FE16-35/4 DxOMarked (with charts)

Discussion in 'Sony Alpha E-Mount Lenses' started by quezra, Nov 7, 2014.

  1. quezra

    quezra TalkEmount Top Veteran

    916
    Aug 22, 2012
    DxOMark just released the results of the FE 16-35/4 and it just squeezes in as the best 16-35/4 on the market (by a fraction, but a significant fraction!):

    At 16mm:

    ugi2kUI.

    Looks about dead even, though the Nikon seems to be fighting a slight field curvature

    At 24mm:

    CtPlHdZ.

    Starting to pull ahead of the Canon version... which mind you is tested on a much lower resolution 5dIII

    At 28mm:

    XoP3ztJ.

    Well the Canon isn't looking so hot.

    At 35mm:

    Pk60ZEf.

    Surprisingly, the Canon comes back into its own, but the Nikon is awful. However, Canon is poorer than the others in its mid-zoom, and Nikon at tele. So for consistency at every focal length, Sony does amazingly well. And in a smaller package to boot.
     

    Attached Files:

    • Like Like x 10
  2. serhan

    serhan TalkEmount All-Pro

    Aug 27, 2011
    NYC
    • Like Like x 1
  3. mcvu

    mcvu TalkEmount Regular

    90
    Sep 24, 2014
    Bay Area, CA
    Minh
    It shows that Zeiss doesn't know how to make good zoom. Their primes are unbeatable, but not the zooms.
    The 16-35mm/4 is not as sharp as both Canon and Nikon versions; 16 PMpix is such disappointed result and sharpness uniformity is not good neither. It will be good on A7s, but on A7r you are losing more than half its resolution capability.
    It looks like it's the same league as the 24-70mm/4.

    I bought the 24-70mm/4 and returned it because I can't stand its low resolving power and sharpness uniformity on my A7r and now my hope for a good wide angle zoom is gone too. Maybe it would have been much better if Sony produced a 16-35/4 G lens instead of Zeiss, taking into account the good results from 70-200/4 G.
     
    • Like Like x 1
  4. mcvu

    mcvu TalkEmount Regular

    90
    Sep 24, 2014
    Bay Area, CA
    Minh
  5. ggibson

    ggibson TalkEmount Regular

    154
    Sep 1, 2011
    LOL, not sure what review you're looking at, but the shots on photographyblog look plenty sharp to me.

    This lens looks great from the early reviews. Great performance throughout, even if it's slightly softer at 35mm. It's definitely on my buy-list.
     
    • Like Like x 1
  6. mcvu

    mcvu TalkEmount Regular

    90
    Sep 24, 2014
    Bay Area, CA
    Minh
    Well, if it's sharp enough and meets your needs then good for you.

    The pictures just confirmed DxOMark tests, which concluded, and I agreed, that peak perceived sharpness is 16 PMpix. That's peak sharpness at some spots, the rest of frame will be lower than 16 MP.

    Again, sharpness is just one aspect of a lens. Color rendition, micro contrast, CA, distortion, bokeh and other characteristics will make any lens unique.
    As long as you're happy, that's what matters.

    Personally, I was so disappointed with the lens performance, especially with the Zeiss brand + price tag.
    The only reason I picked A7r over A7 is resolution.
    Looks like I'll only use prime lenses on A7r. If I needed zooms, they will be on A6000, especially the good 10-18/f4 lens. The only zoom I'll consider for A7r is the 70-200/f4 G. But it's too big for me.
     
  7. quezra

    quezra TalkEmount Top Veteran

    916
    Aug 22, 2012
    That's a strangely selective approach.

    The 10-18/4 rates worse by DxO standards: 9 MPix on either NEX-7 or A3000. But most important is to look at field maps (which take resolution into consideration):
    UUJL56Y.

    WSk6Ow1.

