Introduction

Leica Summicron-R 90mm f/2 is one of five telephoto lenses that Leica offers in its lineup of SLR range of cameras. The first version of the lens was released in late 60s but the company has revised its design a few times with the most recent one, APO Summicron-R 90mm f/2, released in 2002. The version tested in this review was an E55 model, which preceded the current APO version of this lens. The serial number of the tested lens is 3342985. The E55 variant of Summicron is quite readily available on the used markets like eBay, with good copies selling around US$450 (as of October 2007).

The optical construction of the lens consists of 5 elements in 5 groups - quite a simple design when compared to some modern alternatives like Canon EF 85mm f/1.8 which has 9 elements in 7 groups or Nikkor 85mm f/1.4D IF which has 9 elements in 8 groups. The build quality of the lens is superb, as expected from a Leica lens and rivals anything you can find on the market. The metal barrel is super solid as are the focusing and aperture rings. The lens looks and feels very sturdy, almost heavyweight, measuring 59 x 70cm (2.3 x 2.7in) and weighing 550g (19.4oz), Although the barrel extends a bit towards close up, adding another 20mm or so to the total length. The minimum supported aperture is f/22, the minimum focusing range is 68cm (2.25ft) and the filter size is 55mm.

Image

As most Leica R lenses, Summicron 90mm f/2 E55 is easily adaptable to a number of alternative mount systems, including Canon and Micro Four Thirds. You can in theory adapt it to Nikon system as well, but you'd need to replace the mount, since Leica's register flange is too close to Nikon's to use a regular adapter. Within the scope of this review, the lens was tested on Canon's 10Mp APS-C type 400D, 12Mp and 21Mp full frame Canon 5D and 5DMk2 cameras. Since the lens was designed for traditional 35mm cameras, if you're using it on APS-C type body with 1.6x crop factor, the EFOV of the lens will be 144mm.

 

Summary
Lens Composition 5 elements in 5 groups
Angular Field ~27 degrees
Minimum Focus 68cm/2.25ft
Focusing Action MF
f-stop Scale f/2-f/22, manual
Filter Size 55mm
Lens Hood None
Weight 550g/19.4lb
Dimensions 59x70mm/2.3x2.7"
Lens Case None

 

Handling

Leica's lenses always feel like a solid chunk of glass and metal fused in a high-tech lab, and Summicron-R 90mm f/2 E55 is no exception here. The feel of precision when you rotate the focusing ring, resistance of the telescopic built-in hood, satisfying clicks of the aperture ring - all leave an impression of quality engineering, particularly when you compare this lens to other lenses in the class. Only Carl Zeiss and to a much lesser degree Voigtlander lenses leave you with similar feelings, but even then Leica reigns supreme in build quality. Not to say that the lens is completely without drawbacks - the metal knurled focusing ring is fairly thin and is a little bit uncomfortable to grip. The ring itself rotates for about 270 degrees when going from the infinity to the closeup, giving you sufficient amount of precision during focusing.

Despite of its 'chunkiness', Summicron-R 90mm balances quite well on mid and large cameras like Canon 5D and 1Ds series, with the center of weight positioned at the base of the camera's mount. On lighter cameras like 400D the weight is shifted slightly towards the front of the lens, tipping the combo down when the lens is extended towards the close focusing distances. With either small or large cameras, hand-holding the camera/lens combo is not challenging, since the lens is not particularly heavy.

I found Summicron-R 90mm f/2 quite easy to focus - with the lens remaining at f/2, the viewfinder of Canon 5D/5DMk2 remained bright and clear. At close focusing distances you can completely rely on your eyes for focusing the lens, but at around infinity you would benefit from an AF confirmation adapter, assuming you're using the lens on an alternative mount of course. Your mileage with the adapter might vary somewhat though, depending on the adapter quality - I've changed about half a dozen Leica-R/EOS adapters before finding one that was accurate enough during focusing. Even then, as you focus the lens, you will discover that in many situations, particularly at closer distances, the camera would give you a focusing confirmation not at a single unique focusing point, but within a small range - the front to back focusing with adapters is quite common and you would often need to practice with the lens a little bit to understand where exactly you are focusing. In my case, I discovered my most accurate adapter front-focused consistently on all Leica R lenses (but did so in a very predictable, uniform way) and so I took a habit of taking the pictures at the rear end of the focusing confirmation range, which consistently gave be best results.

