Equipment

Sony FE Planar T 50mm f/1.4 ZA MTF and Variance Testing

 

OK, you probably know I don’t like ultra-long lens names. I don’t like lenses that are expensive (who does?). So when you throw out a 50mm lens with a long name, like the Sony FE Planar T* 50mm f/1.4 ZA, and price it at $1,500 I’m a bit predisposed not to like it.

Seriously, that’s a lot of money for a 50mm f/1.4 lens. Let’s take a quick price comparison of some common 50mm lenses.

Zeiss Otus 55mf/1.4$3,990
Sony FE Planarf/1.4$1,500
Canon L 50mmf/1.2$1,350
Sigma 50mm Artf/1.4$950
Sony Sonnar 55mmf/1.8$998
NIkon 50mm Gf/1.4$447

I say all this just to point out that going into this test my thinking was it had better be really good or I’m going to mock it. For those of you who don’t like to read the articles, I’m not mocking it anymore. This lens is really, really good!

MTF Tests

This is our usual ‘average of 10 copies, each tested at 4 rotations’ graph. In order to keep the graph sizes reasonably large I’ll do a series of comparisons with the Sony FE 50mm f/1.4 MTF chart on the left, the comparison lens MTF on the right. We’ll start with the one that shut me up, comparing the Sony to the Zeiss 55mm f/1.4 Otus Distagon lens.

Compared to the Zeiss Otus 55m f/1.4

ADDENDUM: The original article contained our original Zeiss MTF curves, done about a year ago. After a polite inquiry and suggestions from Zeiss, we redid the MTF tests for the 55mm Otus and did, as they thought we would, bet better results. Here is the MTF graph showing the difference with new testing methods (on the right), which is, indeed, better than our original methods.

OlafOpticalTesting, 2016

OlafOpticalTesting, 2016

This doesn’t change my conclusions that the Sony has a better MTF in the center than the Otus, but does show that the Otus is better in the middle third of the image. I have left all other text and images unchanged and it continues below.

 

OlafOpticalTesting, 2016

OlafOpticalTesting, 2016

The Sony is phenomenally good in the center; just absolutely superb. From a resolution standpoint it is clearly better than the Otus in the center, and just as good as the Otus away from the center as well. There is a little sag of the Sony’s MTF halfway to the center but then improvement out to the edges.

One note about that ‘halfway sag’ for both of you who actually read the articles and don’t just look at the graphs. This may be an artifact of sensor cover glass. We are limited to a full mm cover glass increments when testing, so for Sony lenses we use 2mm, which is close, but not exactly the same optical thickness as what covers the sensor. (It’s probably about 0.25 thicker). We generally don’t think of a 0.25mm difference as significant for SLRs, but that may not be absolutely true with the short back-focus distance of the FE mount. In other words, the Sony lens may actually be a bit better than what we see here. Or they may not, we aren’t certain yet. 

Compared to the Sony FE 55mm f/1.8

Remember for this comparison the aperture difference is significant. The f/1.4 lens has been tested wide open and that half-stop (Actually 2/3 of a stop. It was late. I was tired.) of aperture makes a lens significantly better.

OlafOpticalTesting, 2016

OlafOpticalTesting, 2016

The Sony FE 55mm f/1.8 ZA Sonnar is a really good lens, but even giving up a half stop of aperture, the 50mm f/1.4 is better in the center and generally nearly as good away from center. This is most impressive.

Compared to the Sigma 50mm f/1.4

We’ve evened up the apertures now, with two f/1.4 lenses, comparing the Sony to the Sigma 50mm f/1.4 A1 lens. The Sigma is significantly less expensive and in my mind the best ‘bang for the buck’ among 50mm lenses.

OlafOpticalTesting, 2016

OlafOpticalTesting, 2016

The Sigma is really good, and I won’t argue if the Sony is worth the financial difference. But from an MTF standpoint, the Sony is better.

