Geek Articles

MTF Lens Tests of the New Tamron 70-200 f/2.8 SP Di VC USD G2

There’s been a bit of rumbling lately about the new Tamron 70-200mm f/2.8 G2 zoom. Not just because it is a frontrunner in the 2017 Most Initials in the Name Award, but because a lot of people are saying it’s really quite good. The previous Tamron 70-200mm f/2.8 was a good lens, but since they only added two more characters to the name, I wasn’t sure this one would be hugely better; yet, people said it was.

When you consider that it sells for a very reasonable $1,300 compared to $2,800 for the new Nikon 70-200 f/2.8 FL ED VR, $2,100 for the Nikon 70-200 f/2.8 VR II, and $1,900 for the Canon 70-200 f/2.8 IS II lens, well, it seems worth evaluating. So we decided to put ten copies on our optical bench and take a look, even though 70-200mm zooms are a pain to test.

I should mention, just to avoid confusion, that this is the G2 Tamron SP 70-200 f/2.8 Di VC USD, not the slightly less expensive ‘not-G2’ Tamron SP 70-200 f/2.8 Di VC USD, nor the much less expensive Tamron 70-200 f/2.8 Di LD (IF) Macro lens. Keeping up with Tamron 70-200mm lenses is only slightly less complicated than keeping up with Nikon APS-C camera bodies.

So About the Testing Stuff

Well, as always, we’ll be showing you the MTF results obtained by averaging ten copies tested on our optical bench. To give you some idea of copy-to-copy variation I’m also going to use the Full Frame MTF displays that we’ve talked about in this article and that one. Of course, if you are already planning your comments to show the tests are all invalid and your favorite brand is wayyyyy better than this, you probably should at least glance at those posts. I mean, you can’t claim the data is all wrong unless you at least understand the methodology. Oh, wait, of course you can.

First, let’s look at the average MTF chart for the Tamron SP 70-200 f/2.8 yadda yadda G2 at 70mm, 135mm and 200mm. If you don’t speak MTF much, just remember higher is better, lines of the same color close together are better, and I’ll point out other little things as we go along.

At 70mm

Olaf Optical Testing, 2017

The most impressive thing that jumps out at me is how the Tamron is maintaining high MTF all the way out to the edges of the image plane. Zooms generally fall off as they approach the edges, but the Tamron doesn’t. There is some astigmatism-like separation of the sagittal and tangential curves off-axis, but it’s not severe. (We say astigmatism-like because it can be caused by true astigmatism or lateral chromatic aberration. From here on out I’ll just call it astigmatism.)

At 135mm

Olaf Optical Testing, 2017

At 135mm the lens is actually resolving better than it did at 70mm. A fair number of 70-200 zooms are weakest here at the mid-range. There is less astigmatism at this focal length, too.

At 200mm

Olaf Optical Testing, 2017

Performance drops just a bit at 200mm, becoming similar to what we saw at 70mm, which is still quite good. You’ll notice at the center the sagittal and tangential curves aren’t together, which they should be in theory. That’s because our bench lines up with the geometric center of the lens. At 200mm and longer, sometimes the optical center is 1 or 2mm away from the geometric center. For those lenses, the ‘no astigmatism’ point is sort of a circle around the center. It would make no difference in a photograph, but it makes a slight difference in the tests.

Overall, though, the Tamron puts in a very good performance. These are excellent MTF results.

How About Some MTF Comparisons?

That seems pretty reasonable. Since most people will be looking at this as an alternative to Canon and Nikon 70-200mm zooms, that seems the comparison to make. Let’s start by putting it up against the Canon 70-200 f/2.8 IS II, which is an excellent lens.

Tamron vs. Canon at 70mm

Olaf Optical Testing, 2017

The Canon is clearly better in the center at 70mm, especially at higher frequencies (fine detail resolution). Away from the center and towards the edges, though, the Tamron catches the Canon and is better at the edge of the image field.