 


Resolution

Please note that MTF50 results for APS-C and Full-Frame cameras are not cross-comparable despite the same normalized [0:1] range used to report results for both types of cameras.

 

Leica Summicron-R 90mm f/2 E55 showcased fairly good overall performance in the lab. The lens sports superb resolution across the frame from moderate to small apertures - from about f/4 through f/11, the lens performs pretty uniformly with no major drop off in quality from center to borders across different cameras, both APS-C as well as FF. Where performance seems to differ is the widest aperture - at f/2 border image quality is noticeable lower (in MTF terms) then the quality in the center. This remains persistent on both APS-C and FF cameras. f/2.8, on the other hand, seems like a transitioning point, where the border quality improves but does not quite catch to the performance in the center. Of course all this seems very abstract since we're dealing with MTF50 measurements. How all this translates into visual terms is a more interesting questions.

 

Canon 400D (10Mp)

Canon 5D (12Mp)

Canon 5DMk2 (21Mp)

 

The corner crops below, which compare borders at f/2 and f/8, do show visually noticeable softness at f/2. It persists on all cameras, but I can't say that the quality at f/2 is completely dismal - softer, but probably is acceptable around extreme borders to many users. Let's see whether this is truly the case in real life example.

 

Canon 400D (10Mp)

Canon 5D (12Mp)

Canon 5DMk2 (21Mp)

 

The crops below were taken with a Canon 5DMk2, using focus bracketing around infinity. A shot with the sharpest center in each aperture bucket was selected and then region crops were compared across aperture settings. You will notice here that the image quality does vary somewhat around borders where we initially detected the difference in MTF charts. In all four corners of the image, f/2 crops are not as crisp as f/8, I am not including crops from other aperture settings, but the difference in quality is practically not noticeable at f/2.8. There is also an obvious difference in quality between the center and borders at f/2 (at f/8 there does not seem to be any), although the difference in the center between f/2 and f/8 is basically non-existent. If the slight softness really bothers you, applying an unsharp mask with r=2 and amount=30%, followed by a minor contrast boost cleans up the image very nicely, further improving the overall definition. The bottom line is that while the image quality towards border areas at f/2 is not quite as sharp as with smaller apertures, most users would probably be reasonably happy with the quality, particularly if they'd be willing to tweak the image a little bit during post-processing.

As is the case with all such field tests, whenever I discover a difference in image quality around borders at wide apertures, I try to check if the drop off in image quality is from field curvature. Basically, while the focus bracketing tests do not allow for measuring the magnitude of the field curvature, they do help detect one when it noticeably affects the image quality. Of course, technically speaking, field curvature does not really reduce sharpness of the image per se, it basically changes the distance the light travels to the sensor plane because of the curved surface of the lens elements. And so to compensate for field curvature, one has to change the focusing to bring the object on the curve into focus. With such a rather lengthy intro, all I meant to say really is that Summicron-R 90mm f/2 does not seem to suffer from field curvature, since the image with sharpest center also retained the sharpest corners.


Center

L.L.Corner

L.U.Corner

R.L.Corner

R.U.Corner


 

Color Reproduction

Lateral chromatic aberration was somewhat of a mixed bag with Leica's Summicron-R 90mm f/2 E55. One one hand, CA levels in the center were pretty minimal, generally not exceeding ~4px. CA around borders, on the other hand, was significantly higher, particularly at wider apertures. In the f/2-f/2.8 range, CA levels were hovering around 1px, and gradually dropping to more manageable levels at lower apertures. Generally this means that your're likely to see some color fringing in high contrast areas at wider apertures, but in practice you need to blow up the image to 100% and carefully examine it inch by inch to notice CA. Look at a couple of images in the sample gallery, as wel las at the image crops above, you really have to try hard to find CA in the high contrast border areas. So the bottom line is that I don't feel that the CA levels would cause too much problems for practitioners.

