LensRentals.com

Finally Got Around Touit

Posted by

I was lucky enough to be invited to New York for Zeiss’ Touit lens release and they were kind enough to loan me a 32mm f/1.8 lens in Sony E mount to experiment with. Unfortunately being away during our busiest repair season got me behind and I’ve only just now had a chance to experiment with the lenses a bit.

This is more speculative than what I usually write. Generally I wait until we have multiple copies to run Imatest, look at sample variation, etc. In this case, though, it became apparent that when we do get full stock in, it’s all going to go right back out on rental since there’s already a waiting list. So I thought it would be worth investigating the single Zeiss 32mm f/1.8 copy we had.

Background

I don’t want to repeat what everyone already knows, but there are a couple of points I think need emphasis:

1) The Touit lenses are going to be slightly different in Sony and Fuji mounts. The Sony’s have electronic apertures, the Fuji’s mechanical. The optics may vary a bit, too. Differences in flange-to-sensor distance and Fuji’s raw manipulation account for that. I mention it because you should look at reviews of the lens in the mount you are interested in. This is about the Sony E mount because, well, that’s what they loaned me.

2) These are autofocus lenses. The AF system has a DC motor, not an ultrasonic motor. It works well, but it makes a bit of noise which would probably make it a poor choice for video. It’s not a horrid noise by any means, lower pitched and much less buzzy than DC motors in other lenses I’ve used, but it’s noticeable when things are quiet.

3) The manual focus ring is fly-by-wire. Turning the MF ring actuates the DC motor, not a mechanical linkage. It’s one of the better implementations I’ve seen and I had no trouble focusing accurately (much better than, say, the Canon 85mm f/1.2). But again, video people pulling focus will have the electric motor making noise in the background.

4) These two lenses will be followed by a 50mm macro lens in the fall. There will be some other lenses in 2014 and at least consideration is being given to a zoom lens.

5) There are no immediate plans for Micro 4/3 mount Touit lenses.

 

The Sony 35mm f/1.8 OSS (left) and Zeiss 32mm f/1.8 Touit . 

And without hoods.

A Few Subjective Observations

First, I always try to say what my point of view is when I start a post, because I know it influences my opinions to some degree. I’ve made no secret I consider the NEX system to be one of great cameras with generally poor lenses.

Before shooting with them I was much more excited about the 12mm Touit than the 32mm simply because there was not a good wide-angle prime at any price for the NEX. After shooting with them I found the 12mm had fairly blurry corners wide open on the NEX-7. It may be better on the NEX-6 but I didn’t shoot it on that camera.

For this test I looked at the Touit 32mm f/1.8, partly because I liked it better, partly because the Sony 35mm f/1.8 OSS provides a good comparison point. I consider the Sony 35mm to be a good, but not great lens (although it’s one of the better NEX lenses). It’s reasonably sharp wide open, becoming quite sharp around f/2.8. I like the OSS and price is excellent.

The Sigma 30mm f/2.8 E-mount lens provides a third option in this approximate focal length. It doesn’t have as wide an aperture, but it is does have an amazingly low price tag.

Just shooting around with the 32mm Touit I was impressed, particularly by center sharpness and particularly stopped down just a bit. I also loved the close focusing distance (1.2 feet minimum) that let’s it almost be used as a pseudomacro with a 1:9 maximum magnification. Wide open, I wasn’t blown away by it, but stopped down it was awesome.

 

Close up the 32mm Touit is wonderfully sharp
Even viewed as a 100% pixel-peeping crop

 

Imatest Results

I want to point out that these results are from a single copy of the 32mm f/1.8 Touit, not my usual multiple copies, because we don’t have stock yet, just the one loaner copy Zeiss was kind enough to provide us.  I see nothing that makes me think this copy is out of sorts, but until we test multiple copies we don’t know that this one is representative.

I compared it to a group of 5 copies of the Sony 35mm f/1.8 OSS lens (since we hadn’t tested that one before I thought it best to go ahead and get reliable results now). The Sony 35 had an interesting sample variation: all copies had a slightly soft corner. Just one corner. Not the same corner. But always one. It wasn’t awful by any means, but MTF values in 1 corner were always 20% lower than the other 3 corners.

We’ll start with the MTF50 graphs for the Sony 35mm f/2.8 OSS, which as I mentioned, is a pretty good lens.

The Sony starts out pretty well for a wide aperture lens, with center resolution of 790 Line Pairs / Image Height in the center and 595 average at f/1.8 on the NEX-7. At f/2.8 the lens sharpens dramatically with MTF50 of 1030 in the center and 805 average. This is slightly better than the Sigma 30mm f/2.8 which has an MTF50 of 985 center and 805 average at f/2.8, increasing to 1040 / 840 at f/4.

