Equipment

Standard-range micro 4/3 Imatest Results

Published May 18, 2012

 

This is the second in a series of three posts about Imatest results on m4/3 mount lenses. The first post covered most of the lenses of 25mm and wider focal length. Most of the zoom lenses in that test were done at, or near, their widest focal length since we were concentrating on the wider angle lenses. This test run started by retesting most of the zoom lenses at 25mm, which we hoped would show a bit better performance since most zooms are weakest at their extremes.

We also did some confirmation testing at 25mm. First retesting the Panasonic – Leica 25mm f/1.4 because it just didn’t seem as great as we expected on the initial set of tests. Then we did some comparison testing, testing the same lenses on both the Panasonic GX1 and the Olympus OM-D E-M5. I won’t bore you with all the numbers, but testing 6 copies of the Panasonic-Leica f/1.4 didn’t change our initial results much. (The results have been updated on the previous article.)

I had assumed, since we test only unsharpened raw files, that camera brand would not make any difference, and it did not. We tested copies of the Panasonic 14mm f/2.5, the Olympus 9-18mm zoom, the Panasonic 7-14 f/4 zoom, and the Olympus 12-50 zoom on both cameras and there was no difference in MTF 50. This doesn’t mean there is not some distortion correction or other manipulation going on when using a same-brand lens on a given camera. It just means there’s no obvious manipulation of the raw image that affects MTF 50.

Testing the Zooms at 25mm

In the last article, we tested the Olympus 12-50 and 14-42mm f/3.5-5.6 II R lenses at 14mm.  I thought that focal length might be weak points for those lenses, so today we retested them at 25mm, more towards the center of their zoom range. We also tested the Olympus 12-50 f3.5-6.3 (we were out of stock during the wide angle test), Olympus 14-150 f/4-5.6, and Panasonic 14-140 f/4-5.8 OIS all at 25mm. Two or three copies of each lens were tested, and testing was done as described for the other article.

I’ll present the numbers in two tables:

In the first table, we’ll look at the MTF50 of each lens at the widest aperture it can use at 25mm

Oly 12-50 f/5 800 665
Oly 14-42 II R f/4.3 690 580
Pan 14-45 f/5 830 670
Oly 14-150 f/4.7 780 640
Pan 14-140 f/4.7 795 680

In the second table, we’ll look at the MTF50 with each lens at f/5.6

Oly 12-50 f/5.6 830 670
Oly 14-42 II R f/5.6 710 630
Pan 14-45 f/5.6 825 675
Oly 14-150 f/5.6 785 660
Pan 14-140 f/5.6 800 690

 

There are few comments to make about the zooms. Using the rule of thumb that a difference of less than 50 lp/ih is probably not noticeable, only the Olympus 14-42 lags a bit behind the others in resolution. The two superzooms, the Oly 14-150 and Panasonic 14-140 were both pleasant surprises. It’s quite a good showing they made, keeping up with the smaller-range zooms. Have to give some props to the Panasonic 14-45, too. It’s a low cost ($275) lens but hangs right there with the others. Finally, the zooms haven’t changed much between their wide end in the previous tests and the middle of their range we tested now.

Testing the 45mm Primes

We have two 45mm lens choices for m4/3: the Panasonic-Leica 45mm f/2.8 and the Olympus 45mm f/1.8. There are a lot of differences in the two. The Olympus gives a wider aperture, is smaller, and $250 cheaper. The Panasonic-Leica has built in Image Stabilization. I’ve presented this table a little differently so you can compare the two lenses at equal aperture.

Pan-Leica Olympus
f/1.8 720 615
f/2.8 1000 875 900 790
f/4 1050 885 925 800
f/5.6 1015 870 900 770

 

In our tests the Panasonic-Leica 45mm does outresolve the Olympus by what is probably a noticeable margin if you compared them side-by-side. Results from both are excellent, though, and I can’t imagine anyone being disappointed with either one.

Oh, and One More Thing

I was just about to break down the testing set up when the receiving people brought me a present: several copies of the Sigma 30mm f/2.8 and 19mm f/2.8 in m4/3 mount. These were spectacular on the NEX system so I had to go ahead and run the tests on these before they went out on rental.

Sigma 19
f/2.8 850 745
f/4 920 770
f/5.6 925 715
Sigma 30mm
f/2.8 825 690
f/4 910 740
f/5.6 895 730

Several things are interesting here. First, the Sigma lenses give excellent resolution on m4/3 mount like they do in NEX mount. No surprise there. Second I would have expected the resolution to be lower on m4/3 mount since we are measuring in line pairs / image height and the NEX-7 has 4000 pixels of image height versus the m4/3 3460 pixels and Imatest measure system (camera and lens) resolution, not just the lens.

As expected, the Sigma 30mm has slightly lower resolution on m4/3 cameras. However, the Sigma 19mm was had nearly the same resolution in m4/3 mount as it did in NEX mount. In m4/3 mount, therefore, the Sigma 19mm and 30mm were had about equal MTF50 (the 19 actually a tiny bit sharper in the edges and corners at f/2.8). (In NEX mount the 30mm was slightly sharper than the 19mm.)

