The new Zeiss ZM lenses 1: theory
The new Zeiss ZM lenses are now slowly being introduced to the market. This lens line represents the latest insights of Zeiss into the design and production of high quality optics for 35mm photography. These lenses therefore deserve a serious study of the capabilities and properties. I did receive from Zeiss, Oberkochen a set of ZM lenses, the 2.8/15mm, the 2.8/25mm, the 2.8/28mm, the 2/35mm and the 2/50mm. The 2.8/21mm and the 2/85mm will be delivered later this summer.
In this part I will discuss the theoretical issues and a full review of the lenses will be published in a month time.
The third generation of lens design.
Zeiss designs have been instrumental in the evolution of optical design. The Planar and Sonnar design types have heavily influenced the construction of optical systems during the last 80 years. Around 1935 Zeiss introduced a series of lenses for the Contax, their rangefinder camera to compete with the successful Leica camera. In those days, optical calculation was done using seven-place logarithms and the laborious calculations of ray paths through the individual lens elements. Third-order theory was considered a crude instrument and chief designers had ray-tracing assistants
Once a lens design was completed, the quality control was done with test plates to check the sphericity and micrometers to test the thickness. Optical shop assistants were very skilled in interpreting the test plate results and could infer from the test results whether astigmatism was present and could even identify the source of the problem (in this case the back air space is too large). The designers tried to reduce the number of glass elements and to avoid complicated glass surfaces because of the difficulties of manufacturing. They also stayed within the traditional and well-understood designs because of the accumulated wealth of knowledge about these lens systems.
Mechanical calculators made more complicated calculations possible and now third order theory became a serious tool for optical designers. In essence however, the Zeiss designs from 1930 to 1960 for the Contax rangefinder camera did not change that much.
During the fifties a new (second) generation of lens design appeared. The introduction of the computer, and, more importantly, the introduction of optical design programs gave designers the tools to create any lens construction they could dream of and now the evaluation of a lens could be done on screen and not in the shop with test plates. Highly efficient optimization algorithms replaced the laborious ray tracing calculations.
The Zeiss lenses for the Contarex system were the result of this second-generation design approach. I have stated it repeatedly, but the size of the lens, is one of the most important parameters for the optical quality. The Contarex lenses are proof of this statement. Every lens was optimized in performance without regard for the physical size and the designer allowed the lens to grow to its natural proportions. The result was a lens line of impressive performance for that time. There is some mystique around the Contarex lenses: they are sometimes described as the best lenses ever made with an optical quality never surpassed. This is not true. Some designs, like the 4/35 and the 4/135, are indeed close to perfection, but given the modest apertures, that is not a big challenge. The mounting and centring of the lens elements is indeed not yet surpassed.
The great revolution in optical designs occurred during the sixties and seventies, when Dr. Glatzel of Zeiss introduced new methods of lens design. In the Leica-centric view of the world, one often assumes that well-known Leica designers, like Dr Mandler, are in the front rank of the world's best optical designers. In fact this is not the case. Persons like Dr Glatzel or Dr. Kingslake have had a more lasting impact on the evolution of lens design. It is only quite recently that Leica lens designs have broken out of the traditional constraints that have restrained the free flow of creativity of the design department.
The approach of Dr. Glatzel is based on two pillars.
The first pillar is the idea of design relaxation. A lens is considered to under strain, when every lens element and glass surface has a major contribution to the final performance. Such a lens can only be manufactured with the most stringent of tolerances and is costly to produce. Every deviation from the ideal surface shape or location of the lens element will significantly reduce the image quality. If we are able to spread out the correction of lens aberrations over a larger number of lens elements, we can accomplish that the contribution of every single surface or element will be reduced and a defect will have a lower impact on the overall performance.