 

Canon 400D (10Mp)

Canon 5D (12Mp)

Canon 5DMk2 (21Mp)

 

Besides the lateral CA that we discussed above, the lens also showed very minor degree of longitudinal CA, which I discuss a little bit more later in the DOF section. Color palette remained well saturated and images looked quite lively, the minor exception being image periphery at f/2. Local levels of contrast at f/2 were somewhat lower here making the image look more subdued. The lower tonal reproduction in borders also accentuated the feel of images being slightly softer around borders, as mentioned earlier.


 

ISO 100, 1/800, f/2, 90mm (Canon 5D)

 

 

DOF & Bokeh

One of the main uses for a fast medium telephoto lens is obviously a portraiture type work. While Leica Summicron-R 90mm f/2 is not the fastest lens in its category (that title belongs to Canon's EF 85mm f/1.2L), the f/2 maximum aperture is supposed to provide enough creative control over DOF when needed. But Summicron actually holds an advantage here compared to the faster Canon lens - its minimum focusing distance is only 68cm vs Canon's 95cm. As I've been saying elsewhere, the MFD influences the DOF significantly more then aperture level (to some extent obviously, since there is a limit as to how much the closer focusing distance can compensate for slower apertures). One of these days I will conduct a more detailed test to demonstrate how the decrease in MFD compares to the loss in f-stop speeds. Until then though, it's enough to say that if DOF control and hence bokeh are of any importance to you, then your ideal lens would have a fairly short focusing distance and large maximum aperture (and no, wide angle lenses don't count). Anyhow, let's start with a simulated DOF scale, which, while obviously not an absolutely correct way to measure DOF, will give you a sense of how shallow it can be. The shot below was done at f/2 and ~75cm, with the camera pointed at 45 degrees to the surface.

 

ISO 400, 1/250, f/2, 90mm (Canon 5DMk2)

 

Ok, so from the look of it, the DOF is fairly shallow, which is quite promising. More interestingly though, you might notice two more things about the shot - firstly, there's minor degree of longitudinal aberration noticeable in the out of focus area around the necklace clip, and secondly, the entire image seems to have lower level of contrast (right mouse-click on the image and then open it in new tab if you want to look at the full-size original). While looking at the out of focus area around the necklace clip, also notice the plain ugly highlights, with sharp, pronounced edges. Ok, so there are three things to worry about, but considering that you will not be shooting rulers located on a table in real life, let's see how all this translated into real life and whether we should really be concerned here.

The series of shots below demonstrate what you can expect from the lens in real life, when you shoot at close and moderate focusing distances. Since bokeh depends on quite a number of factors, including the distance not only to the subject in focus, but also to the background objects, you should not automatically expect exactly same bokeh in your shots. In these tests I try not to vary the distance to the background and keep it at close to the lens's infinity (about 10m to the closest palm tree and over 50m to the facade of the farthest building). The farther is the distance to the OOF background object, the better is the chance of the lens smoothing it out nicely when rendering the bokeh. As you can observe from the shots, f/2 and close focusing distance render the most pleasing results - the object in focus stands out and the background is rendered into a shapeless mask of subdued colors. There is no sign of harsh edging around OOF objects whatsoever, the highlights actually look quite neutral/pleasing to my eye. Even with stopped down aperture, the lens still produces a fairly decent bokeh - there is obviously more detail and so it feels somewhat busier, but not completely distracting. OOF highlights continue to retain neutral feel.

Move away from the target and the bokeh feel changes as well. At f/2 the lens still tries to retain some of its best characteristics like neutral highlighting and low contrast background, but the foreground does not stand out as much because of the increased clutter in the background - significantly more detail is carried in into the frame at this level and if you continue to stop  down the lens, the bokeh will start to feel too busy. Ok, so basically all these tests did not reveal any significant longitudinal aberration seen in the closeup shot above, nor harsh edging around highlights, which basically tells me that both can occur only under some extreme circumstances. What extreme situations? My guess is that the objects that are on the immediate border of the focus place might have somewhat harsher definition to them. Let's do another test to check whether this is the case and move on.

