MTF Tests of the Tamron 17-28mm and 28-75mm f/2.8 Di III RXD Lenses

Tamron has a nice set of three f/2.8 zooms for Sony E mount cameras. Today I’m putting out MTF results for the 17-28mm f/2.8 Di III RXD and the 28-75mm f/2.8 Di III RXD. They offer a reasonably priced alternative to Sony-branded lenses, and Tamron has been making good products lately. I’ll admit to a bit of a soft spot for one of them; the old Tamron 28-75mm XR Di f/2.8 was one of the first lenses I owned when I got into photography, and I loved it.

The Di III RXD lenses are priced attractively ($899) and are reasonably sized (under 4″ long and under a pound for the 17-28mm; 4.65″ long and just over one pound for the 28-75mm). They’re smaller and less expensive than the Sigma E mount lenses, and less than half the price of the Sony GM lenses.

But, (and there’s always a but) we know, because we do science, that you usually get two of the three of cost, size, and image quality. We have a low cost. We have a reasonable size. So we expect average image quality. We also know, because humans, that people (including us) are hoping for magic, and we’re going to get low cost, small size, and great image quality. I do have some hope for magic here; the Tamron’s use hybrid and molded aspheric elements, which saves on cost. A few years ago, ‘hybrid’ and ‘molded’ meant ‘low-quality,’ but the technology for these has improved quite a bit.

Images courtesy Tamron, USA.

Let’s see how that quest for magic turns out.

Tamron 17-28mm f/2.8 Di III RXD

Our practice is that 3X zooms are tested at both ends and the middle, while 2X zooms are only tested at the two extremes. (Before you ask, 10X zooms are not tested at all because I don’t have enough crayons and time to explain that yes, they really do suck that much. Don’t get me wrong. I travel with one sometimes; they have their uses, but high-resolution isn’t one of them.)

At 17mm

In the center at 17mm, this thing is ridiculously sharp. For a wide-angle, it maintains sharpness pretty well away from the center, but there’s a lot of astigmatism-like separation between sagittal and tangential lines. Some of this is because of lateral chromatic aberration (astigmatism and LCA have the same effect on MTF), but some are actual astigmatism., 2020

At 28mm

We lose some center sharpness at the long end, but it’s a reasonable trade-off. Astigmatism / LCA effect is completely gone at the longer end., 2020

Overall, this is an excellent performance for an f/2.8 wide-angle lens, but that’s easier to see in some comparisons. We’ll start at the wide end.

17mm Comparisons

vs. Nikon 14-24mm f2.8 ED AF-S at 18mm

This isn’t really a fair comparison; the Nikon is a much wider lens (wider is more difficult to design) and an older lens. Still, the Nikon is the first ‘really good wide zoom’ lens and remains a comparison gold standard. The Tamron is better at this focal length., 2020

vs. Canon 16-35mm f2.8L III USM

The Canon 16-35mm f2.8 III USM is a newer design and a similar focal length. Split hairs if you like, this is the internet after all, but I’d call this pretty damn even. Given the Canon costs more than twice as much, I’d call the Tamron a win here., 2020

vs. Sony 16-35mm f2.8 GM

Of course, this is the crucial question for a lot of folks. How close can you get for half the money? The answer is pretty damned close. The Tamron is sharper in the center; the Sony has less off-axis astigmatism. This would come down to what you prefer or value the most. (Yes, I know it’s the money, but let’s check out the other end before you plop that down.), 2020

28mm Comparisons

vs. Sony 16-35mm f2.8 GM at 24mm

The Sony is at 24mm here, vs. the Tamron at 28mm, because those are the numbers I have on file. This is a pretty good showing for a lens at half the price., 2020

vs. Canon 16-35mm f2.8L USM III

The Tamron holds its own against Canon’s similar wide-angle. 2020


Tamron 28-75mm f/2.8 Di III RXD

This being greater than a 2X zoom, we’ll do it at three focal lengths.

At 28mm

Superb center sharpness at 28mm, and good off-center performance., 2020

At 55mm

In the middle of the range, we see not quite so good center performance, but it maintains off-center performance quite well., 2020

At 75mm

At the long end, however, the lens is quite average. Acceptable, but nothing that’s going to wow anyone., 2020


I’ve stopped using the variation graphs very much because people abuse them so, but in this case, they help explain some (not all) of the weakness in the 28-75 as we leave the short end. First, let’s look (left to right) at the variation of the 17-28mm at both ends, and of the 28-75mm at the 28mm end. Don’t overread this; just look at the left side (the center) of each graph and take away that there’s a small range there. Not a lot of variance., 2020


Now let’s look at the variance graphs of the 28-75mm f/2.8 at 55mm and 75mm. Looking just at the left side (center), you can see there’s a lot more variation. Since we’re showing you the average of 10 copies, well, there’s a pretty wide range of what you might see in an individual copy. Don’t get me wrong; none of them are ‘wow awesome’ at the long end, but some are pretty good, and some are kind of soft., 2020

It’s Still Pretty Good

Before we get into complete comparisons, let me talk some of you down off the ledge. The 17-28mm, which was just great, probably set you up for some disappointment. The old version 28-75mm f/2.8 for SLRs was a very popular lens; as I mentioned, I carried it for years. This new lens is WAY better than the old one was.

