The new Summilux-M 1.4/35 ASPH. adds the floating element feature to the basic optical design of the 1994-edition of this very classical and popular lens. The f/1.4 aperture was married to the 35mm focal length in 1961, a first in those days when all other designs had a maximum aperture of f/1.5 or smaller. The 35mm focal length changed guard with the 50mm lens as the primary lens for Leica M users in the period from 1955 to 1965 when the style of documentary and human interest photography asked for a close encounter with the subject matter. The 35mm focal length is eminently suited for this visual frame because you can have a format-filling main subject encapsulated by a meaningful environmental statement. A wide aperture gave the opportunity to take pictures in low ambient light where claire-obscure pictorial composition was needed.
The Summilux 1:1.4/35mm stayed in the catalogue till 1995, but in 1989 a new version with two aspherical surfaces was added to the range of Leica lenses. The production process was quite complicated and therefore expensive and a new somewhat simplified version was introduced in 1994. That version had one aspherical surface and was in production until recently and has now been superseded by the new SX35 FLE. The technology of floating elements where a whole group of lenses moves dynamically according to the distance setting was introduced in the M-line with the SX50 ASPH in 2004. Since that year, Leica has upgraded several high speed wide angle lenses to the FLE construct, notably the SX21, SX 24 and now the SX35.
The main task of the FLE construct is to compensate for the loss of contrast that happens with stationary lens groups that have been optimized for longer distance photography. The spherical aberration (SA) which is always present in a lens is responsible for this behavior. In very high speed lenses this same aberration also causes a focus difference when stopping down. It is evident that the aperture stop and the focusing movement are not physically linked and the FLE construct cannot be used directly to battle the focus difference. But when a designer knows that (s)he has more design parameters, the overall design can be improved to correct more than one problem area.
Below left: Summilux-M 1.4/35 ASPH, (1994)
Below right: Summilux-M 1.4/35 ASPH. FLE (2010)
SX35 FLE
The physical dimensions have not changed much, the length is the same with 46mm and the front ring diameter is still E46. Weight is identical with 320 grams, but the internal lens mount has been substantially modified to make room for the floating element construction. Leica also changed the rangefinder cam of the focusing mount from the classical brass color to black. Romantically inclined Leica aficionados may deplore the fact that you now cannot see the individual adjustments, but with the computer controlled machining the black mount is at least equally precise. And the black version is more effective in evading reflections and flare. The focusing mount has a reduced diameter from 56mm to 53mm. In the slimming process the focusing level has not been dropped. The lens livery follows the style of the other lenses in the wide angle and standard focal length range.
The distance settings on the focusing mount are changed a bit: the FLE version has 1; 1.2; 1.5; 2; 3; 7; infinity markings where the SX version has 1; 1.2; 1.5; 2; 3; 5; 10; infinity markings
Performance in the close range
Leica states that the SX35 FLE is optically identical to the previous version. A comparison of lens design and MTF graphs clearly show the validity of this statement. Basically then the evaluation of the SX 35 ASPH. is equally vald for this lens, as long as you restrict yourself to the infinity setting or in practical terms to the object distances of 4 meters and longer. Many users of the 35mm focal length do operate in the distance range from 1 meter to 3 meters. This is the range where the wide angle lens can show its qualities.
The two questions to answer are: is the focus shift reduced and is the close range performance improved. The short answer is to both questions is an unhesitating yes. But there is a caveat as is often the case. First the test results.
The M9 was positioned at a distance of 1.35 meter for a test chart (the well-known Siemens star made by Image Engineering). I have noticed over a longer period that the print quality of the test chart is of paramount importance for the results. More about this in the S2 review. The IE charts are the best you can buy. The camera was put on a device where you can change the distance by millimeters. At first the camera was optimally focused (an act that demands patience and accuracy). Then there was a range of pictures made without changing the lens focus, but with the camera moved by 1 cm over a rage of 10 centimeters. This was done for both lenses, an SX with and without FLE. Two apertures were used: f/1.4 and f/2.8.