Compared to the Nikon 50mm f/1.4 G

This is another unfair comparison, the Nikon 50mm f/1.4 G AF-S costs only a fraction of what the Sony does, but I thought it was a good “what you get for your money” comparison. The Nikon is a very nice, decent quality, usable lens. But the Sony is dramatically better, which at the price it should be.

OlafOpticalTesting, 2016

OlafOpticalTesting, 2016

 

Quick Summary:

The Sony Planar 50mm f/1.4 is expensive and not everyone needs to plop that much money down for a 50mm lens. But, if you do need one, it’s worth the money; it’s really superb. The center sharpness in particular is unheard of in a 50mm lens. This seems to be a pattern we’re seeing with some of the new Sony lenses, too: The fine resolution (at higher frequencies) is higher than we’re used to seeing.

From a pure value standpoint, the price is reasonable. Best quality prime lenses tend to cost well over $1,000 and some up to $2,000, and this is a best quality prime lens. Your shooting may not require a top-of-the-line 50mm lens and there are lots of other options in that focal length for FE shooters. But if you require the best one, then this would be the one you buy, at least based on bench test results. (Remember, I never suggest buying a lens based solely on test results. Go check out pictures, too.)

Optical Field

You may have been, like I was, impressed with how well the Sony maintained a high MTF away from center. The optical field helps show why, and it is also very impressive.  Both the sagittal and tangential fields are almost perfectly flat from one side to the other. For some photographers that will be more important than absolute sharpness. But for those of you who aren’t aware, a field this flat is really an accomplishment. We rarely see it.

OlafOpticalTesting, 2016

OlafOpticalTesting, 2016

Variation

There was a time, early in the history of FE lenses, when I used to wince mentally before I pulled up the Variation graphs for a new lens. That time seems to be past, with excellent results from the more recent Sony lenses. So I was looking forward to seeing how the Planar 50mm f/1.4 lens did.

I have two things to remind you about. First, remember our variation graphs now show only 1 Standard Deviation, rather than 1.5 that we used months ago. Also remember that our bench cuts off about half of the 20mm (edge) readings on Sony FE lenses, so take the extreme edge variation with a grain of salt. I’ve put a 50% gray box over the questionable area to help clarify this.

The first comparison I pulled up was against the Sony 55mm f/1.8 ZA Sonnar, which is one of the better and more popular Sony primes.

OlafOpticalTesting, 2016

OlafOpticalTesting, 2016

The new 50mm f/1.4 ZA Planar clearly has less variation than the 55mm f/1.8 did. It’s really nice and consistent.

To give you a more general comparison, here is the 50mm ZA Planar matched up against the Sigma 50mm f/1.4 Art lens, another really good 50mm lens with very reasonable sample variation. The variation is very similar; both lenses are quite consistent. (As always, a reminder to please not confuse autofocus accuracy with optical consistency. They have nothing in common.)

Sony-Sigmavarsml

 

Overall, the Sony FE 50mm f/1.4 Planar has very reasonable sample variation, about what we see in most other high-quality 50mm prime lenses from other manufacturers.

Summary

MTF tests like this give you a good idea of how sharp the lens might be and how much sample variation you should expect. On the basis of these, the Sony 50mm f/1.4 Planar is as good as anything available. For those of you who need a really high-quality 50mm lens on an FE mount camera, it looks like your best choice.

The very flat field with almost no curvature is another plus, and one that may appeal to photographers every bit as much as the excellent MTF does.

But of course the proof is in the pictures. We’re starting to see some sample images online and that will explode shortly, giving you a chance to evaluate the bokeh and the lens’ performance in various lighting conditions. But unless those greatly surprise me, I think a lot of people are going to love this lens.

Roger Cicala, Aaron Closz, and Markus Rothacker

Lensrentals.com

July, 2016

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
  • S.Yu

    Great, thank you!

  • Very good points Daniel, and that is exactly how it looks.