Imatest results for the Zeiss Touit 32mm were even better.

At f/1.8 the results were similar to the Sony 35mm, with MTF50 of 800 in the center and 630 average. The Touit sharpens up more quickly, though, having MTF50 results at f/2.2 (1190 center, 830 average) that are much better than the Sony. AT f/2.8 through f/5.6 the Touit has significantly higher MTF50 than the Sony. I had mentioned online that wide open I thought the Sony and Touit were fairly similar, but that stopped down just a bit the Touit was clearly resolving more. These numbers support that.

I should mention also that the Sony lens, even on the NEX-7 does not reach maximum resolution until f/5.6, while the Touit maxes out at f/4. The Sony exhibited less distortion, with 0.5% barrel distortion compared to 2.1% for the Touit in our tests.

It has been stated the 32mm also has a lot of field curvature, but we didn’t see this in our testing. Corner sharpness was not improved greatly by corner specific focusing. Our tests were done at about 13 feet shooting distance. Things may be different at other shooting distances.

Conclusion

This is a great example, I think, of why test numbers are just one thing to consider in choosing a lens. According to the test results, if you want the best bargain and shoot at f/2.8 or smaller, the Sigma 30mm f/2.8 for $200 is just about impossible to beat.

The Sony 35mm OSS gives you slightly better performance optically, a wider aperture, and image stabilization for $449. At f/1.8 it’s pretty equal to the Touit in resolution and has less distortion.

Stopped down just a bit, to f/2.2, though, the Touit gives much better optical performance than any of the other lenses – numerically and by the eye test. Worth $900? To some people certainly. To others, not at all.

For me, taking actual pictures with it confirmed what the numbers suggested. The center resolution is probably the best I’ve seen from any lens on an NEX camera. That’s enough to make it a future purchase for me.

Roger Cicala

Lensrentals.com

May, 2013

26 Responses to “Finally Got Around Touit”

jseliger said:

There are no immediate plans for Micro 4/3 mount Touit lenses.

Which I’ve been asking Zeiss about on their blog but haven’t gotten an answer. Perhaps you have some ideas? Based on my understanding, it’s not terribly hard to make a lens for a camera with a smaller sensors.

Here’s what I left; it starts from with a Zeiss blog quote:

We want to offer lenses of very high quality. The image quality of the system also hinges on sensor size. Therefore, we intend to concentrate initially on the biggest sensor size in this segment which is APS-C. It would be possible for us to make lenses for Micro 4/3-bajonet as well but we have not made a decision on that yet.
If current APS-C lenses were also used for the Micro 4/3-bajonet, we wouldn’t be able to achieve the best tradeoff between lens size (weight), and image quality.

And I replied with this:

Zeiss’s explanation for why they’re not interested in a micro four-thirds mount doesn’t make much sense: “lenses of very high quality” can, on their own, be produced in any size, as long as they’re large enough to cover the sensor. Then Zeiss’s explanation shifts from the quality of the lenses to “the image quality of the system,” but the quality of the image produced can be independent of the quality of the lens.

The shifting ideas in the first three sentences make me wonder if their post was translated from another language, because right now it reads like an LSAT question regarding flawed reasoning, which depends on conflating the quality of the lens itself with the size of the sensor.

It’s true that their lenses would be larger than they need to be in micro four-thirds—but so what? If they’re great lenses, people will use them. More importantly, from what I understand the highest cost in lens development is in the glass itself, not in producing different editions for different mounts.

So: is the cost primarily in the glass, or in the mount and software? Because right now, this: “In order to stay true to the high quality standards of the ZEISS brand, it was decided to focus on cameras with an APS-C sensor” sentence still doesn’t make much sense (unless I’m missing something, which is very possible).

(Also, as a minor note, there’s a typo in the last sentence: “it a futur purchase for me.”)

jseliger said:

(All my blockquotes and formatting got removed, which makes the preceding comment very hard to read. Sorry!)

LensRentals Employee

Roger Cicala said:

The one thing that struck me at the Zeiss Touit introduction was that Sony and Fuji were well represented and all made it clear there was complete cooperation between their companies and Zeiss in designing the Touit lenses, including sharing of autofocus algorithms, etc. I would assume from that that Panasonic and Olympus aren’t interested in working as closely with Zeiss, despite Zeiss being a member of the 4/3 Consortium.

I can’t imagine there is a technical reason for not modifying the lenses slightly for m4/3. The flange to sensor distance is different, so possibly that would require some redesign, but I can’t imagine it being a big problem.

Roger

den_sh said:

Would love to see similar comparison of Fujinon 35mm f/1.4 vs Touit 32mm f/1.8 on Fuji X-Pro1/X-E1. It would be also interesting compare performance of Touit on Fuji vs the same Touit on NEX (the same lens, different sensors.)