The second pillar is the idea of the extended design. Wide-angle lenses, especially the retrofocus designs, have often very large front lens diameters. Such designs are relatively short, but quite wide. Excellent performance is possible with this design approach. Zeiss however, wanted to avoid the mistake of the Contarex lenses where every lens needed a specific mount and every lens had a different size, which is ergonomically unpleasant for the photographer. Zeiss then tried to elongate the lens mounts: the lens barrel will be longer, but smaller in diameter. The optical 'trick' to accomplish this is to position the entrance and exit pupil locations farther apart and farther away from the front (back) lens surfaces. The designer has to break up the middle section of the lens design: the section around the aperture stop. If we split up the lens elements around the stop in a symmetrical way, we can extend the design, but we need more elements to get the rays through the longer pipe. Today this is not a big problem with new glass types and effective coatings. The knowledge of zoom lens designs is of course quite helpful in this respect.
The lenses for the Contax RTS are based on this thinking and are indeed quite uniform in size and performance. The Contax RTS system however did not succeed in the market. It is of some historical interest to remark that the European camera makers are more successful with rangefinder designs, where the Japanese makers excel in single lens reflex cameras.
The Contax G series was also based on these design principles. The G-series of lenses deserve high praise, but they again did not survive in the market. It might be the role of a future historian to analyse the reasons why this type of camera did not make it.
The ZM lenses
With the introduction of the ZM lenses, Zeiss seems to have learned their lesson. With the adoption of the M-mount bayonet, they assure themselves of a potentially much larger market. And the decision to outsource the manufacture of most lenses to Japan (the Cosina factory where the Voigtlander lenses are manufactured) will hold the prices to an acceptable level.
The ZM series of lenses are new designs, not related to the G-series. The back focus of the G-body is 12mm and the Leica M has a back focus of 15mm.
The ZM designs are evidently targeted at the equivalent Leica designs. The exception is of course the 2.8/15mm lens. This lens is equipped with an aspherical surface and internal focusing. There is no range finder coupling and focussing has to be done by the old-fashioned distance guessing, hardly a problem with the generous depth of field. Zeiss has always stressed the fact that lenses should be designed for photographic use, not as specimens for lab testing. The requirements of photographers may be not related to the parameters that can be tested on a numerical basis.
It is an irony of history that Zeiss has introduced the MTF graphs for better understanding of the properties of a lens and for general enlightenment of the public and now have to play down the importance. Currently we are in the stage of extreme number fetishism, driven by photo magazines that are claiming that the more numbers you crunch, the better should be the result. A test result, produced by a computer program that evaluates two million measurements, is always better than a result, produced by a modest 25 measurements. That at least is the claim!
If you cannot relate the figures and facts to the requirements of the working photographer, no amount of number crunching will improve the assessment of a lens and its practical validity
Zeiss (and the same is true for Leica) follow a sensible approach: MTF measurements are required reading for the person who can read and interpret the graphs. The amount and quality of the information are the best single source to study the optical behaviour of a lens design.
But the working photographer is not that much interested in the level of residual aberrations. Once the lens design does ensure that contrast and definition are beyond what is needed for accurate reproduction of the subject, other parameters are of paramount importance. Even illumination, flatness of field (the famous Petzval sum), reduction of flare and reflections, halo suppression around light sources, the relation between maximum sharpness and sharpness spread over the depth of field region, the performance in the near focus range, colour reproduction and the performance at the medium apertures are aspects that are often more important than the sheer wide open performance.
Excellent image quality at the maximum aperture can only be guaranteed by a meticulous standard of engineering. It is not uncommon to hear complaints about the performance of a lens that can be directly related to false focusing (by the eye or by the engineering quality of the rangefinder system). One should really not underestimate the error range in handling and mechanics when assessing the quality of a lens.
The Zeiss claims
Here we enter a fascinating topic. Zeiss states that the ZM lenses provide even illumination and flatness of field and close focus performance of a higher level than the competition (read Leica). They also have been trimmed the design to provide a harmonious spread of unsharpness over the depth of field when recording subjects with extended depth. These characteristics are design choices, partly related to the philosophy of the extended design, discussed above. The use of more lens elements and a mainly symmetrical construction allow for a specific quality of imagery. The smaller front lens diameter might help the reduction of flare and reflections, but the absence of aspherics restricts the performance at maximum aperture.