    The 16-35/4 is substantially better than the 10-18 all across its range (and bearing in mind APS-C f/4 is really f/6 in FF terms), only weakening at 35mm (which the 10-18 doesn't reach)
     
  8. mcvu

    mcvu TalkEmount Regular

    90
    Sep 24, 2014
    Bay Area, CA
    Minh
    I didn't say that the A6000/10-18 combo is sharper than a7r/16-35.

    Peak sharpness of the A6000 combo will be less, but I expect sharpness uniformity (across the frame) will be better because the 10-18 lens covers FF size, and aps-c crops and uses the middle good portion; so it avoids soft corners and edges. NEX 7 is known to have bad corners with wide angle lenses.

    For landscape uses, I'd rather have sharpness uniformity than much sharper peaks and noticeable uneven regions elsewhere.
    Once you have sharpness uniformity, it's much easier to sharpen the whole picture post, using software. It's much harder and more time consuming to fix uneven sharpness otherwise. It's the same problem that Sony tries to solve with the curved sensor.

    The larger the sensor, the much harder it is to design good lenses.
    Because of its tiny sensor, the RX100 M3 produces superior sharpness uniformity with its zoom, than any bigger sensor + lens combo, even better than a7r + FE 55/1.8.
     
  9. WoodWorks

    WoodWorks Super Moderator

    Dec 12, 2012
    Ashland, OR, USA
    David
    Unless I'm completely misreading them, the charts Joel posted above seem to show the opposite.

    And having used (and now sold) the 10-18 on the A7, I can assure you that it only covers FF between ~12mm and ~16mm, and there's significant smearing at the edges even within that range, as this post shows. Livable perhaps, but if even sharpness across the frame is your goal, the 10-18 is not going to come close.
     
    • Like Like x 1
  10. mcvu

    mcvu TalkEmount Regular

    90
    Sep 24, 2014
    Bay Area, CA
    Minh
    That's why I'm not using the 10-18 on A7 series. On the A6000, with the aps-c crop, it's a different matter.

    The chart Joel posted is for NEX 7 which is known to have bad edges & corners with wide angle lenses.
     
  11. ggibson

    ggibson TalkEmount Regular

    154
    Sep 1, 2011
    Wait, why is the 10-18/4 "good", when the "Perceptual MP" DxO result on that lens is similar to the FE 16-35? They rate the 10-18 a 9MP on the 24MP NEX-7. So why is that lens good, but the 16-35, which has a higher perceptual MP rating on the A7r is not good?

    On the topic of the 10-18/4, I'll be curious to see a 10-18/4 on A6000 vs 16-35/4 on A7 comparison. Should be a good race. :biggrin:

    Personally, I like my UWA zooms to go to 35mm for walk-around purposes. That's why the little m4/3 Olympus 9-18mm has been my favorite lens for the last few years.
     
  12. quezra

    quezra TalkEmount Top Veteran

    916
    Aug 22, 2012
    Which is why I posted the 20 MP a3000 comparison as well, showing the same acutance drop-off. As I said, you appear to be selectively cherry picking some results and handwaving away others.
     
  13. mcvu

    mcvu TalkEmount Regular

    90
    Sep 24, 2014
    Bay Area, CA
    Minh
    I'm after uniform sharpness here, not peak sharpness.
    You take a lens that can cover FF and crop away bad edges and corners on aps-c, you should have better uniform sharpness.

    You can't do that on the A7 series, i.e. create lenses that cover more than FF, then crop; because the E-mount was originally designed for aps-c size sensor and now when Sony uses it for FF, the mount opening can barely cover the whole FF sensor. In order to design a really good FE lens, they would need to do lots of corrections, with more lens elements/groups, thus making bigger, heavier lenses and much more expensive; similar to the Otus series.

    My prediction that A6000/10-18 will have better uniform sharpness than A7r/16-35 might not be true and we can't verify it until somebody test both combos. One thing I can verify and did do test shots even though not extensively is that my RX100 M3 has much better uniform sharpness than A7r with either FE 35/2.8 or FE 55/1.8. But then there're different tools for different jobs. For portraits, or where shallow DOF is preferred, uniform sharpness is a moot requirement.

    Anyway, we're off track here and I'm sorry for steering the thread that way.
    Lets not talk about other lenses or cameras. Go back to A7 series and the 16-35mm lens.

