Article Index

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.

 

Like most fast primes, Leica Summilux-R 80mm f/1.4 showed a dual personality, with somewhat softer results at wider apertures and top notch resolution at smaller ones. The f/1.4-f/2 apertures have a statistically measurable (and visually noticeable) difference compared to resolution from f/2.8 and on. This is particularly the case for the corner quality on full-frame cameras, which unmask even the slightest flaw. Center image quality is not as much of a problem and the MTF50 charts claim that the quality is adequate (not steller though) on either full frame or APS-C cameras. Naturally, how much this might really impact one in real life will be examined next, and in the meantime, take a look at the MTF50 charts, which were recorded at the focusing distance of 8m to the target (the test range was 2m-8m)

 

Canon 400D (10Mp)

Normalized raw MTF50 @ 80mm

Canon 5D (12Mp)

Canon 5DMk2 (21Mp)

 

Below you can find the target's corner crops, comparing border image quality at f/1.4 and f/8. Visually, you can see the difference across all cameras, but the results are not particularly bad, meaning that while the quality around borders is indeed visually worse at f/1.4, it does not completely fall apart and might be tolerable to some users.

 

Canon 400D (10Mp)

Image borders @ 80mm (100% crop): f/1.4 vs f/8

Canon 5D (12Mp)

Image borders @ 80mm (100% crop): f/1.4 vs f/8

Canon 5DMk2 (21Mp)

 

Let's actually take a look at how the lens performed in real life situation. Below, you can find the 100% crops taken with a 21Mp Canon 5DMk2, comparing image quality at f/1.4 and f/8 across three areas of the picture frame. The  lens was focused in live view mode first and then the pictures were taken using focus bracketing at around the  original focusing point (around infinity). You could notice a fairly consistent behavior in these tests. There was a slight difference in image quality in the center at f/1.4 when compared to images at smaller apertures, but it is not significant enough to cause concerns. Border quality, on the other hand, did vary more noticeably, and f/1.4 and even f/2 corners were visibly worse in all image corners. This confirms the results from the MTF tests and hence is not particularly controversial. Image quality degrades slightly more as you move to closer focusing distances, but it's impossible to quantify this degradation due to the increased shallowness of DOF at closer focusing distances. For most practical purposes though, this should not be a problem, particularly if your primary use for this lens is going to be portrat type work.

As is the case with all focus bracketed tests, I try to compare the bracketed shots for any sign of field curvature that the lens might exhibit. Naturally, this is not a very scientific experiment, but it at least gives a glimpse into whether one should even worry about field curvature to begin with (and for most normal and tele lenses, this is a non-issue) . Basically, we want to look for shots that look sharper at different focusing ranges, that is if an image taken at f/1.4 and 2m focusing distance is the sharpest from the series, then 2m should be the registered distance and all other apertures should also produce sharpest images at 2m. If that is not the case, then the lens is likely exibiting some field curvature. Bottom line is that this is a rather long introduction for saying that Summilux 80mm did not show signs of field curvature in these tests.

 

Center

L.L.Corner

/

U.R.Corner

 


Color Reproduction

Chromatic aberration (lateral) with Summilux did not vary much from camera to camera - with APS-C body, the aberrationhovered at ~1.1px around borders at the widest aperture level, slowly drifting towards 0.4px by f/11, and was only ~0.3px in the center, remaining pretty much constant throughotu the tested aperture levels. CA was about within the same range on full frame cameras, showing very little variance in behavior.


Canon 400D (10Mp)

Chromatic Aberration (APS-C) @ 80mm

Canon 5D (12Mp)

Chromatic Aberration (FF) @ 80mm

Canon 5DMk2 (21Mp)

 

The lens showed very minor signs of longitudinal chromatic aberration, which can be noticed in image periphery with wide open aperture, but it was fairly mild and one needs to literally pixel peep to find it. Interestingly, the lens showed varying degree of contrast - at closer distances the contrast seemed somewhat lower at wider apertures (f/1.4 and f/2 specifically) then at infinity. This made color pallette look slightly subdued. But more on that in the next section.