We may compare this approach to the current Leica M designs, where the emphasis is on maximum performance at the wider apertures, the employment of aspherics and special glass types to improve the overall performance. Leica also tries to reduce the number of lens elements to the absolute minimum that is required to do the job. It is clear that the Zeiss and Leica designs share much common ground, but differ in their final choices to provide a specific quality profile that they assume will suit the working photographer best. It will be most revealing to study the quality differences that can be seen in pictures made with both lens systems, based on design approaches with subtle differences in flavor and philosophy.
The new Zeiss ZM lenses 2: the opto-mechanical aspects
In this second installment I will focus on the optical characteristics and the mechanical aspects. In the third and last installment I will analyse the imaging capabilities. This installment will address the measurable characteristics. The next part will concentrate on the more subjective aspects, without, however, neglecting those characteristics that are quantifiable.
The optical performance
Biogon-T 2.8/25 ZM
This is a nine-element design in seven groups. At full aperture we note a high contract across the whole field. It is almost free of distortion and there is hardly any curvature of field. In this respect it is better than the Leica Elmarit-M 2.8/24mm. There is a trace of longitudinal and lateral chromatic aberration. In fact this is an indication that the lens is highly corrected. With many lenses you do not see any chromatic aberrations, not because there are none, but because the primary aberrations (coma, spherical etc) are covering up the chromatic aberrations. On axis the Elmarit 2.8/24mm has slightly higher contrast. Stopped down the Biogon equals the Elmarit in most respects.
There is no decentring to speak. Mechanically it is very well finished.
This lens is outstandingly good and rivals the Elmarit in its optical performance.
Biogon-T 2.8/28 ZM
This is an eight-element lens in 6 groups. There is some play in the mount, but the centring of the optical cell is good. Wide open there is some flare on axis, indicating the presence of spherical aberration. Also a trace of coma can be detected, reducing the overall contrast somewhat. At aperture 4, the coma is gone and the lens improves quite visible in contrast. The comparison Summicron-M 2/28mm has wide-open better contrast, less coma and generally an improved performance.
The Biogon is an excellent lens, exhibiting the characteristics of the Elmarit-M 2.8/28 of the current generation.
Biogon-T 2/35 ZM
This is a nine-element lens in six groups. At full aperture there is a fair amount of coma, but hardly any curvature of field. The coma does reduce the overall contrast quite visible. At f/4 coma is gone completely and the lens from there is an outstanding performer. Compared to the Summicron-M 2/35 ASPH the ZM has better curvature of field, but less contrast and crisp definition wide open. Performance on axis is equal between both lenses. This Biogon is a bit overstretched as a high-speed design. The design itself does not support high-speed lenses and it would have been better if the marketing people of Zeiss had restricted themselves to an aperture of 2.8.
The original 35mm lens for the Contarex has an aperture of 4 and was already at that aperture almost diffraction limited. Had this lens been a 2.8 design, it would have much better performance. Stopped down as it is now the performance from aperture 4 is beyond reproach, but at wider apertures it is at the level of the fourth generation of the Leica Summicron 35 lens.
Planar-T 2/50 ZM
For several generations the Planar design has tried to challenge the Summicron 50mm and never became as good. Now at last we have a lens that equals the Summicron-M 50mm and is even a trace better in the curvature of field area. The optical performance of the Planar is simply as good as that what can be expected form the Leica Summicron. The Double-Gauss design has been studied exhaustingly and it is now possible to equal but not surpass the Summicron design as long as you stay within the D-G limits. It is worth some study to note that the curved elements of the Planar bring no significant improvements in comparison to the many planar surfaces of the current Summicron.
This conclusion makes the claim of some Leica collectors, that the current Summicron is a lesser design than the all-curved predecessor, somewhat hollow.
Distagon-T 2.8/15 ZM
With this lens Zeiss throws in all technology they have: with eleven elements in nine groups, an aspherical surface, a floating element construction and exotic glasses, this lens is an example of modern optical design. There is some decentring, but on axis the performance is outstandingly good, when we account fro the very wide angle that has to be covered. You cannot compare the absolute quality of this lens with a 21mm lens, so you have to use other 15mm lenses to get the proper reference. The Leica-R 2.8/15mm was used as comparison and both lenses proved to be equal. This implies that from now on the Leica M user can own and use a 15mm lens of excellent performance and bring into Leica M photography the perspective and composition possibilities of an extreme wide-angle lens.