See the table below.
distance | SX FLE 1.4 | SLX FLE 2.8 | SX 1.4 | SX 2.8 |
0.05 | 16.6 | 18.3 | 19.1 | 17.4 |
| 0.04 | 18.5 | 18.8 | 20.2 | 17.6 |
| 0.03 | 19.6 | 19.6 | 20.9 | 18.1 |
| 0.02 | 20.7 | 20.3 | 20.9 | 19.4 |
| 0.01 | 21.7 | 21.2 | 20.6 | 20.2 |
| 0 | 21.8 | 21.8 | 19.9 | 20.8 |
| -0.01 | 21.9 | 22.5 | 19.1 | 21.4 |
| -0.02 | 21.4 | 22.5 | 18.4 | 21.9 |
| -0.03 | 21.1 | 23.4 | 16.6 | 22 |
| -0.04 | 20.2 | 23.4 | 15.6 | 22.4 |
| -0.05 | 19.1 | 23 | 15.1 | 22.9 |
The number in the columns is a merit value. The higher the number means the better the performance. The distance setting indicates that the 0.05 position is at the 1.40 meter and the -0.05 position is at the 1.30 meter setting. The first conclusion is that the SX FLE at f/1.4 has its best position at the indicated distance setting (plus/minus 0.01mm!). The SX has its best position slightly closer to the indicated setting. Remember that we are talking about differences of at most 2cm out of 135 cm or a percentage of 1.5%. This is really accurate! If we look at the optimum setting wide open the FLE has a value of 21.9 where the SX brings 20.9 on the scale: the claim that the wide open performance in close up range has been improved can be validated. For the focus shift the conclusion is clear: SX FLE at f/2.8 keeps its performance at nominal distance setting and there is a small focus shift to the front over a total distance of 3 cm, which is as expected. The SX on the other hand has a focus shift of 5 cm, almost twice the range of the FLE version. This is a big improvement. Again as a percentage it is 3% and we should put this in perspective: a slight horizontal movement of the body during focusing/shutter release creates a bigger difference. Remember too that we are talking about a lab setting where great care has been taken to get the optimum results. In general work, even on tripod, the differences will be smaller. But the experimental results indicate that the focus shift is substantially removed and that the performance wide open and stopped down has improved in the close range domain from 1 meter to 3 meters. It depends on the quality of your own work if and how good you see the differences, but the improvements are real and visible.
Overall performance
Wide open the SX FLE delivers a clean and crisp image with excellent definition of very small details. See below a small part of the full image taken wide open at close range.
The corners and edges are of lower contrast and details are depicted with softness. At f/2.8 you see the improvement based on better control of the focus shift. The graph of the 40 lp/mm has improved contrast at the center of the image, compared to the SX version. At f/5.6 the FLE holds the lead and now the performance exceeds the possibilities of the CCD capture element of the M9/M8.
The lens does improve when stopping down: the increase in overall contrast gives images a clean and transparent look and fine detail is seen with crisp outlines, emphasizing the excellent life-like representation of three dimensional objects. Generally one could remark that the optical and mechanical improvements are best seen when you use the digital workflow. When film is employed in the M3/MP range the differences are a bit masked by the emulsion thickness, but more visible when using the current mono-dispersal emulsions.
There is some coma visible at wider apertures at the longer distance settings. The effect is quite small, but visible. Secondary reflections and flare are still present under adverse situations. It is less evident than with the SX design. It this domain Leica has made good progress and the delay in the production has partly been caused by final improvements in the flare reduction by coating technology.
Bokeh is a bit unruly in the background with specular highlights. On the other hand the unsharpness gradient is wide open and at close range very pleasing. On the other hand the unsharpness gradient is wide open and at close range very pleasing.
Flare is visible when you force the lens to capture very strong lights in oblique position. Night pictures wide open show excellent contrast in the shadow areas and good containment of high-light areas.
Vignetting
In the days of AgX photography the property of the lens to illuminate the negative area evenly into the far corners (the absence of light fall off or vignetting) was an important parameter. It was a bit overvalued as a slight darkening of the edges hardly bothered the subjective quality of the image and in many cases was even an integral part of the scene composition. With digital files as the basis of imaging it is easy to calculate the amount of vignetting and to correct this in the post-processing stage. Leica has this information and uses it to suppress the amount of vignetting. In order to evaluate the effects, I made a number of pictures with the SX35 and SX35FLE wide open and with the lens detection on and off. The conclusion is evident: both lenses have the same amount of vignetting, which is not surprising and the effect of the lens correction in post processing is visible too.
You can see a slight change in color correction of the JPG files too in the AUTO-mode of the camera.