LensRentals Employee

Roger Cicala said:

den_sh, you can’t do that comparison: the lenses are either NEX or Fuji mount, not both.

den_sh said:

Roger,
It’s the same lens design with different mounts. I can’t give the source but it must have been mentioned somewhere by Zeiss.

Rodrigo said:

If Zeiss doens’t get the AF algorithms they will not make the lens. The glass design is easy, for them, to adapt.

cputeq said:

It’s a damn shame these are so-so AF and focus by wire instead of nicely made MF lenses.

Am I the only person on the planet who completely hates focus-by-wire? I don’t care who does it – it stinks. Even cheap Minolta MD-mount MF lenses are worlds better :(

That said, IQ seems to be there from the samples I’ve seen. Unsure if I see that “Zeiss look” though.
Thanks for the testing Roger.

Gene said:

While this early test is interesting and valuable, I would have preferred to see the Touit compared to the Zeiss Biogon 35mm f/2.0 ZM. You have already admitted that you have an adapter, so no excuses. This comparison should be closer in optical performance, at least I hope so. And prices are comparable. One factoid I would love to know would be the trade off between optical quality and AF / electronics. Thoughts on relative build quality would also be interesting.
Being a Nex-6 shooter this would be my ideal choice, but a Nex-7 would be OK to, even if it’s an older design. The additional resolution may show the differences between lenses more clearly.
Do you think this would be a worthwhile comparison?

thanks

David said:

Thank you for comparison, interesting. I agree with den_sh that it would be interesting to see the Fujinon 35 1.4 vs Touit 32mm f/1.8 on X-Pro1/X-E1. I think a lot of Fuji shooters are waiting for such a comparison. Because a lot of Fuji owners told that there is almost no difference between the two.

Would really appreciate such a comparison. I know that there is not a high probability that you will do one :)

Trenton said:

I’d like to second @den_sh and @David: it would be really great to see Touit compared to Fujinon 35/1.4, from both Imatest and handling perspective.

Kevin Purcell said:

Zeiss excuses for FourThirds seem a bit weak and seem to focus on “the smaller FourThirds sensor is not good”. It’s 2dB down in overall signal to noise ratio compared to APS-C (that between you can just notice and you can’t notice without ABing the same scene with two different sensors). So that doesn’t seem like a valid reason.

One technical issue with FourThirds is the pixel densty (for the same number of pixels) is a third higher so that makes lens issues more obvious. But that’s true for all FourThirds lenses so why would Zeiss worry about that.

On Roger’s comment: “The optics may vary a bit, too. Differences in flange-to-sensor distance and Fuji’s raw manipulation account for that.”

I doubt the optics will vary at all — after all that’s the difficult bit for Zeiss to design and get perfect.

The mount design will take account of the different flange to sensor distances. It’s only 0.7mm difference in this case 17.3mm on the X mount and 18.0mm on the E mount. That’s easy to accomodate in the mount design i.e. design the lens for the longest “last optical element to sensor” distance you need to accomodate then “shim” the “shorter” mounts.

The microFourThirds flange-to-sensor distance is 19.25 mm. I wonder if they didn’t want to be constrained by the need to put the last optical element another 1.25mm further from the sensor. That might be a part of the “quality” argument though the “quality” in this case is the quality of the lens on APS-C cameras. Just an idea.

The 2.1% distortion seems rather high to me for a “standard lens” from Zeiss. Has Zeiss decided that software correction of distortion is a useful trade off to improve other properties of the lens? Is that a first?

Roger said: “After shooting with them I found the 12mm had fairly blurry corners wide open on the NEX-7. It may be better on the NEX-6 but I didn’t shoot it on that camera.”.

I doubt it. The NEX-7 has problems with oblique rays in the corner (giving “wacky” colors) but the NEX-6 microlenses seem to deal with that (as do the same sort of sensors in the Ricoh GR and the Nikon A). This is a telecentric lens so it should be the same independent of the sensor.

It will be nice to see the “statistical” test when you get a group of them. How reproducible is manufacturing of these lenses? Better than the Japanese lenses from Sony and Fuji?

The Zeiss PR is almost certainly translated from German (as is almost all of their website). It often turns a little wooden and clunky (which I find odd as there are so many good English speaking Germans out there).

CarVac said:

CPUteq:

Focus-by-wire is beneficial in one circumstance: superzooms at full telephoto, up close. My Canon 18-135 non-STM is irritating to MF at long focal lengths, because it’s too quick. My friend found (yes, found!) an STM 18-135, and the focusing is a lot more usable at the long end.

Wagner said:

this clearly shows AF is not zeiss’ strong point. also the af focus noise is a showstopper for me, forget video! and at that price point i see no significant stand out points for purchasing a heavier lens.