The Voigtlander 4.5/15mm lens was also brought in as a comparison and proved to be quite good. The gain of one and a half stop and a better image quality brings in a much bigger lens, the size of the Noctilux. If you only need a 15mm occasionally, the Voigtlander is the one to buy, but if you are serious about very wide-angle photography, the 2.8/15mm ZM is the only serious choice for an M user.
The ZM lenses have aluminium mounts and are not available in chrome (silver and black). The black versions are painted. The mounts are made as similar as possible to save on the tooling. The rangefinder curves follow the Voigtlander approach: the curves are painted black and not individually machined to match the movement of the cam follower. With the ZM lenses a different technique is used: the double helicoid is matched in pairs with a screw in a slot to adjust the distance. It is a matter of taste whether you like the naked look of the Leica lenses where you can actually see the rear mount or the Voigtlander method where the rear mount is disguised. Theoretically the Leica solution might induce some additional flare, but in practice the causes of flare are so numerous that this single one might not be important.
The mechanical infinity setting on the ZM lenses has a tolerance range between 1/100 and 3/100, where the comparison Leica lenses have a tolerance range from 1/100 to 2/100.
Most interesting is the adjustment for the optimum focus plane. In theory you want the true focus plane to coincide with the location of the film plane in the film gate. But the true (or paraxial) focal plane does not necessarily give you the best overall contrast at a target of let us say 20 lp/mm. All ZM lenses had an adjustment in the minus direction, which is in front of the film plane. The Leica lenses were adjusted to the plus side, that is behind the (theoretical) film plane. From these figures it is clear that Zeiss assumes that the film surface is bulging outward, where the Leica designers assume that the film is curved inwards, or at least that the optimum location is in the emulsion layer and not slightly forward. We are talking here is quite small dimensions, a thickness of a handful of (human) hairs, but it may be significant. This type of adjustment may be an additional factor in the misalignment of the Konica lenses for the KM mount when being used in the M environment.
Optically the Zeiss lenses can be divided in two groups: those challenging or equalling the Leica designs and those that are as good as the non-aspherical Leica lenses of the same focal length. Generally the ZM lenses offer improved curvature of field, but less definition on axis. The next article will discuss whether these measured differences are important for practical photography.
Mechanically the ZM lenses are a major improvement over the Voigtlander lenses: the Voigtlander designs are optically quite impressive, but the theoretical performance is not always available to the user, due to a higher level of mechanical tolerance. Especially the V-lenses with high apertures (1.2 to 1.9) are prone to deviations because of the wider tolerance band during assembly and quality control. On reflection this is not surprising: there must be some relation between manufacturing cost and selling price and quality.
The Leica lenses are immaculately finished and mounted: here one can be assured that wide open every lens performs as specified and you pay for this level of quality assurance.
The ZM lenses are located between these two extremes: they are a living proof that high quality manufacture and good quality assurance can not be paired to a low cost production.
The ZM lenses share many parts and their mounts are quite similar. This is not only an element of cost reduction but also of a clear approach to haptics and ergonomics. When you change lenses and can handle them in the same fashion, you need less time for individual adjustments and can operate lenses in the same manner. This is the Zeiss approach and is a reaction to the Contarex designs that were very different lens for lens and here operators complained that they had to adjust to the handling for every individual lens.
The Leica approach offers a different perspective. The Leica mounts follow a general pattern that is the same for all lenses, but every mount is individually optimized such that the character of the lens is visible. When looking at a lens, you see immediately the personality of the lens.
The ZM line of lenses offer some very interesting alternatives to the Leica suite of lenses and the introduction of these lenses may help promote additional interest in the rangefinder concept and silver based photography.
The next installment will look at the actual results that can be achieved with these lenses.