Tamron vs. Canon at 135mm

Olaf Optical Testing, 2017

At 135mm, where the Tamron is strongest, the two lenses are pretty even. The Tamron is a bit better at the edges, but otherwise, there’s little to separate them.

Tamron vs. Canon at 200mm

Olaf Optical Testing, 2017

The story is much the same at 200mm. From a pure MTF standard, the Tamron is as good in the center and perhaps a bit better at the edges.

We need to provide a comparison for Nikon shooters, too, and unfortunately for the Tamron that means a comparison with the Nikon 70-200mm f/2.8 FL ED VR, which is currently the best 70-200 zoom on the planet.

Tamron vs. Nikon at 70mm

Olaf Optical Testing, 2017

OK. The Nikon’s better at 70mm. What did you expect?

Tamron vs. Nikon at 135mm

Olaf Optical Testing, 2017

And there’s a big difference at 135mm, too, particularly in the higher frequencies, although the Nikon does fall off at the edges enough for the Tamron to catch it there. But let’s face it, most people are most interested in center sharpness at these longer focal lengths, and no zoom in this range is as sharp as this Nikon FL ED VR (including the VRII Nikon).

Tamron vs. Nikon at 200mm

Olaf Optical Testing, 2017

At 200mm things are much more even. The Nikon is a tiny bit better, although the central false astigmatism of the test masks it somewhat. But it’s pretty close, it really is.

Let’s Look at Variation (a Bit Differently)

At 70mm

Over the years we’ve tried to summarize variation a lot of different ways, none of which worked perfectly. So I’m just going to start showing you the actual tested copies with actual data so you can see how each varied. Here are thumbnails of the MTF full-frame displays at 70mm for all ten lenses (dark blue is best, green adequate, yellow-green getting noticeably soft).

One thing to mention, this is actual data, and you’ll notice some white areas. This is where we removed data points, something we rarely have to do. This lens has a very easily moved vibration reduction element that does not lock down. That has no effect when you’re taking a picture, but if the lens is on a very delicate optical bench and a truck drives by or someone drops a 50-pound box of supplies in the next room, the vibration showed up in the readings very obviously, so we deleted that data. We would usually have repeated the run, but in this case we were under time pressure to get these back in stock and I didn’t look at the data until after the testing was completed.

Olaf Optical Testing, 2017

If you want my verbal summary, three copies are pixel peeping perfect (all blue), three are photography perfect (dark blue with some off-axis light blue), one (top middle) is adequate. Three have some issues at 70mm. They’re OK, but if you chart-tested carefully you could probably tell the three weren’t quite as good as the others.

Before you scream ‘throw this one or that one out,’ let’s explain this a little further with a couple of important points. I’m showing you ten lenses right out of the box, as they ship and that’s what we do. Second, let’s look at them at other focal lengths because they’re zooms. And third, I’ll give you some comparison zooms in an addendum so you can see what 70-200mm zooms vary like in the best of circumstances. Then you can scream ‘throw them out’, just scream it about all zooms as I do.

At 135mm

The same copies are in the same positions so you can look and see how they change at different focal lengths. Don’t try to kill yourself reading the shrunk down test run numbers in the thumbnails.

Olaf Optical Testing, 2017

Overall, you can see there’s less variation at 135mm with this lens, they’re almost all pretty good. That lower left one, that was weak at 70mm is still weak at 135mm, though.

At 200mm

You get to see the more ugly underbelly of my testing life because my last testing copy was grabbed for rental before I could complete it’s testing at 200mm, so we only have nine results here. Other than the gap, they are again in the same location by copy. This is not what usually happens when we test, but usually, we have at least many dozen copies and getting them for testing is not a strain. With this lens, we have less than two dozen right now, and they’re in and out of stock pretty quickly.

Olaf Optical Testing, 2017

If nothing else, this should give you some understanding about why I roll my eyes when someone says they want a perfect copy of a zoom. By my reading, there’s one of these nine that is excellent at all focal lengths. It’s the nature of zooms. But the good news, since most people shoot 70-200mm lenses at the longer range, is that at the majority are good at the long end, in fact the variation is quite low.