Herman said:

Thanks for sharing your first immpressions!

The center resolution of the Touit is awesome indeed, but the corner resolution is a little disappointing in my view (remarkably similar to that of the Sony lens).

How is the Touit doing with bright lights, what about lens flare?

And, yeah, a comparison of Touit 32 mm and Fujinon 35 mm would certainly appreciated by many.

WillemB said:

I want to see the difference in Bokeh between these lenses. The FF Zeiss 1.4/50 had a horrible Bokeh!

Euell said:

I would be interested to see how the Zeiss compares to available rangefinder lenses as, for example, the 28mm F2 Voigtlander. A comparison with Leica lenses would be interesting, but they are so much higher priced than the Zeiss that it would be pointless.

sdchew said:

Out of curiosity, how did you determine that the AF uses a DC motor and not a SSM type motor? Do you know what AF motors are used in the other e-mount lens?

LensRentals Employee

Roger Cicala said:

It’s in the Zeiss specs – but we also took one apart and looked inside. Definitely DC.

sdchew said:

Thanks for the reply Roger. Any idea if the typical e-mounts are also DC or SSM type?

Florian said:

Do you have any word on whether and when a firmware update will be available for the touit e-mount lenses that will allow for phase detection autofocus to work on the Nex-6?

Florian said:

I just went and got the Touit 32 for my Nex-6. Phase detection autofocus works out of the box, provided you have firmware 1.01, as I was able to test in the store. Look and feel of the lens is quite a step above the Sony equivalent, as far as I am concerned, though I am sure they will both find their market.

Florian said:

Correction: while I convinced myself for a moment that phase detection of Nex-6 does work with new Zeiss Touit lenses, it does in fact not work as of yet. So, here’s to hoping that the Nex-6 firmware update rumored for August will fix this.

Ken said:

Very useful information, as always -thanks a lot Roger. Now, if there’s one lens I’d want to see tested by you on the NEX-7 it’d be the legendary 45mm/f2 Contax G Planar. I love my copy to bits.

Florian said:

After spending a few days with the Zeiss Touit 32mm f/1.8 lens on my Nex-6 I took it back to the store and got the Sony 35mm f/1.8 OSS instead. And while I love the look and feel of the Zeiss lens and the pictures I got are superb, I decided that it is not for me and here is why: it doesn’t focus fast enough and it hunts in the process of acquiring focus. This does only cause missed shots, e.g. when your subjects/objects are moving at too fast a pace for it to keep up, but it takes the fun out of taking pictures for me. The Sony lens, on the other hand, focuses relatively quickly and without any perceptible hunting. And never mind that it does so silently and with OSS at half the price.
At first I thought the superior focusing performance of the Sony lens has to do with the PDAF assisting the CDAF but after switching off the PDAF, the Sony lens still acquired focus a lot more quickly and w/o hunting. It appears to “talk” to the Nex-6 body in a way the Zeiss lens simply can’t. Whether the way in which the Zeiss lens acquires focus is inherent to it’s design or whether it can and will be fixed with a firmware update on Sony’s part is anybody’s guess, but it would appear to me that the Zeiss lens would greatly benefit from PDAF assist (for Nex-5R and Nex-6 bodies) and possible gains in CDAF performance (for all other Nex cameras) as well.

Alex said:

About the wide open performances and irregularities with Rogers tests:

Kurt Munger found out in his tests that the Zeiss sel24 1.8 was sharp wide open already but the sel35 1.8 he called soft wide open but sharpens up at 2.8 and the next big step is f4. According to his crops the sel24 is noticeably sharper then the sel 35 both at 1.8f.

When I went to my local photo store I tried the Touit 32 and shot in RAW some ads with Text in the store to see the corner performance and centre too. I already was thinking how much worse my sel24 is going to be but when I went home, surprise!!! wide open the sel24 is sharper in the centre and about the same in the corners!
One thing that the Touit 32 does better ist the CA control … much better then on the sel24…

So I dont see how Roger got these results for wide open at 1.8 for the e35 here.

But there is more funny stuff going on with Rogers Tests:
According to Roger the SEL50 1.8 is supposed to have almost as good corner as center performance on the nex-7 (corners and center in the 650 lines range)! At photozone.de the sel50 1.8 has been tested and it has huge problems on the Nex 7 … corners are 60% worse then centre, according to ephotozine.com this is also the case even on a nex6! When u look at slrgear.com and also kurt munger´s tests they all test the sel50 with farfar worse corners then center!!!

I also have the sel50 and I can def say it is not good in the corners wide open, far worse then the sel24. (it doesnt bother me in portraits and stopped down to f4 its magnificent)

So whats going on here?

Same thing with Sigma 30mm 2.8 on Nex-7 at photozone.de

Leave a Reply