  • Edna, that is a very good point. With the vast majority of lenses that’s exactly how our copies are obtained: we buy everything from US retailers and there’s no way the manufacturer could pick our copies. In this particular case, though, you are correct and I should have pointed that out: these are Sony’s own copies, not mine, we just had the opportunity to test them. This is the first time we’ve done that but I was excited about getting into them. It will also be the case with the 70-200 f/2.8 GM.

    On the other hand, and I don’t mean this to sound snarky but it’s true: Sony doesn’t have the capability to test to our standards or even close to them. They’re as interested in our results as anyone else. That’s not saying they might not have cherry picked to some degree.

    Roger

  • S. Yu. I understand and that’s a pertinent point. In the fall, when things slow down for us a bit, I’ll try to do a short series with the common lenses at both 1 and 2mm to at least give some ideas. I’m still getting used to the increased sensitivity to stack thickness that the short backfocus distance of the mirrorless mount gives. We did our initial work with SLR lenses and things weren’t as sensitive.

  • Daniel

    Looking closely, the Otus seems to be a clever tradeoff: The mtf curve rises (almost) until the 12mm mark (the short edge of the sensor, after that the covered area stops increasing with r^2 due to the cuts in the circle). It looks like a carefully designed optimum of sharpness per area.

  • Daniel

    I’m sure you’re aware of it, but you didn’t mention, that the mtf chart is misleading, if you forget to think about the sensor area covered by a certain region on the chart. The values between 0 and 2 cover an area on the sensor of about 12,6 mm^2, the values between 8 and 10 cover an area of about 113 mm^2.That’s probably why Zeiss sacrified some center performance, I guess the Otus is still sharper by a good margin if you consider the area.

    Thanks by the way for your excellent and entertaining blog!

  • S.Yu

    That’s why it would be best for reviewers to test using samples acquired at random retailers, preferably offline, it would be much harder to rig an entire area’s inventory than a single order 🙂

  • S.Yu

    lol ok, I just wanted the mean value of the few tested lenses at the data points at the dip actually.

  • Thinkinginpictures

    I had the 50mm 1.4 ZA in A mount. It was a pleasure, but wide open performance was lackluster. Excuse me…it actually sucked. Outside of that however, the colors, contrast and bokeh rendering where top notch. It too was 1500.00 fat ones. Since selling it off at a massive loss, I’m not sure I’d give Sony another go at that price point. So many releases these days with mirror less…best to wait it out a bit and do what I call “The 30 percent rule.” Buy it used when you can grab it for 70% of retail. With Sony, it’s the most practical thing to do.

  • Edna Bambrick

    RC, Any thoughts around the idea that production quality and variance are highest in the first production run or later as times goes on? Differences between manufacturers?

    As a side note, I shoot the 7mm REM MAG Sendero and the rumor was at the time of the release that the first run was all supervised and QC by the Custom Shop. I believe it. I shoot 1/4 to 3/8″ MOA groups with my own hand loads.

  • Edna Bambrick

    You got that all wrong. It’s the dancing girls that work on us. : )

  • Edna Bambrick

    Roger, I appreciate the response but the DPR article brings in more doubt to the overall story here. What I ventured over there as a possibility (and a credit to LR) that it’s possible that lenses sent to LR are cherry picked knowing your advanced testing capabilities and the popularity of this blog. If I were the manufacturer, I’d go the extra step to ensure that the supply chain for LR is fed with the best examples. (a good reason that you will be my retailer going forward)

    If DPR gets their lenses the same way as most consumers do then their results more closely resemble what the average photographer may get.

    Although, I understand DPR was provided the lens as part of a media event. So one would expect that lens to be an exceptional copy.

    Since getting OLAF have you taken any of these exceptional lenses and tested then with Imatest or other testing to confirm their performance inline and correlated to OLAF results?

  • Steve,

    It’s a really good point, but remember how lenses are made: the assembly line sets up, makes a run of a given lens, then tears down and sets up for a different lens. It may be months (and in some cases we know it’s over a year) before that lens is made again. With the Sony lenses I would guess months, but at any case, there won’t be another cycle for quite some time, so we have to do the best we can.