And for those of you who want to make rash statements about what you expect of your zoom lens (and remember, expectations are a down payment on disappointment) take a glance at the addendum where I’ve put similar charts for a very, very good lens; the Canon 70-200 f/2.8 IS II. And for you fanboys, no the Nikon and Sony 70-200s aren’t better than the Canon. Zooms are zooms, not matter how much you want to bang your fist on the table and demand that they shouldn’t be.

So What Did We Learn Today?

That if you’re looking for a 70-200mm zoom and the Nikon or Canon offerings are a bit too pricey, the Tamron 70-200mm f/2.8 G2 is worth looking into. From an MTF standpoint, it’s about as good as the Canon 70-200mm f/2.8 IS II, which means it’s as good as the Nikon 70-200mm f/2.8 VR II. It’s not as good as the Nikon 70-200mm f/2.8 FL ED VR, but then again, nothing is. And the price difference could buy you a nice second lens.

As always, I’ll repeat, the MTF results make it worth a look. I’m not a lens reviewer, and many things like handling, color, flare, autofocus speed, and accuracy are going to matter as much or more as MTF results. But there are plenty of lens reviewers who will tell you that stuff.

But as far as optical resolution goes, the Tamron is excellent at a really excellent price.


Roger Cicala and Aaron Closz


April, 2017



Addendum: Canon 70-200 f/2.8 IS II

I just put this here, so you have a comparison. Again, each copy is in the same position in the graph for each focal length, so the top left corner is one lens, etc.

I just used thumbnails of 9 randomly selected copies because nine fits in the images better. I’ve got dozens and dozens, and these are good representatives. And if I replaced them with similar ones from other 70-200 zooms, it would look about the same. But making these takes time and time isn’t something I have enough of. But if you notice one that’s perfect at every focal length, feel free to point it out.


At 70mm

Olaf Optical Testing, 2017

At 135mm

Olaf Optical Testing, 2017

At 200mm

Olaf Optical Testing, 2017

Author: Roger Cicala

I’m Roger and I am the founder of Lensrentals.com. 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 Geek Articles
  • ipdouglas

    I believe what most consumers/customers/photographers want is some assurance that what they buy is a good sample and does not have any faults. We cannot measure other than by using an item for its intended purpose. Additionally most of us are not in any position to compare like with like?
    I have a Nikon 70-300mm VR that I always felt was poor and never reflected magazine (and such-like) tests. My wife has a Nikon 70-300mm VR which is outstanding? It is significant that my wife’s copy has a USA serial number. This indicates that Nikon are streaming the ‘best’ to USA or that poor version (in the factories) are not being fixed or destroyed allowing poor versions onto the street.
    It is significant that Nikon seem unable to carry out effective Quality Control and Quality Assurance in recent years with many recalls and other unacknowledged issues plaguing Nikons.
    I am not sure how us the consumer/customer/photographer can be confident of buying lenses unless we pay much higher prices for premium/professional products?

    As they say … you get what you pay for.

    What we need is a bulk buyer that sells on tested ‘good’ versions at a price premium

  • Patrick Chase

    OK, here’s one option that Roger is probably too circumspect to raise:

    LensRentals screens their lenses pretty well, so if you want a known non-lemon one option is to rent one, test it just to be sure, and “keeper-ize” it. I just did that with a 16-35 f/4L IS.

    I’ve also used rented lenses to serve as reference points for purchased ones, for example I weeded out a much worse than average 24-70 f/2.8L II (the only truly bad “L” lens I’ve ever purchased) that way.

  • Tony Thompson

    Thanks very much Roger.

    Yes it’s pretty uniform and to be honest I like how it’s not super loose the zoom

    This is definitely optically way better than my first copy so I’m glad I don’t need to return it

    Thanks again !! 🙂

  • Tony, I think that’s the normal sound for this lens. The VC doesn’t lock in place so when the electrons go through the magnets it clicks into place. Seems to work very well though.

    I never worry about stiff zooms as long as it’s fairly uniform through the range. When there’s a ‘sticky place’, as in the resistance increases right at 135, or somewhere, that’s a problem sign sometimes.

