Tokina AT-X 14-20mm f/2 PRO DX lens tested at DxOMark

Tokina-14-20mm-f2.0-AT-X-Pro-DX-lens
DxOMark tested the Tokina AT-X 14-20mm f/2 PRO DX lens ($899) or Nikon F-mount:

tokina-at-x-14-20mm-f2-pro-dx-lens-tested-at-dxomark-2 tokina-at-x-14-20mm-f2-pro-dx-lens-tested-at-dxomark
DxOMark's conclusion:

Although the range of focal lengths is somewhat limited, especially when compared with its closest rival, the Sigma 18-35mm (28-45mm equivalent) F1.8, the wider 14-20mm F2 (21-31mm equivalent) Tokina is an entirely different proposition. Indeed, with that f/2 maximum aperture, there really isn’t anything else to compare it with. If the choice of 21 and 24/28mm focal lengths appeal, then the Tokina’s optical performance won’t disappoint.

This entry was posted in Nikon Lenses and tagged , , . Bookmark the permalink. Trackbacks are closed, but you can post a comment.
  • Henri De Vreese

    21-31 is very short imo, I don’t see many people choosing this lens over the Sigma. I think a Tokina 11-16 combo with the Sigma 18-35 is a better choice for people interested in this focal lenght.

    • David

      Maybe a lot of people don’t choose. I had the 11-16 and a Sigma 17-50, and I imagine people chasing the best DX image quality will pick up both. Gotta say, when traveling (for instance, the Grand Temple in Bangkok) I would have liked a little more flexibility. I would have been willing to snip off a few mm at the wide end for more at the long end, even if it was only a few mm. It’s a lot easier to crop from 30mm equivalent than 24.

  • fanboy fagz

    weird focal range. not wide, and not middle. numbers look nice but as I have seen, dxomark numbers mean jack shit from what ive seen on the screen. im sure the lens will be useful for some though.

    im sure with that range it will perform nicely. kinda like the sigma 24-35 which too has a short range but very sharp and also between wide to mid.

    • Eric Calabros

      Yea, their numbers dont even represent their own test results
      18mm wide open

      https://uploads.disquscdn.com/images/b8f4f363d5b91679b8fb0bb620ebbf02d7e132287669e1c9cedd7162b3e48dac.jpg

      • El Aura

        Which of their numbers are you referring to? The vignetting, CA, distortion or transmission values? Or the (integrated) resolution stopped own (P-Mpix)?

        I don’t think that the vignetting, CA, distortion or transmission values contradict the graphs you posted. Nor can graphs of the performance wide-open really contradict resolution numbers measured stopped down.

        • Eric Calabros

          P-Mpix is rated at f/2. So they’re saying Tokia is just a little bit softer than Sigma wide open. But the graph shows the difference is more than 17 vs. 18.

          • El Aura

            You are confusing the DxOmark Score with the P-Mpix rating. The latter is for the aperture at which the maximum resolution is recorded: “Then, for each focal length, we select the maximal value of sharpness over the range of available apertures. We average this value over the whole range of focal length to obtain the DxOMark resolution score that we report (in P-MPix).” (https://www.dxomark.com/About/Lens-scores/Metric-Scores)

            The DxOmark Score is grading on a scale. A given resolution (maybe other metrics are also included) is assigned a higher score the faster the f-stop is (and probably the more challenging the focal length). Thus at which f-stop the maximum DxOmark Score is achieved depends how the two curves of resolution over f-stop and score per resolution over f-stop intersect. Recently, most lenses achieved their maximum f-stop wide-open. But a lens that is mediocre wide-open can easily get it maximum DxOmark Score stopped down somewhat.

            • Eric Calabros

              Even your defensive explanation proves that their numbers are misleading. The sharpness gap between these two lenses is bigger that what the numbers say.

            • El Aura

              Defensive? Who is trying to defend his incorrect apples to oranges comparison here?

            • David

              blah blah blah, harder to design a wide angle zoom, blah blah blah.

              Look, the Sigma is a friggin masterpiece, and we all wish Sigma would do a new version of the 17-50 or something and give us something that starts at 16 and goes to something acceptable. The Tokina may not come close to the level of greatness the Sigma possesses, but the reality is that if you shoot Nikon you pretty much HAVE to go Tokina if you want a decent wide-angle ANYTHING, much less a zoom. The real question is: is it really THAT much better than the 11-16 or 11-20 (plus the wider aperture) that it makes up for such a restricted zoom range?

            • Eric Calabros

              its not apples to oranges. Its apple number via apple graph. At all focal lengths, at any aperture, Sigma is much better performer than what so called P-Mpix suggests. End of story.