The new Zeiss ZM lenses 3: the practical aspects
This is the third and last installment of the report about the Zeiss ZM lenses. In this part, I will look at aspects that are important for the practitioner of photography and are hard to capture with a technical analysis alone. I will make only passing comments about the bo-ke of the lens. Some writers have chosen to make bo-ke the centerpiece of lens analysis, because of its assumed pictorial quality and its influence on the subjective qualities of the image. All of this may be true, but you can overshoot your aims. Just as with resolution figures that can become an obsession to some, we should keep a balanced eye on the total range of image parameters. However we may dispute this basic fact, but a lens in its core business is an opto-mechanical tool for accurate reproduction of the scene in front of the front glass element.
Quite recently I read a cultural study where the differences in viewing between the Anglo-Saxon and the Japanese world were discussed. This had nothing to do with photography by the way, but focussed on movie art and drawing art. The writer claimed that the Japanese eye is automatically drawn to the background and from there looks at the centre of the image. The Japanese approach is one of holism, where the whole defines and conditions the subject. The Anglo-Saxon eye is narrowly focussed on the main subject that defines the scene and where the background is used to complement this definition. This may indeed explain why the topic of bo-ke is more important in Japan than in the western world.
The practical performance
Biogon-T 2.8/25 ZM
The Fuji Velvia 100F slides show very clean colors with a great amount of subject detail and subtly defined hues.
Wide open or stopped down does not make a visual difference in the recording capabilities of the lens, but stopping down does reduce internal reflections and gives very fine detail just a bit more contrast to rise above the grain pattern. There is hardly any distortion to be seen. At wider apertures you can notice some light fall off in the corners. Internal reflections, even when shooting straight into the sun are very well controlled and slides keep the deep black shadows, often becoming darkly greyed by unwanted stray light. Some moon shaped secondary images can be detected in adverse conditions, but being at the edge of the image, could be caused by reflections from the chrome front ring.
A fine test is always to shoot tree branches with the sun directly behind them, to see whether the light is spilling over the edges and will grey out the tiny branches. In this test the Biogon did extremely well, but at wider apertures the lower contrast was quite visible. Close up performance is very good, and at close distances the contrast reduction was quite modest, delivering crisp detail. On the other hand I could notice some shadowing at the edges of the frame.
Background drawing is pleasant with good retention of outline shapes, but quick blurring of details, giving the eye a rest from absorbing detail information.
Specular highlights are well recorded without star like reflections or haloing, but overall the images do not have the brilliance of the current Leica lenses.
Biogon-T 2.8/28 ZM
Zeiss and the focal length of 28mm have a strange relationship. The original Contax series had a 28mm lens, but with the Contarex line they dropped the 28mm in favour of the 25mm and added an 18mm, but with the Contax RTS and the G series, the 28mm became a prominent lens in the range.
At medium distances the lens captures fine detail with crisp delineation of the subject outlines and neutral colours that are on the edge of becoming vivid. The lens draws almost distortion-free. With backlighting, there is visible veiling glare, especially when the pictures are a bit over exposed. Correctly exposed we see clean solid black shadows and excellent retention of high lights. Specular highlights on the other hand show strong but localized haloing.
Close up performance is very good with crisp definition of very fine detail till the edges without any darkening at the edges of the frame. The smooth gradation into the unsharpness zone adds to the three dimensional illusion of the image. Background blur is a bit restless in its reproduction, but the light spots are smooth and round, thanks to the shape of the diaphragm.
Biogon-T 2/35 ZM
Wide open the lens records fine detail with smooth delineation of subject outlines form centre to corner. There is hardly any vignetting and stray light is well suppressed. Background unsharpness washes away all detail and holds the major subject outlines, but these are quite blurred. Stopping down to smaller apertures you get outstandingly good imagery with very crisp images and excellent definition of small detail that holds even in specular highlights, giving the images a powerful expression.
Colors are neutral and hold subtle hues quite well.
At closer distances the lens records fine detail with good contrast, but more with finesse than with crispness.
In contre-jour lighting, the Biogon 35 retains the definition of detail in the shadow areas and suppresses halo-ing around highlights very well. Only wide open the edges of tree branches become grey-ish.