Here are some old versus new comparisons; it’s obvious the new RXD smokes the classic old lens, as it should. It’s also clear that the old version also was weaker at the longer end. Despite this, back in the day, many of us found this a good lens, including me. (Granted, that was on cameras with lower resolution.), 2020, 2020, 2020


We have three focal lengths to compare this time, so I’m just going to do a couple.

vs Canon 24-70mm f2.8 L  II USM

This is not a ‘which will I choose’ comparison. Instead, I’m just trying to make a comparison to a widely known standard in this focal length.

Wide End

The Canon is at 24mm and the Tamron at 28mm; make of that what you will. This is near enough to identical that I doubt you’d notice any difference testing them on cameras., 2020

Long End

Like most 24-70mm zooms, both are a bit weaker at the long end, although the Canon clearly resolves better in the center half of the image., 2020

vs. Sony 24-70mm f2.8 GM

Let’s remember the Sony GM is more than twice the price of the Tamron and is at least as good as any 24-70mm f/2.8 made.

Wide end

Slight advantage to the Sony here., 2020

Center range

The advantage is more pronounced now. I will point out we pick ‘mid-range’ mostly based on the focusing ring throw, but the Sony is at 35mm compared to the Tamron’s 55mm., 2020

Long end

As with most 24-70 zooms, things are weaker at the long end for both or the more so for the Tamron., 2020

So What Did We Learn Today?

A couple of things, I think. First, from a resolution standpoint, the Tamron 17-28mm f2.8 Di III RXD is really good; amazingly good when you factor in price and size. If I was looking for a wide zoom in E mount, and 17mm was wide enough, I’d give it strong consideration. The 17-28mm, in particular, is about as close to magically violating the “price, size, optical quality” pyramid as you’re likely to get.

The Tamron 28-75mm f2.8 Di III RXD is good lens at a reasonable price. If I needed a zoom at this focal length and price was a prime motivator, it’s a great choice. I’d consider it more strongly if I also carried a 70-200mm lens or an 85mm prime that I could use at the long end. You could certainly buy this zoom and a good 85mm prime for less than the price of the brand-name zoom.

Some people aren’t going to be happy with the resolution at 75mm, at least wide open. But as a walk-around or vacation lens, I expect the vast majority of folks will consider it good enough at a bargain price. Let’s emphasize that in the comparisons, almost every lens I compared to the Tamron was at least double the price. From a pure resolution standpoint, it’s nearly as good as lenses costing double the price.

As always, this was a resolution test. It was only a resolution test. If you’re interested in these lenses, you’ll want to look at what reviewers have to say about all the other factors that should be considered. The one thing I WOULD take away from this resolution test, though, is there’s some very real sample variation at the long end of the 28-75mm. Some reviewers are going to have a good copy; some aren’t. So I expect you’ll see some differences of opinion among them.

If I carried the two zooms in my bag, I’d try to use the 28-75 rather than the 17-28mm at 28mm. Knowing which lens does what best is a big part of what makes these tests worthwhile. If you’re going to collect the whole set of Di III RXD zooms; hopefully, tests of the 70-180mm lens are going to show it’s a standout at 70mm, which would nicely cover you at that range.


Roger Cicala and Aaron Closz

August 2020

Author: Roger Cicala

I’m Roger and I am the founder of Hailed as one of the optic nerds here, I enjoy shooting collimated light through 30X microscope objectives in my spare time. When I do take real pictures I like using something different: a Medium format, or Pentax K1, or a Sony RX1R.

Posted in Equipment
  • Kambiz Dardashti

    Hi Roger. are all lenses tested wide open? are MTF figures here all shot at f2.8? do you ever also shoot them at f8? given these lenses are often used for landscape, that may be very usefull also.
    thanks for all your work.

  • Roger Cicala

    It’s like looking at a house as you walk past and saying “Wow, that’s a well built house because it’s got bricks on the outside”.