Focus shift
Some time ago the topic of bokeh became a hype and you could not discuss the behavior of a lens without mentioning or demonstrating the effects of this phenomenon. In recent times the interest in bokeh is less great, but there now software programs where you can emulate bokeh or even introduce the effect where there is none in the image.
Another topic that has generated much discussion is the phenomenon of focus shift. In all lenses where the spherical aberration is still present (and that is the case in almost every lens) the plane of sharp focus wanders when the lens is stopped down. The total length of this shift can be calculated and the axial positions of the focus plane can be established too for every aperture. You can now decide to select the best focus for the maximum aperture and let the plane drift when stopping down. Or you can select a position between the two extremes to minimize the drift. This option has been used with good effect in the Nokton 1.1/50mm. Leica uses a different solution and here too the effect is clearly seen in the positive sense of course.
The focus shift of the old Sx35 ASPH has been discussed in the internet discussion groups and websites. As usual with any complicated phenomenon both extremes exist: there are some persons who have never detected the focus shift and there are some who declare that the effect in the SX35 is dramatically large. Without visual proofs or measurements made in the lab this last statement has no value. That you do not see a certain effect is no proof that it does not exist, but it is a sign that the effect might be too small to detect in normal photography. To give some foundation for the debate I have selected a series of pictures from the ranges I have made to study the focus shift of both SX lenses. I have selected the pictures that are made with the lens in the position where it is when you focus with the rangefinder. See table at the start of this article. This is the 0-position in the range of pictures from -5 to +5 that I made. When you stop down you do not change the lens focus, so the location stays the same. To emulate this effect I used the 0-position for both lenses at maximum aperture and at f/2.8. If you stop down to smaller apertures the focus shift increases but is compensated by the larger depth of field and the overall effect is minimized. At f/2.8 the depth of field is still small enough to see the effect. The pictures below are the original DNG files without any post processing, the true raw negatives so to speak. When you use sharpening the result will be different, but as noted in the X1 review a comparison is most meaningful when you reduce the number of factors that can influence the result or effect.
The quality differences are small and only visible after very critical scrutiny. I also added the picture of the SX35 with best quality wide open and again you see small differences in contrast and definition of very fine detail. The conversion from DNG to web-based JPG also reduces the visible differences. An important restriction is the limit of resolution of the M9 sensor so the best way to look at the examples is to find the reduction in contrast.
The conclusion then is that the focus shift in the two SX 35 lenses is measurable with appropriate equipment and is reduced in the FLE version. It is an overstatement to designate the effect as "dramatical". You can of course try to emphasize the effect by force, but that is not the goal of this exercise.
Conclusion
Finally I would like to ask attention for the fact that most modern Leica-M lenses (those designs that have been designed starting with the Kölsch era) provide excellent image quality. In many cases they are much better than what the photographer in the field can use, certainly with the M8 and M9 sensor limits of resolution. Reviews of Leica lenses are becoming a bit boring so to speak as the general quality (contrast, definition of fine detail and so on) at all apertures and distances is so good that you really have to search for meaningful differences. Improvement could be created in the one field where the Leica lenses are a bit weak: that is the performance in the zonal area at an image height from 14mm to 18mm and beyond 20mm image height.
The consistent high quality of the modern Leica-M lenses may the cause of the fact that most discussions about Leica lenses are drifting to extreme and often mythical, if not mystical proportions. The facts on the ground however do not support these extremes. The new SX35FLE also shows that Leica now can claim that they are able to design and produce lenses with a very consistent image quality of a very high order. One of the points of critique of the older SX35 ASPH was the occurrence of flare and in this area the Zeiss competition was ahead. Now Leica has significantly improved the control of flare. As the examples above show you can force the lens to exhibit flare, but you must do this with some determination. In normal situations the flare is effectively reduced to small proportions.
The images made for this comparison might indicate that the differences between the SX35 and SX35FLE are small. In fact the main difference is enhanced contrast.
The new SX35FLE is a definite improvement over its predecessor. But the main advantages are to be found in the focusing range from 1 meter to 3 or even 4 meters. There the performance gain is visible and to be appreciated. If you already own an SX35 ASPH and work mainly at larger distances, you might not see much improvement. For closer range subjects it is a matter of critical distinction. If you did not find faults with the SX35ASPH image quality or do not need the ultimate in quality, the urge for an upgrade is less pronounced. If you own a non-asph 1.4/35 or even an 2/35ASPH and want more punch wide open and up to f/4, the new SX35FLE is a very tempting proposition.