 

DOF @ 80cm

ISO 100, 1/4000, f/2, 90mm (Canon 5DMk2)

ISO 100, 1/250, f/8, 90mm (Canon 5DMk2)

DOF @ 2m

ISO 100, 1/4000, f/2, 90mm (Canon 5DMk2)

ISO 100, 1/160, f/8, 90mm (Canon 5DMk2)

 

The shot below, also done at f/2 and 80cm focusing distance (the point of focus is on the cat's nose). Here I do see very occasional harsher edging around OOF highlights, but it is not particularly extreme. I also see minor color shifts in the OOF highlights, particularly those that were rendered around the close to in-focus areas around cat's paws and necklace. The bokeh though still feels quite pleasing though as these minor distractions do not affect overall look.

 

ISO 200, 1/800, f/2, 90mm (Canon 5DMk2)

 


Flare

I was quite surprised to to discover how much the Summicron-R 90mm f/2 E55 flares. The two shots below compare the behaviour at f/2 and f/8, with the sun positioned right above the building, in the left upper corner of the picture frame. You can see that with both wide open and stopped down apertures, the lens shows noticeable glare, ghosting and color shifts, along with the generally reduced contrast across the entire frame. Of course, this is probably one of the worst possible cases and I don't expect many users will be composing their pictures with the sun (or any other light source) shining directly into the lens, but if you do end up in such situation, use the built-in telescopic hood - while it is fairly shallow, it's sufficient to cut all the flare out for 99.9% of cases out there.

 

ISO 200, 1/2000, f/2, 90mm (Canon 5DMk2)

ISO 200, 1/640, f/8, 90mm (Canon 5DMk2)

 

 

Vignetting

Given its large maximum aperture, Leica Summicron-R 90mm f/2 E55 shows a pretty moderate amount of light falloff. On an APS-C type body with smaller sensor, the lens shows very minimal amount of vignetting even at its widest aperture - ~0.4EV at f/2, which then slowly drifts to lower levels. Vignetting on full frame cameras is higher, reaching ~0.8EV at f/2, but is still manageable from practical perspective.

 

The shots below demonstrate what vignetting will look like in real life. On a FF camera, you will see mild darkening around corners in light colored scenes, so you might want to either activate your in-camera vignetting correction or do that during post-processing. On APS-C cameras, there is basically no darkening of corners at all. In either case, vignetting is too problematic from my perspective.

 

Vignetting @ f/2 - full frame vs 1.6x crop

 

 

Distortion

Leica Summicron-R 90mm f/2 E55 showed basically no distortion - at ~0.03% barrel distortion is completely superficial and will not be visible in real life.

 





Alternatives

Assuming you want to stick with Leica lenses, you might want to evaluate the latest 90mm variant of this lens - APO Summicron-R 90mm f/2, which by many accounts improves the image quality even more when compared to the older E55 version, but unfortunately would also cost you a whopping US$3,400 if bought new. If you need an even faster lens, then Summilux-R 80mm f/1.4 would offer that extra speed but at an even higher price - US$4,000 if bought new. For those of us with limited resources, the latest Carl Zeiss Planar T* 85mm f/1.4 in ZF/ZS mounts or its older variant in Contax/Yashica mount might be a viable choice. If speed is not very important, then Carl Zeiss Sonnar T* 85mm f/2.8 (review) for Contax mount could prove an excellent choice.

 

Recommendation

Leica Summicron-R 90mm f/2 is a very solid overall performer. The lens showcases excellent image characteristics - sharpness, color, contrast, distortion, all are in the right place (yes, sharpness around borders at f/2 is a little bit softer, but it should not matter for practical purposes). The lens sports superb build quality and although the pricies for the lens have been slowly creeping up, they are still more or less reasonable for what the lens offers. With its fast maximum aperture, the lens would be an interesting choice for portraiture type work. Of course you still need to deal with manual focusing, and this might be the only drawback to some users.


Sample Images

 

 


User Reviews