  • I recently read a review that concluded that a lens had “great build quality” and “enormous copy variance. We evaluated four copies to find one good one” and I could only roll my eyes. I really don’t understand why so many people equate weight with build quality. I view build quality as the reliability of a lens to arrive optically aligned and stay that way.

  • iKonOkLasT

    Some edits are needed to the posted graphics and text.

  • Trey Mortensen

    Hey Roger is back! I can’t wait for the inevitable R5 teardown. I’ll have my popcorn ready for the comment section 😀

  • Turniphead

    Hi Roger,
    Excellent work as ever; thank you for posting this!
    One correction; in the section entitled:
    “28mm Comparisons
    vs. Sony 16-35mm f2.8 GM at 28mm”
    The posted image pair is for the lenses at 16/17mm, not at 24/28mm.
    Thank you,

  • Nick Podrebarac

    It’s interesting how much variation there is with the 28-75. I wonder if that’s at least a part of why opinions on it seem so polarized (staying out of the bokeh holy wars…). It’s impossible to say where my copy falls in that spectrum, but I’ve never been left wanting for (center) sharpness, even at its weakest focal length (75mm). If I have a ‘good’ copy, though, then I could understand why some people aren’t impressed; if I have a ‘bad’ copy, I suppose I’ve just got low standards 🙂

  • Roger Cicala

    We plan on it. Things are slow because we’re working with a short staff these days.

  • Athanasius Kirchner

    I’ll add my voice to those saying that we missed these tests. I’m sure that you and Aaron (and the poor hapless victims that end up being recruited to help you in the “mad scientists’ lab”) have had plenty of things to do, but please, never forget that you have loyal fans out there ?

    Is there any chance that you might test the Tamron primes for E-mount, Roger?

  • Athanasius Kirchner

    Ooohh, I’m already hyped for that one!

  • Athanasius Kirchner

    What you wrote is coherent with my experiences with both. I’d take the Sony, even if it resolves less in pure terms, because of the OSS and larger range (well, and that smooth metal barrel that adds literally nothing, but feels so good ?). People wanting a faster aperture or higher center resolution should take the Tamron, which is outstanding for the price.

  • Notix
  • Matti6950 .

    Hi Roger, i missed your tests so much! Hope we get more of this soon

    I have the Sigma 14-24mm DG DN and it’s absolutely amazing. Always been curious how it will perform here, but i suspect well. It definitely beats the crap out of my nikon 14-24mm (who might have been a dud i think, often soft images without explanation.)

    I also now own the Sigma 24-70mm DG DN (next to Sony 24-70mm GM). A few findings:

    Sigma has much less chromatic aberrations (those special elements work)

    Sigma has a lot of vignetting at 24mm (at f8 it starts to be ok), Sony better here.

    Close to camera, Sony a little sharper, far away from camera Sony a lot sharper. ‘landscape sharpness comparison:

    24mm: Sigma much sharper in corners stopped down to F8, DX frame sharp wide open on Sigma.

    28mm: wash, both excellent

    35-46mm: Sony is prime sharp here. Sigma is also excellent, but sony has lead here.

    70mm: My Sony copy is terrible here (i think even worse then Tamron 25-75mm). Just cannot get any meaninfull sharpness at 42 megapixels, stopping down doenst help much. Sigma on the other hand is very sharp here, not prime sharp, but sharpest among many 24-70mm’s i used. Sharper in center then nikon 24-70mm VR, and about same in corner (stopped down a bit).

    I’m also a 24mm lover, so Tamron lens (28-75mm) is not for me. I also like build quality and buttons to switch stuff, wich Tamron has not included. For that reason the Sigma is very good lens. It cost 1199 euro (200€ less then DSLR version – although that one had OS). With 10% discount i payed 1080€. Stunning heh for a lens costing half price of Gmaster, it’s actually better for me?

  • Chris Jankowski


    Tiny, nit picking correction suggested:

    This title:

    17mm Comparisons
    vs. Nikon 14-24mm f2.8 ED AF-S at 16mm

    should read:

    17mm Comparisons
    vs. Nikon 14-24mm f2.8 ED AF-S at 18mm

    if the title on the graph just below is to be trusted.


    Thanks again for the enlightening test and the explanation of the results.

    By the way, do you have now enough of the various EF lenses to start test ting them? Nudge, nudge… (:-)).

  • As a nightscape photographer who stares at infinity-focused, pin-dot subjects all day/night, this lab test confirms my real-world findings/suspicions: the 17-28 is INSANE in the center, but falls apart just a little bit off-center and in the corners.

    If you’re looking for a perfect wedding journalism/portraiture lens to take you wider, without taking up a ton of space in your bag or $$$ from your gear budget, the Tamron is the absolute champ.