            • Nikita

              he wasn’t defensive, you are.
              play fair

            • Eric Calabros

              Playing fair? You mean what DXO refuses to do? 🙂 they tested a bad copy of 24-70E which their review was based on, which was clearly contradicting with Roger’s test results. Thats how they play.

            • Nikita

              No, I meant accusing El of being defensive.
              Name calling is not a fair debate tactic.

            • Eric Calabros

              What name calling? Its describing the way someone argues. I’m offensive, and he is defensive. Thats it.

    • El Aura

      The first real wide-angle zooms were 20-35 mm. This very much is a wide angle zoom.

      • fanboy fagz

        Wide, yes, ultra wide no. 21 is wide but its not sitting im the 3 ranges we have. 17-35/24-70/70-200 (variances of focal length are available but just gave a rough example)

        • El Aura

          21 mm sits inside the 17-35 mm range. Not sure what you are referring to.

          • fanboy fagz

            it is not a an ultra wide angle lens. it only goes to 21mm. it isnt wide enough for an ultra fl and not midrange. leave it. youre missing the point.

            • El Aura

              I grew up the definition that 28 & 35 mm were moderate wide-angles and everything below that were extreme wide-angles. And until recently (roughly ten years ago), very few people ever had a lens going wider than 17 or 18 mm (the only thing on offer below that in almost all lens portfolios was a 14 or 15 mm prime). I am fine with with calling 24 mm a transition WA. But everything from 21 mm and below falls into the extreme WA category.

            • fanboy fagz

              move on…

  • Photobug

    Yea, this is a weird and short focal length. Don’t understand the marketing when you have the Sigma 18-35 lens.

    • David

      Because the Sigma is awesome but I like something wider. If the Sigma had been an 16-24 I think it might have been more popular. Considering my old D7100 had a ~15mp 2x crop it would have been like having an insanely sharp 16-35 on DX.

  • TwoStrayCats

    What exactly does the “sharpness – 17P-Mpix” mean?

    • Jeffry De Meyer

      It is the averaged sharpness of a lens over its entire focal range and all apertures.
      17 on a 24megapixel aps camera is a good score for a wide lens like this.

      • El Aura

        It might be fun to guess what P-Mpix might mean, but when you pass off your guess as fact, you are deceiving the audience. It takes exactly two clicks to get from dxomark.com to the page explaining what P-Mpix represents:
        It’s a resolution value extracted from a (full) MTF curve integrated over the whole frame, giving less weight to the corners. The reported value is the best one found over the range of tested aperture. For zoom lenses, the average of all tested focal lengths is calculated.

        The key here is that the P-Mpix value represents the performance at the optimal f-stop and thus says little about the performance wide-open (except for slow lenses where due to diffraction, the optimal aperture is very close to wide-open).

        • Jeffry De Meyer

          Pretty much what I wrote except for the apertures, it was from memory as the popup thingies didn’t work on the ipad

    • dabug91

      The P stands for perceived or perceptual megapixels. The P-Mpix will never be quite the actual MP count of the camera body it is tested on, but generally the closer P-Mpix is to the camera body’s MP, the sharper the lens tends to be throughout the frame.
      So for instance, a lens may score 17 out of a camera’s maximum 24MP, but if you put the same lens on a 16MP camera body, its score may change to 10 or 11 P-Mpix for that particular body.

  • Lex Cross

    This math is pretty bad. 18-35 = 27 to 52.5. NOT 28-45. And 14-20 = 21-30. NOT 21-31.

    • Nyarlathotep

      APS-C Crop factors are not consistent. It depends on the camera APS-C sensor. Over the years Nikon APS-C crop factor has ranged from 1.558x to 1.519x, 1.532 being the crop factor for the sensors in the current Nikons. Enough to vary 20mm to either 30 or 31 and 18 to 27 or 28. Plus Canon also throws things for a loop since there crop is ~1.6x. And being the 800 gorilla, the numbers can often be derived from their crop.

      Of course the 35 crop conversion to 45 was just a straight mistake it appears.

  • tobi

    Looks really nice.. may be worth considering an upgrade to my 12-24 F4 …
    Especially for Night cityscape ..

    LOL and I see the normal questions re Dxo’s PMP and the usual confused and false replies 😉 .. My advise to the people who want to know what PMP is is to go to the DXO website and read their explanation. and if you are still confused then rage and say that DXO is crap. Most of the noisy ones do 😉

  • Nikita

    pretty sharp, it looks like.
    still, I wish they’d do an UWA prime, 10mm 2.8 (w/removable hood)

  • Back to top