    For A7 series owners, we don't have any native FE wide angle lens now, except for the 16-35mm zoom and I'm just pissed at Sony for producing mediocre zooms. Both of the most popular zooms for any system, ie. the 24-70/f4 and 16-35/f4, can't stand up to A7r resolution.
     
  14. MAubrey

    MAubrey TalkEmount Top Veteran

    The last sentence here is certainly true. However, the size of the mount opening has nothing to do with creating quality lenses, whether they cover FF or more than FF. And If they did cover more than FF, that portion of the image circle won't hit the sensor regardless of how large the mount is, so I'm not sure how that's really relevant to question. There's no definite correlation between mount diameter and lens quality. The Zeiss Contax/Yashica lenses have a mount diameter that's about the same as E-mount and many of those lenses are absolutely spectacular.

    I'm assuming your uniformity statement there is without downsampling the 36MP of the A7r to the 20MP of the RX100?

    Well I'm not super sure there's much to complain about. There are few ultra-wide zooms that do stand up to 36MP sensors. The best is still the Nikon 14-24mm f/2.8...which still only achieves 23MP according to DxO. The 19MP of the 16-35mm f/4 on the D810 compared to the 15MP of the 16-35mm f/4 on the A7r is really only an 11% increase in resolution (15 = 41.67% of 36 vs. 19 = 52.78% of 36). My experience is that you'd need a much larger increase in resolution before you'd even be able to tell the difference between the two...especially if you're printing.

    But I'll still rather be using primes any way...
     
  15. ggibson

    ggibson TalkEmount Regular

    154
    Sep 1, 2011
    • Like Like x 2
  16. mcvu

    mcvu TalkEmount Regular

    90
    Sep 24, 2014
    Bay Area, CA
    Minh
    What I meant is that there're many ways to have better uniform sharpness. One way is to have more corrective elements/groups, complicated design.
    The other is to have a much bigger image circle, similar to 10-18, then crop using the sensor size. The latter can't be done on A7 because of the E-mount.
     
    • Like Like x 1
  17. ggibson

    ggibson TalkEmount Regular

    154
    Sep 1, 2011
    OK, I finally had a chance to look at the files here. I focused on subject 1, 2, and 5, mostly comparing like apertures. Even in these unscientific tests, the A7+16-35mm clearly beats the A6000+10-18mm.

    In the center, I found the difference to be the smallest. There the A7 files have only a slight bit more micro-contrast in the edges. Just a bit more definition to everything. The difference is there at f4 through f8, probably the least difference being at f8.

    In the corners, however, the 16-35mm is far and away better, no matter what aperture. It's obvious it's the 10-18mm just losing resolution towards the extreme corners of the frame. More smearing in the foliage, less detail in the bricks and building. The close-up in subject 2 is particularly impressive on the 16-35mm. It just blows the 10-18mm out of the water. The brick texture is incredibly sharp on the former, whereas in the latter all that detail turns to mush away from the center.

    Of course, this is the result that I would expect given the A7 combo is essentially twice the price. All in all, the A6000+10-18mm results are still quite close and I'd guess that for most practical viewing purposes, the difference would go unnoticed. I tend to agree that sharpness as a quality in a lens is quite overrated. The main reasons that I like the 16-35mm better are 1) the focal range is better for a walk-around lens, since it extends to 35mm, and 2) the full-frame format allows you to get shallower DOF at f4 (especially since you can shoot at 35mm as well). Clearly you pay the price in weight and cost, but I'm looking forward to picking up this lens as soon as my wallet can afford it :biggrin:
     
    • Like Like x 2
  18. jai

    jai TalkEmount Top Veteran

    589
    Feb 4, 2013
    Yep, sharpness is overrated. Corner sharpness is way overrated.

    Both combos look really good!
     
  19. quezra

    quezra TalkEmount Top Veteran

    916
    Aug 22, 2012
    And now for the opposite kind of review from DxO rating: Steve Huff's take on the 16-35/4 (he likes everything about it but the size).
     
  20. Lisandra

    Lisandra TalkEmount Veteran

    216
    Jan 28, 2015
    It just fell short compared to the 24-70 f2.8 when I compared them side by side. And theyre almost as big :(