    If you’re looking for the absolute champ of nightscape photography, though, the Sigma 14-24 2.8 DN is the real pixel-peeper’s home run you were looking for… (Why not compare the Sigma DN against the Tamron 17-28?)

  • Roger Cicala

    Yes, our 20mm is a bit longer than the lateral edge, but not quite to the corners. There could also be some field curvature going on, I didn’t run curvatures on these.

  • Roger Cicala

    It means it varied more. I don’t make any other conclusions unless I actually know the cause.

  • Vladimir Gorbunov

    Well, does it mean that 16-35GM is badly underdesigned with its wild centering figures at 35 mm?

  • Calphate

    hmm interesting… I’ve done the comparison myself and noticed that Tamron 17-28 is generally very sharp except for the extreme corners. At the corners it drops like crazy and it’s certainly not reflected in your MTF chart. But then I noticed that your results only cover <20mm image height, so maybe that's the reason?

    Anyway, I really don't care about the four little corners that much so I sold my 16-35GM.

    And your test at 28mm is exactly what I was seeing: Tamron is a lot sharper than GM that when you crops it to equivalent 35mm it's still quite comparable. GM does have two advantages: its entrance pupil is larger so there is visibly more bokeh; the sunstar is so much better than the Tamron. Other than that I don't really miss it. The flare/ghosting control is not so good which is deadly for a landscape lens!

  • Roger Cicala

    75-180 next if I can get enough. We don’t usually test 5X or greater zooms, the variation is so large they have to be tested at multiple focal lengths and really should do 20 copies. It’s just too much of a time investment.

  • The_Incomparable_Douche

    Roger, please test the Tamron 35-150mm next!

  • Roger Cicala

    I thought the 17-28mm was really low variation, particularly with centering. Tilt, though, that’s almost inevitable in a wide-angle, wide-aperture zoom.

  • obican

    So, Tamron is slightly sharper at center and the extreme corner tangentially but Sony performs better at sagittal lines and Tamron drops faster in the mid field? I guess Tamron should easily win at f/4 and I’d expect them to perform similarly at f/8 anyway. Long end performance seems even closer but Sony’s FOV goes further at both ends. Am I missing anything important?
    I guess Sony ZA is still a great choice for people who want a slightly wider focal range and OSS while Tamron has a faster aperture and is lighter and performs better optically, but I don’t think it’s by that much.

  • Vladimir Gorbunov

    Thank you for that information, it’s really appreciated!

    How did you like the variation of 17-28 regarding its centering/tilting? My own copy has one slightly soft corner at 28 mm, which is visible on A6500 (sic), but not visible on A7III. But when I looked at your measurements of 16-35GM, I realized that ideally centered fast UWA zoom is something fantastic.

  • Roger Cicala
  • Roger Cicala
  • Roger Cicala

  • obican

    Useful as always, team Lensrentals. You are a blessing to the photography community.
    The lenses however, were a huge disappointment for me from the moment they were announced. Surely the performance is great, regardless of the price, but as an interior photographer, I need every extra degree of FOV at the wide end and 17mm just won’t really cut it. Before anyone says 16mm vs 17mm won’t matter much, it does when it does. I’d really rather see something like a 15-25/4 instead of a 17-28/2.8 and if I were to stick with my Sony, I’d probably re-purchase my beloved 16-35/4. Do you have any data for that lens btw? That’d be a nice addition to this test since it’s quite similarly priced to the Tamron, iirc.
    I’m not in the target audience for the 28-75/2.8 anyway but if I were, again, I’d make the same comment. 28mm is just not wide enough for a general purpose zoom for me. I get that everything is a compromise in lens design and wide end (and the long end with their 70-180, which is once again the more useful end) is what Tamron sacrificed to keep everything else on a higher level but I’d really rather see this trio of lenses with a more useful range and possibly smaller size while sacrificing the aperture.
    But I’m probably the odd one out here, not going for the fastest aperture available.

  • Roger Cicala

    Thank you Clayton, I’ll get on the graphs. And yeah, I agree with you, that’s a good point about quality. Unfortunately, too many internet experts confuse ‘build quality’ with ‘it’s got a metal shell’.

  • Clayton Taylor

    Great to see new lens tests, and nice job, Tamron!

    BTW, I have always thought that there were 4 aspects to consider when evaluating a lens – to “cost, size and image quality” add “build quality” (aka ruggedness). To me, that is where your tear-downs and reliability reports have become possibly more important than the MTF charts.

    One thing – up in the 17-28mm lens section, at 28mm vs. the Sony 16-24, you have the Sony 28-70 vs Tamron 28-75 graphs displayed.

Follow on Feedly