Both get better when stopping down but still may require some post work. No problem as it takes but a few seconds to fix in ACR 5.2. I spend more time tweaking contrast and other adjustments.

    For weddings and sports, I stressed over using the 70-200 on fx until I actually put it to use. I usually frame the subject on center out to thirds give or take, right where the image is sharp, and the corner softness just adds to bokeh/selective focus. Same for the vignette usually as well. In short, its characteristics match my uses very well, and I feel it's well worth the money. The last hockey game I shot, 3-4 guys with D3s, I think one or two with Canons, all against the glass in the corner with 70-200s. Same for basketball now in the corners of the court. Nobody seems to be complaining.

    If you are however shoot telephoto landscapes exclusively, I don't think you'll be satisfied with the 70-200. If money were no object, I'd think 200 f/2 or 200-400 f/4 (probably the latter). This is exciting however that a landscape shooter isn't talking entirely about the 14-24.
     
  21. Slightly asymmetrical lens properties are common rather than a "defect" when it comes to zooms. Having a lens realigned may temporarily reduce the asymmetry but unfortunately the lens is likely to drift out of alignment over time, so it's pretty pointless (and expensive). I have discussed this with a senior Nikon repairman and the only conclusion that could be made is that if I am expecting anything else, my expectations are too high.
     
  22. Of course, there are also defects that can be repaired. But Shun's first vignetting test image also shows asymmetric brightness in the corners (the lower left side is the brightest, the lower right corner the darkest).
    Many teles that came out between approximately 2001 and 2006 are DX-optimized and have vignetting on FX - older teles typically have less since they were designed for film. The comparison should be made at the same apertures to be meaningful for practical use - i.e. 200/2 at f/2.8 vs. 70-200 at f/2.8 etc.
     
  23. I took Shun's vignetting example image and did some pixel peeping and analyzed the asymmetry in a simple way.

    The mean of the pixel values is 111 (I converted the original to an 8 bit grey image for simplicity).
    I rotated the image 180 degrees and took the difference of the two images.
    The mean of the difference is 3.9 out of 111. The spread is small.


    If I flip the other axis the difference is 3.8 out of 111.
    Since in 8 bit both values would come out as 4 this "asymmetry" is nothing that would concern me.
    If we have to turn to 16bit images to detect asymmetry I would call this is a well aligned lens.


    BTW: I use this lens for 3/4 of a year and am very pleased with it - but please consider my zoom expectations as
    stated above.
     
  24. ShunCheung

    ShunCheung Administrator

    Ilkka, I am afraid that repair person is right on suggesting that you have unrealistic expections. I used to have the 80-200mm/f2.8 AF-S, and I compared that to the 70-200mm/f2.8 very carefully with both Velvia film and my D2X. As far as performance goes, those two lenses are very similar. I tested their respective vignetting issue on 35mm film and they are also very similar. I sold that lens last year so that I cannot test it on FX any more. Here is a test shot with my 300mm/f2.8 AF-S, 1st generation. This lens was introduced back in 1996, years before any DX DSLR was introduced. I bought it in 1998. This lens is arguably the best Nikon tele I have ever owned. As you can see, even a high end prime has some vignetting issues. If you actually use these lenses to take pictures, I think they are just fine.
    00Ref3-93687684.jpg
     
  25. I cannot remember where I read that this lens works perfect on the DX sensor cameras because it is using the "sweet spot" of the lens. By "sweet spot" they were meaning the center part. So, the vignetting and soft issues were not present on the DX. Maybe individuals that have used it on the DX sensor cameras expect the same "sweet spot" performance on the FX sensor cameras. I don't think any company makes a better alternative to this lens. The lens does not work as good on an FX as on a DX camera. But it is the best there is.
     
  26. Thank you guys for your very good discussion and all your wonderful contributions. This is more important to me than any discussion or analysis in DPReview.com or elsewhere. This is like a master's class in vignetting with teles and zoom teles, and something I actually can 'take to the bank'.

    Thanks again.

    John (Crosley)
     

Share This Page