  • Greg Dunn

    Not all Sigmas. The 50-100 does not zoom in the Canon direction, as it constantly reminds me every time I shoot with it. Maybe the others do, but I don’t own any of them.

  • Tony Thompson

    Maybe more like a very low clicking sound as it engages/disengages.

  • Tony Thompson

    Roger I’d really appreciate some help if you can

    I’m on to my second copy of the lens (first was soft as mush)

    Second seems really good upon initial shots though the zoom ring is noticeable stiffer and before the VC engages it mades a very short audible rattle esque sound and after it disengages, is this something you have noticed on your copies?

    I had a Sigma 105 macro that done the same thing when the OS engaged but ALOT louder!

    Thanks for any help

  • milkod2001

    Price is very close, no point to go with Tamron if you use Canon body. With Nikon’s latest 70-200 at: €3199.99 Tamron is fantastic deal.

  • Sergiu Mosoia

    Me too (and I had the 5DS as well) – btw, Sony is not “better” than both Canons, at all. Different, yes. But this is not about cameras, but about lenses – and I can tell you that I am experiencing some strange issues with the 24-70 f/2.8 II on the MK IV… the same lens having none on 5D III…; 70-200 II works the same on both.

  • Sergiu Mosoia

    Me too – btw, Sony is not better than both Canons, at all.

  • Thanks Roger.

  • Julian, we constantly doing that – it’s part of the reason we test them when we get them. We rarely see a change optically unless the lens has either been dropped or we see something else going wrong, for example it feels a little rough zooming or focusing, etc.

    But very consistently we have a customer tell us they dropped the lens, but it seems fine and then testing shows it’s not fine optically. The same thing when a lens comes back with a dented filter ring, etc.

  • silmasan


  • As always Roger, thank you for MTFing us -Again. I love your honesty and humor. One thing I would love to see is a follow up in about a year or two on how these lenses, actually all the lenses you test, hold up to usage. It is one thing to crack open a new box and bench the babe but quite another to see how the lens fares, mechanically and optically, after it has been out in the real world. Thanks.

  • Mike A

    Downsides of a 5DS R – compared to what? Compared to a 5D Mk3 – same DR, better AF, down-sampled it has superior DR and ISO performance. Compared to Mk4? The 5DS R has visibly more detail, sharpness and overall “pop”. It appears you don’t shoot them side by side and compare. For reference, I own a 5DS R, 5DMkIV & A99II. BTW – the Sony is better than both Canons…

  • Joshua A

    Thanks for the reply mate.

    It’s a terrible shame that thorough lens testing like you do here with OLAF is so time consuming and expensive. I’m an avid Pentaxian with a significant investment in the system, and use it professionally as my main system, and its a shame that Pentax doesn’t have the market share of the other Big 3 mounts do (Canon EF, Nikon F and Sony E).

    I’m just intrigued to know how the new DFA* 70-200mm f2.8 measures up to the competition. From personal testing and use, the lens seems incredibly sharp across the field, similar to the Tamron 70-200mm f2.8 G2, but with an equalised test like OLAF, its near impossible to actually compare it to it’s peers under lab conditions.

  • Patrick Chase

    Olaf Optical also drives rental traffic in a nontrivial way. You can safely assume that none of my expenditures would have happened without your blog. I’d had bad experiences in the past with rental houses that had very little understanding of the products or control over their stock, and your blog was what made me rethink my resulting aversion to rental.

  • Troy Phillips

    Very good thanks for all you do.
    I bought the Nikon 70-200 after reading your review and several others. I’m good light I really like it. It’s replacing my old Nikkor 80-200 ads-d. The motor went out. In low light the keeper rate is about the same. But with the new lens I get fewer tack sharp keepers. The old lens either hit solid or missed good. Lol if that makes since. I shoot live music in low light a lot. The new lens gets totally confused with smoke it the air.
    Guess I’m still in love with the old afs-d . Oh and at close focus the old 80-200 afs-d is a true 200mm and the new fl lens is about 180mm .

