New Items

Zeiss Milvus Optical Bench Tests

Published November 9, 2015

When Zeiss announced the new Milvus lens lineup, there was a mixture of excitement and yawns. In part, I blame Zeiss’ marketing department for the yawns. OK, I shouldn’t be harsh. I shouldn’t, but I will. I’m certain a lot of people are perfectly comfortable just knowing that the new lenses will ‘redefine the limits of my creativity’ and provide me with ‘freedom to use the focus position as an artistic tool’ while having great ‘haptic appeal’., 2015

There’s no question the new housings are much nicer than the old ones. That certainly makes a difference in how the lenses handle. It might also make a difference in how much copy-to-copy variation there is since newer housings may allow more accurate assembly or greater ability to optically adjust the lenses. New optical coatings can’t hurt, and Zeiss has always done great things with coatings.

Me, I’m a geek, so I want to know things like are these really different than the previous lenses? We know that the 50mm f/1.4 and 85mm f/1.4 Milvus are completely new optical designs, replacing two of the lenses in the ZE/ZF lineup that really were due for an update. The other lenses in the Milvus lineup have new coatings and are in completely new housings. It doesn’t appear they have been changed optically, although I’m not absolutely certain they haven’t been modified.

We don’t have any method of quantifying the lens’ haptic appeal for you, and can’t provide you a numeric example of how much the limits of your creativity will be redefined, but we sure can put some of these new lenses up on the optical bench and see if there are optical differences we can detect. We expected to see big changes for the Milvus 50mm f/1.4 and 85mm f/1.4 lenses. We weren’t expecting to see optical changes in the other Milvus lenses when compared to their ZF/ZE predecessors but thought we’d double check by testing the Milvus 100mm f/2.0 Makro.

Testing Methods

By now I think that most of you who read these articles know the drill: we tested 10 copies of each lens, each copy tested at 4 different rotations. Before starting the testing, we did trials with and without optical glass in the path of the lenses. The only one of the lenses that showed any difference with glass was the 85mm f/1.4. It was slightly better with 2mm of optical glass in the pathway and was tested that way. The others were tested with no glass since it made no difference in their MTF and it’s simpler to test without glass.

MTF Results

Milvus 50mm f/1.4

The new optical design of the Milvus 50mm f/1.4 is readily apparent when comparing it to the ZE/ZF 50mm f/1.4. The Milvus has a slightly better resolution in the center and maintains that resolution further out into the image circle than the older design.



Just because I know someone’s going to want to call the Milvus ‘just as good as the 55mm Otus’ or something else silly, I’ll post that comparison, too. The Milvus is a clear improvement on the older 50mm design, but it’s just as clearly not as good optically as the Otus, either.


Roger Cicala and Aaron Closz, OlafOpticalTesting, 2015


Milvus 85mm f/1.4

The new design for the 85mm lens shows a similar improvement. At the higher frequencies, the improvement is even more dramatic than it was for the 50mm lens.

Roger Cicala and Aaron Closz, OlafOpticalTesting, 2015


Milvus 100mm f/2

If I understood all of Zeiss’ release verbiage correctly, we would not expect any major changes in the MTF chart of the 100mm lens, and indeed, it is basically identical to the ZE/ZF MTF. The MTF graphs aren’t quite identical, but that’s nothing other than the slight variation we’d see if we tested another 10 copies of the ZE or Milvus and compared them to each other.

Roger Cicala and Aaron Closz, OlafOpticalTesting, 2015


Sample Variation

With MTF, we had expected to see some differences in the 50mm and 85mm lenses, but not in the 100mm, and that was indeed the case. With sample variation, we really didn’t have any expectations. The original ZE/ZF lenses were all quite good as far as having copy-to-copy consistency, but these are entirely new housings and that could make a difference in variation.

Roger Cicala and Aaron Closz, OlafOpticalTesting, 2015


Roger Cicala and Aaron Closz, OlafOpticalTesting, 2015


Roger Cicala and Aaron Closz, OlafOpticalTesting, 2015


Remember the absolute numbers we use for the Consistency score are a very blunt tool and only evaluates the 30 lp/mm (green) variation. I encourage you to look at the actual graphs rather than just spout the number. But as far as the Consistency number goes, the Milvus lenses do seem to have slightly less variation (higher consistency number) than the ZE/ZF lenses. I can’t say if this is because the optical design is more tolerant, the optomechanical housing more consistent, or a combination of both factors.

Particularly with the 100mm and 50mm lenses, the graphs seem to show less center variation, which indicates center sharpness is very consistent from copy to copy. In both cases, the differences when looking at the graphs seem even more impressive than the numeric differences. The 85mm has the biggest improvement in Consistency number, but I think this sort of demonstrates the shortcomings of the ‘one number’ summary measurement. If you look at the graph, it is much more consistent at the lower MTF frequencies, but doesn’t seem much better at higher frequencies. Still, it is clearly improved over the ZF/ZE versions.


The new Milvus lenses are going to appeal to a lot of people because of their more elegant housing. I poked a little fun at the term ‘haptic appeal’ but I think it’s a very real thing and a lot of people are going to love these lenses for that. But that’s not something I can measure; you’ll learn a lot more about that from actual lens reviews and handling them yourself. Although I will say even simply setting them up for optical bench testing I love the feel of the focus throw. It’s remarkably smooth and precise.