I’m not sure why the difference, so what follows is purely speculation and I welcome other people’s ideas. In looking at our m4/3 results overall, it seems there is a tendency for wider focal lengths to resolve at least as well as longer focal lengths. In most systems we’ve tested that hasn’t held true: wider lenses, with a few notable exceptions, tend to be a bit weaker. Perhaps the m4/3 sensor format makes designing wider lenses easier or more effective. Perhaps the m4/3 sensors, particularly the newest 16-megapixel sensors have a better or more effective microlens array? Perhaps there is some sharpening or contrast boost going on in-camera to the raw images? Note: Please read Thom Hogan’s excellent comment below: it makes more sense than any of my ideas did. 

It’s also possible there is some defect in our testing set up that’s giving an advantage to wider focal lengths on m4/3 cameras. I don’t think this is likely, we’ve run 5,000 lenses through our Imatest lab and I think we’ve worked the kinks out. But it’s always possible. I’ll look forward to seeing the results from other testing labs to see how theirs compare.

Part 2 Summary

Having been through most of the m4/3 lenses now I think there are some very obvious conclusions: 1) there are a lot of good prime lenses available for m4/3 from 14mm to 45mm. It’s interesting that the new Sigma lenses stand out as superb compared to other NEX lenses, but are just another good prime lens on the m/4/3 system.

Things are a bit different on the zoom side. There are some good consumer quality zooms, but other than the Panasonic 7-14 f/4, there are not any that I would consider ‘top’ or ‘pro’ quality zooms. Several of the zooms have good resolution, but they are all aperture-impaired. I think this is a big hole that needs to be filled, especially now that we’re getting reasonably fast autofocus in mirrorless bodies. An f/2.8 range 35-100mm would kick some serious butt, even at a premium price. Even a reasonably long wide-aperture prime, something like an 80mm or 100mm f/2.8 would be welcome.

The biggest gap in the m4/3 lens lineup, though, is certainly going to be at the long end and we’ll cover that in the third part of this series. I’d love to find an exception, but having spent a weekend shooting every one of the m4/3 mount long zooms and comparing them to what’s available for SLRs, I don’t expect a lot of good results.

Roger Cicala

Lensrentals.com

May 2012

Addendum: Wouldn’t you know, the day after I wrote my little diatribe about needing good quality f/2.8 zooms for m4/3 the first preliminary reviews of the Panasonic 12-35 f/2.8 OIS come out and are very positive. Now, just give me that good telephoto, guys, and I can stop juggling mirrorless systems 🙂

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 Equipment
  • Sam Atherton

    I know this old but isn’t there some correction that goes on in camera that could of also had an impact on this? Do the Panasonic lenses resolve better on native bodies? And then on the flip side Olympus lenses resolve worse on a Panasonic? Do these numbers still hold true today?

  • There seems to be a huge difference in the performance of both Sigma lens when you read different reviews. I purchased both lens for my EP3. The sharpness of the 19mm was inferior to the Oly 14 42 Kit lens. I tried unsuccessfully to sell it for $50 and ended up tossing it into the bin.

    The 30mm was better but I couldn’t understand why it would take wishy-washy shots one day and very sharp the next. I discovered that at F2.8 and F4 the shots were poor, but shut it down to F7 and greater and shots were very sharp, but what is the point of having a F2.8 lens and only using it at F7? – may as well use the kit lens – which I do.

    Disappointed in both Sigma lens – wasted $200. The Panny 20mm is far superior to both these lens.

  • Roger Cicala

    Steve,

    We shelved the telephoto part for two reasons. First Imatest isn’t giving us accurate numbers over 400mm full-frame equivalent. Second I was so unhappy with everything over 150mm in m4/3 format I sort of lost interest. I use the m4/3 cameras as SLR alternatives but when I did side-by-side comparisons there was nothing even equal to the Canon 100-400. In fact I got more detail with 5D3 at 400mm than with OM-D at 300mm (600mm equivalent).

    So I decided to wait until better options come up.

  • Steve Luck

    Hi Roger,

    Any update on part 3? Really looking forward to it.

    Thanks for the great information,
    Steve

  • Paul R

    Mr. Hogan, I understand the argument you’re making about the aggressiveness of lens designs for different formats. People have been saying the same thing for years. But I think the situation is quite a bit more complex than this.

    Companies like Schneider and Rodenstock are extremely competetive with the performance of their large format optics (or at least they were, back when the market was bigger). Lf lens designers don’t slack just because the negative is big. In fact, they are forced to work with the more difficult constraints posed by a large image circle. The image circle of many large format lenses could swallow an entire dslr … this coverage comes at a price, especially when LF photographers demand good performance all the way out to the edges in order to have movements.

    It’s true that for bigger formats the designers tend to emphasize MTF at lower frequencies than for smaller formats, but this is really about tweaking the shape of the MTF curve rather than just giving up on resolution.