  • Troy Phillips

    Love it ….. lol.

  • Duncan Dimanche

    the canon is 1800$… the newest one that is

  • Patrick Chase

    Here’s a concrete example: The Sigma 85 has slightly less “exotic content” than the 70-200 f/2.8 with 4 exotic elements in the Sigma vs. 6 in the zoom.

    If you look at Roger’s measurements the Sigma at f/1.4 matches the zoom at 70 mm and f/2.8 in the center and spanks it off-axis. By the time you stop the prime down to f/2.8 there is absolutely no comparison. That doesn’t mean that the Canon is a bad zoom, it just means that variable focal length isn’t free and you shouldn’t expect leading-edge performance from any remotely affordable zoom.

    Which reminds me: IS isn’t optically free either.

  • Patrick Chase

    It’s all relative.

    Adding exotic elements in zooms makes them sharper than zooms without, and expands the range of applications they can serve. Adding similar exotic elements to primes makes them sharper still.

    Today’s zooms are indeed sharper than primes from, say 15 years ago (see Roger’s 70-200/2.8L II vs 200/2.8L II comparison from a while back), but that’s not a relevant comparison because we didn’t have 50 MP FF bodies back then. At any given time the highest-res bodies on the market (like the 5Ds R) tend to be well matched with the highest-res lenses on the market, and for the most part those are primes. That’s particularly true if you care about corner performance as you do, as that’s where zooms suffer most.

  • LensNut

    Yup the 9-thumbnail MTF figures at 70/135/200 just like what you have done for Canon 70-200 IS II as part of the addendum

  • Horshack

    If zooms weren’t intended to be used on higher MP bodies then why are lens designers using expensive elements to improve the resolving ability of their lenses, with particular attention to full-field sharpness?

  • Timur Born

    So the optics are good, but in the past I wasn’t convinced by the mechanics and electronics of Tamron lenses. It’s 1000 EUR less than the Nikon, but only 80 EUR less than the Canon. So it seems like a good idea to wait for prices to come down a bit anyway.

  • Patrick Chase

    No, but the high resolution of the high-res body certainly disappears when you put a zoom on it, which was my point.

    If you’re obsessive enough about IQ to accept the downsides of a 5Ds R (which is slower, produces more moire, has less dynamic range, worse high-ISO performance, worse AF, worse video, etc compared to its lower-res counterparts) then the lack “versatility and flexibility” with primes should not be a concern to you.

    Zooms do make sense for many uses for the exactly the reasons you give. High-res bodies make sense if resolution is what you’re after. Sticking a zoom on a high-res body and then complaining about lack of ultimate corner sharpness is just daft, though, because it means that you used the wrong tool for the specific job (or you’re pixel-peeping where it’s not warranted). If you doubt me just look back through Roger’s blog at the many field diagrams and OLAF shots he’s published from “top quality” zooms.

    I use zooms a significant percentage of the time, including the one you’re griping about. I’m happy with their performance, but that’s probably because I don’t ask them to do things that I know they cannot. I have primes for those jobs.

  • One. Chances are that one is going to be excellent for photography with no noticeable defects unless you measurebate. The odds of that one being excellent may vary from 75% to 95% depending on the exact model in question, but odds are any single copy is going to be fine.

    Now if you want one that is pixel peeping perfect at all focal lengths and all focusing distance, with no measurable tilt, etc. it will vary by brand but it will certainly be in the thousands of copies. We have many thousand copies of zooms and I’ve never seen one.

  • Horshack

    The flexibility and versatility of using a zoom doesn’t disappear when it’s mounted onto a high-resolution body, nor does the flexibility and versatility of a prime increase.

  • Patrick Chase

    If you’re looking for maximum possible quality out of a 5Ds R in the corners then IMO you shouldn’t be shooting a zoom at all. Wrong tool for the job and all that…

  • Horshack

    The Canon’s corners sharpen up @ f/5.6 but they’re still visibly softer than the center, at least on the 5DSR.

Follow on Feedly