Optically, there’s no question the new 50mm f/1.4 and 85mm f/1.4 lenses are improved over their predecessors. We’ll get around to testing the other Milvus lenses one of these days, but I expect they’re going to be pretty much identical to their ZE/ZF counterparts, just like the 100mm f/2.0 was in these tests. From an optical standpoint only, then, there’s no real reason to upgrade, Although as I said, there may be some very good non-optical reasons to do so.


Roger Cicala and Aaron Closz

November, 2015


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 New Items
  • Max Manzan

    “The Milvus 25mm seems to outperform the Otus 28mm across the board”

    I don’t know where you have this, bit it’s very clearly not the case. The Otus 28 is optically superior in nearly every aspect of its performance, the only exception being the sharpness in the frame centre at wide open for far objects, where the Milvus is a bit better while the Otus shows an incredbly uniform sharpness over the whole picture right into the corners, which is unparalleled for an f/1.4 wide angle lens.
    This is a far cry from “…outperform the Otus 28mm across the board”.

  • Arthur Meursault

    Being a Zeiss fan has been a little frustrating lately. The Milvus 25mm seems to outperform the Otus 28mm across the board and for half the price. Being an early adopter of all of the Otus lenses, it is frustrating that Zeiss continues to produce lenses for half the cost and with the same cosmetic design that sometimes beat their ‘top of the line’ lenses. The lesson Zeiss fans are learning, is it’s better to wait than to buy.

  • William Crane

    That the Milvus 50/1.4 is better than the old Planar 50/1.4 wide open is no surprise. How do they compare at f5.6? Thanks.

  • Any chance of running the new Tamron f/1.8 VC lenses through OLAF?

  • Tuco

    I buy 1.4 to use the lens open (most lenses are great at f4 – f8). And, focus is critical open the much, so how do you do it? What use are these lenses put to?

    I’m curious as these seem all to be “tripod lenses” (or video, perhaps) at modest focal lengths, best used for product and food (and other staged macro-like use). Maybe landscapes, that need 1.4? :/

    I’ve used a pretty complete Hasselblad kit for many years and appreciate the rendering and quality of Zeiss’ great coatings/glass, but I can’t figure out where I’d need *fast* manual focus lenses anymore, other than studio work.

    I think I’d buy into Zeiss again, if they could incorporate with a good AF system. It’s kind of the elephant in the room.

  • Roger Cicala

    Probably Lee, but we haven’t done field plots yet. It could also be that it has less peripheral aberration, although I’d bet field.

  • Lee Saxon

    Since the MTF falls off more slowly I assume the 50 Distagon has a flatter field than the 50 Planar?

  • Roger Cicala

    Alan, we did for all of them.

  • Allan Sheppard

    Hi Roger and Aaron,
    I have the Zeiss 50mm F2 macro and find that I have to be careful handling it as the surface is so smooth and I don’t want to drop it. While the focus knurling can be useful when installing on the camera you tend to hold the lens close to the mount.

    Do the new Milvus models have the same (extra) smooth surface?

  • Alan

    Roger, another excellent series. I think Zeiss is dialing in the manufacturing process with these Milvus. IMO, it makes sense that higher frequency would see less improvement since they’re more sensitive.

    One question: Did you try 1mm and 3mm for these tests (as well for the Nikkor test)?

  • Dibyendu Majumdar

    Hi – The purpose of the sample variation tests I assume is to look at the variation as products roll out from the manufacturer, therefore the sample variations tests that you post only make sense if you test new samples only. Clearly two tests where one set of samples is new and the other isn’t is really not comparing like for like and is misleading.


  • Roger Cicala

    Lynn, newly arrived. We’ve done repeat test with many lenses over the years and there isn’t a lot of change over time — but remember we test them after each rental, do optical adjustments when needed, have our own service department and use factory service frequently. If we just rented them out for two years with no maintenance and adjustment, then I’m sure there would be deterioration.

  • Roger Cicala

    Arun, the manufacturer claims under 1% but we think it’s around 1-2% as we operate it.

  • Arun

    Hi, do you know what variation there is in the measurement system (Olaf and operator)? This will help quantify what portion of your copy to copy variation comes from Olaf as opposed to the lenses themselves.

  • > we tested 10 copies of each lens …

    I’m curious is that is 10 copies when newly arrived, or after they have been rented out several times.

    (I’m guess new, as I believe the Milvus line was recently introduced.)

    It might be interesting to see how much, if any, the IQ of lenses deteriorate over time. Rental lenses might tend to receive rougher treatment than those owned. Also, it might indicate how well LensRentals repairs are doing.

  • bdbender4

    Language wrangling redux: my beloved Fuji Touits are happily haptic, perhaps matching Milvi? On the little-known anomalous partial dispersivity index, though, the Touit 12mm rates a three-element score to the Milvus 100mm’s two. The 32mm Touit is, alas, a plain zero planar lacking anomalousness.

    Marketspeak – bah humbug. Nice lenses though.

Follow on Feedly