    Another place to look is at the lenses for medium format technical cameras (also made by Schneider and Rodenstock), which, as far as I can tell, outperform lenses made for any other format, ever. It’s difficult to know for sure, since the MTF curves published by Canon and Nikon seem hypothetical at best, but the photographic results I’ve seen from these lenses make my Nikkor optics—which are awesome—look sad in comparison. I suspect this is possible because the constraint of price point has been lifted for these lens’s designers. The MF technical camera crowd expects to pay $4K to $6K for a lens, because that’s how they roll.

  • Roger Cicala

    Hi Nick,

    I was so disappointed by the Imatest results on the m4/3 telephotos that I decided I needed to do some real world shooting to confirm the results (Imatest is testing them at 20 – 40 feet which makes me uncomfortable – they may perform differently at longer distances). The IQ isn’t really very good on any of the long m4/3 glass when tested in real life either but when I started writing it up I was told our optical bench, which tests at infinity, will be here in a week or two, so I decided to wait for that so I could get numeric results at infinity. I’m afraid it’s going to be August before I get it all done.

    Roger

  • As someone who just jumped on the OM-D bandwagon, and has been loving it + 45/1.8, I have been finding these posts very interesting. Any ETA on part 3? I would love to know which is the best of the bad tele lenses, as I need to pick up a tele before I head on a trip to europe.

  • Jim Thomson

    Roger,
    The Sigma’s aren’t just good prime lenses, they are also significantly less expensive than the OEM primes.

    I hope Sigma follows up with some more lenses for the format.

    Maybe they could make some of your long lenses as neither Panasonic nor Olympus has shown much interest in the long end.

    Regards
    Jim

  • Steven Runyan

    would also like to see a test of current production x 14-42 after having experienced the 1/60 to 1/160 issue (which my resolution test at slower shutter speed didn’t show up)

  • Roger Cicala

    Those are good points, Sam. I used the Olympus as our base testing camera originally planning to compare Panasonic and Olympus bodies. There was so little difference in the ones we tested that I didn’t explore further. I had assumed there wouldn’t be a difference in 4:3 ratio and that only the GH2 could hold it’s own (given the oversize native sensor) in the other ratios.

    I’ll try to do that comparison after the holiday rush is over, though. Assumptions are never a good thing 🙂

  • adventsam

    Roger, the highest spec. Panasonic is the GH2, is there a reason why you use the GX1 instead of the GH2? Also, it would be very interesting to see the GH2 3:2 ratio tested as opposed to 4:3 ratio. On the GH2 this is a native 15mp sensor, I use it a lot along with the 16:9 ratio(14mp) these are unique to the GH sensor’s but in my testing the 15mp 3:2 raw out resolved the Nex5n for instance by some, but I only had the nex5n 18-55 kit lens c/w the 14-140 of the Panasonic. Even so, I would be interested in your findings, the GH2 sensor technically is the best sensor Panasonic make for stills camera’s not the G3/GX1 though.

  • In general, as the imaging circle and capture area go down in size, lens designs must actually be better to perform well. I suspect what you’re seeing is a lot of compact camera lens design experience moving upwards.

    Nothing’s different than it was with film, actually. In the film world, the MF makers were a bit more lax on lens design than the 35mm makers had to be. The same thing is happening in digital: the larger formats are not having to be as aggressive about lens design as the smaller formats. Of course, with 24mp DX and 36mp FX sensors now with more pixels in the future, that’ll change a bit in the future, as demand for better lenses will happen from the user base.

    But it’s interesting to watch. Both in sensors and in lenses the big engineering tasks are happening in the smallest sizes. This is partly a quantity game. We’ve got individual smartphones selling in the 10’s of millions, compact cameras selling in the millions, mirrorless selling in the hundreds of thousands, and so on (annual basis). The low-end crop sensor DSLRs are the only ones that stick above that declining demand curve, and you’ll note that those kit lenses for them are pretty darned good. In short, the optics makers are putting their money where the demand is. Since the bigger demand is low end, that’s where all the high end design action is.

    For the two m4/3 players, they don’t do bigger stuff, either, which means all their optical action is in smaller imaging circle captures. It’s the companies like Canon, Nikon, Sony, and maybe Pentax that are spread thin in lens designs (from compact cameras to full frame or bigger). Sigma made an interesting choice: they basically designed the same two lenses to cover three imaging circles. In other words, they built a larger demand by consolidating designs. Sony’s choice of APS for NEX may have been determined by this, too. Nikon, on the other hand, now has four basic imaging circle sizes they have to design for. They’re spread thin and it shows.

  • rrr_hhh

    Thanks for sharing those results with us. The fact that you can test several samples of the same lenses is a real plus. In other tests I’m always wearing about sample variations. It would be nice if your ables included the number of samples you ave been able to test.

  • Jessy Plames

    Thanks for the tests. I was hoping you had the Panasonic x vario 14-42mm to test. From all accounts it is supposed to be as good as the 14-45mm.

  • MikeH

    You might want to add that the Panasonic 45 is a macro lens. So it gives you some options that the Olympus 45 cannot. The Olympus 45 is speedy and brings in more light so it has a few tricks up its sleeve too. 🙂 Both excellent lenses for m43.

  • These articles are fantastic! Super helpful comparisons.

Follow on Feedly