    While it is a limited finding and not done with the Sony brand yet, we have, with some other lenses, done repeat runs many months apart. We haven’t noticed any real difference in general. On the other hand we do know that sometimes things have been changed in different runs so I hesitate to generalize too much. For example, Canon 85 f/1.8s have different electronics in newer runs, although we haven’t seen optical differences. There have been similar minor changes in other lenses.

  • Steve Waldstein

    Roger, Aaron, and Marks thanks for the work. I do have a question about your sample variation testing. With a new lens in limited supply how do you ensure your not testing lenses that came from the same manufacturing cycle that will tend to be fairly close to each other. In my industry you need to take 3 samples form 3 to 5 independent manufacturing lots to validate process variation. Anything manufactured on the same day is disqualified.
    I just don’t see how you get a good representation of true variation on a new lens. Maybe after 6 months of buying samples to use but the lens just became available.

  • S. Yu, I apologize, I didn’t realize you wanted statistical analysis of significance. We didn’t do a multiple copy anaysis of the sensor stack difference – we’ve done it on some lenses in the past as we fleshed out technique but we don’t have the extra 10 hours to do that on every lens. We simply take a single lens, check that it’s a well centered copy, then measure it with and without glass in place to determine if it is better with glass. Then proceed with our testing.

    Roger

  • Leslie Feigin

    Would I see a difference between these lenses in a 20″ x 30″ print hanging in my hallway from two feet away photographing a scenic at about f/8 ?

  • S.Yu

    …and significant means p<0.05? So the mean difference would be?

  • It’s significant off axis.

  • Edna, I can’t tell you that because we don’t know Zeiss’ testing metrics. Do they use white light? Photopic light? Green light? How many points do they measure? More importantly are their MTFs the average of numerous copies? We have some Otus lenses that are about as good as what the Zeiss graphs show. It’s not all of them, though, and when we average multiple copies this is what we get. Or if I test with monochromatic light, the MTFs are better. Even using a different reticle (pinhole versus cross versus separate lines).

    You bring up a good point though: what is spec? We don’t know. No manufacturer releases what their acceptable spec are. (I know a few but I’m under nondisclosure about them, and I don’t know Zeiss’ at any rate.)

  • S.Yu

    How much worse is 1mm?

  • Edna Bambrick

    How are the Otus results so far off of the Zeiss advertised measurements and consistently so? If an Otus lens performed as badly as OLAF is indicating then those lenses would be dramatically out of spec.

  • Thank you Carleton. I’m working on dancing girls.

  • Carleton Foxx

    I hope both your customers AND your investors realize what an amazing value you provide with this kind of reportage. The only way you could improve upon it would be to mount OLAF in a Sprinter van and take your show on the road. Dancing girls and the Marine Corp Band optional.

  • 1mm is worse and 3mm is worse, but that doesn’t show us that 2.3mm wouldn’t be better than 2mm, I’m afraid.

  • Joel, we can’t directly measure transmission and vignetting, our bench has a work around but it’s time consuming as can be.

  • Paul Gero

    Thanks Roger for your test. I’ve been fortunate to have one to test and shoot and the look is so nice….thanks for reaffirming what I suspected — it’s a FINE piece of glass.

  • S.Yu

    Any chance on testing a few copies on 1mm sensor glass? If we see a much worse dip then that pretty much explains it.

  • Joel Benford

    Do you happen to measure transmission and vignetting, Roger?

    Some of us available darkness dwellers are interested in 50/1.4s more for the light gathering more than the shallow DoF. 🙂

  • Lee

    I have some questions for Zeiss.

  • disqus_FQr0oPsUDS

    Thanks for the super-quick review of this new Sony Lens. I can imagine this would also be a great ’75mm’ portrait lens (via Sony’s APS-C mode on their A7 Cameras)… Can’t wait to rent it!!

Follow on Feedly