 

ISO 100, 1/4000, f/1.4, 80mm (100% crop)
ISO 100, 1/4000, f/1.4, 80mm (100% crop)

 

DOF & Bokeh

There is a common misconception floating around the net which claims that one needs a super fast lens to achieve the best separation between the foreground and background and/or achieve the smoothest bokeh. But in reality, you can often compensate a slow aperture with another attribute - moving closer to the subject or using a longer focal length would do the trick. Obviously, a combination of fast aperture on a medium telephoto lens would be ideal and as can be seen from the shots below, Summilux-R 80mm f/1.4 shows a shallow enough DOF. The first sampel simulates a DOF scale - you can see that only a very small patch of the frame remains in focus at f/1.4, and even slightest move away from the focusing point results in a fair bit of detail loss.

.

 

ISO 200, 1/500, f/1.4, 80mm (Canon 5DMk2)

 

The shots below show what you can expect from the lens when it comes down to the depth of field in real life situations. The first series were made at 80cm (the minimum focusing distance, while the second was made at about 2.5m. As is the case with all bokeh/DOF tests, I would like to warn the reader that bokeh is as subjective item as it can every be and depends on many factors, including the distance to the subject, the distance to the background, the aperture used, etc. As a rule of thumb, the lens would produce the shallowest DOF at its minimum focusing distance and with wide open aperture. As you can observe from the shots below, the lens renders background into a shapeless mask at close focusing distance and wide open aperture. We benefit here from a fairly long distance to the background objects (bunch of trees, tents, building) - different distance and the overall look would be different. With stopped down aperture, the lens starts to bring in more detail into the frame, which is not that unusual obviously, and as you move away from the target, subsequently you get even more detail and busier background. Again, not particularly surprising and so the widest aperture and close focusing render the most pleasing to my eye bokeh.

What is interesting to me in these sample shots, however, is the fact that the images at wider apertures have a fairly noticeable drop in contrast across the frame. Ok, it's really hard to tell from the samples below whether the contrast in the center is much worse because the focusing target is not ideal. But arguably, the lower contrast in periphery is not necessarily a bad thing - if the lens manages to maintain a fairly high contrast in the center along with fairly high resolution, then the photographer can achieve very nice, crisp definition to the subjects in focus. This is the key ingridient for the so called '3D effect' when the subject clearly stands out in the frame.


DOF @ 0.8m

ISO 100, 1/640, f/1.4, 80mm (Canon 5DMk2)

ISO 100, 1/160, f/8, 80mm (Canon 5DMk2)

DOF @ 2.5m

ISO 100, 1/4000, f/1.4, 80mm (Canon 5DMk2)

ISO 100, 1/125, f/8, 80mm (Canon 5DMk2)

 

As mentioned earlier, the lens showed minor variance in contrast reproduction, which seemed to depend on the focusing distance. At closer focusing ranges, Summilux produced slightly lower contrast both in the center as well as in the periphery, giving images a softer feel. As you moved the lens further away from your subject, you could see a noticeable boost in contrast, and as you'd stop the lens down, you'd see even more boost. This seems to be a particular characteristic of the optical design of the Summilux, particularly when you compare it to Summicron-R 90mm f/2, which produces a more contrasty and much sharper images at wider apertures. Of course, depending on the photographic needs, one might argue that this is not a bad characteristic in a lens - softer skin tones in portraits for example, might actually boost the pleasing feel about an image taken with Leica Summilux-R 80mm f/1.4.

As you examine the pictures in this sections, note that smoothness of OOF areas, particularly at wide apertures. The OOF highlights are evenly illumited, giving the background a nice, smooth feel to it. If you zoom in to 100% on the image below (taken at 80cm, with the focusing on the cat's left eye), you can notice very minor CA in the OOF areas, but it's barely noticeable and does not create any distraction.


 

ISO 100, 1/125, f/1.4, 80mm (Canon 5DMk2)