Planar-T 2/50 ZM
Wide open the lens shows excellent neutrality of colours with amazingly good retention of fine colour hues. Very fine detail is recorded with good clarity, but with less crispness than the Leica counterpart. It shares with that lens the weak suppression of secondary reflections, due to the reflections at the edges of the rear mount. The background blur is on the harsh side.
The transition from the sharpness plane to the unsharpness regions however is quite long, giving a fine impression of depth and extension. The lens is especially good at recording detail in extended shadow zones, when you take pictures at dusk or at night.
The background blur shows the major outlines of the subject shapes, more sketching than drawing so to speak. Close up performance is excellent from centre to edge without any vignetting and distortion.
The Planar wide open is a potent performer and at smaller apertures becomes a master at reproducing with a life-like three dimensionality, that was the hallmark of the G-version of the Planar too.
Distagon-T 2.8/15 ZM
This lens needs some de-learning as there is no focusing aid. You look through the additional viewfinder and then have to guess the distance and set it manually. Of course you will often forget to do this, but then the great depth of field will cover up with forgiveness your lack of attention.
The lens records the scene with a good overall contrast. Contre-jour pictures show excellently defined subject outlines of quite fine detail and black shadows. There is quite visible vignetting wide open and for wide screen landscapes, one should stop down. Colours are vivid, slightly more so than with the other ZM lenses. The lens is also a bit more prone to internal reflections and veiling glare, but given the wide angle and big front glass element, the results are excellent.
Distortion is very low and when the lens has been held level you will hardly see any falling lines at the edges. Depth of field is impressive and you can indeed use the lens as an artistic tool to give the impression of great depth and space. On the other hand one can record quite fine detail and use the lens for accurate depiction of scenes and buildings. With good composition and exploiting the large view of the lens, one can create pictures where the viewer is literally drawn into the image. The viewer really stands with his feet at the edge of the picture and can step into the scene.
Any 15mm lens will show a certain amount of horizontally drawn out subjects and objects or persona at the edge of the frame will become slightly flatter and drawn out than they really are.
This lens is a welcome addition to the M lens world.
With the Contarex line of lenses, Zeiss wanted to create the Hasselblad-like system in 35mm format. The lenses were carefully colour corrected and were evidently designed to emulate the pictorial smoothness of the medium format imagery. This design philosophy clashed with the then popular acutance era in lens design and film emulsion concepts. And Zeiss revised their views with the Contax RTS lenses that followed the contrast paradigm of the day.
But with these lenses Zeiss did not overshoot their aim for a homogeneous and balanced quality and they sacrificed some wide-open performance to improve aspects as distortion and even coverage over the whole frame.
The new ZM lenses stay into this tradition and the main overall impression when looking at the slides is an association with medium format imagery.
The slides do possess that fine quality of smooth tonal resolution that is the characteristic of large format pictures. We can define the quality of images over two dimensions, the spatial and tonal resolution. Leica is evidently the champion of the spatial resolution and the elimination of aberrations at all cost, with a very accurate definition and very crisp drawing that extends to the limit of the modern emulsion technology (or capture technology to include the sensors of digital cameras).
Zeiss favors a type of tonal resolution that brings rich colours and a smooth gradation over the whole image, not only from corner to corner, but also into the image from foreground to background.
Of course the differences are not a simple or clear-cut as described here. Spatial and tonal resolution are two sides of the same coin and can not be separated as two competing dimensions. If you have good spatial resolution, then tonal resolution is good too. But you can shift the balance and the relative weighting of the two. We are discussing lens lines that are quite capable of excellent imagery, but with a different design approach, that does become visible in practical photography.
The Leica lenses in general at all apertures create images with more sparkle and detail coverage than the Zeiss lenses, again as a general statement. The Zeiss lenses are quite good at the spatial resolution and sometimes reach the diffraction limit at medium apertures, but Zeiss will consider this aspect quite simply as the result of good optical design, that automatically flows from the optimization of the other aspects, like distortion, even coverage and good performance over the whole range of distances.
One cannot say that the differences of lenses disappear when you stop down to smaller apertures, as is often claimed. The lenses described here do show differences even at smaller apertures, but these are of the qualitative nature described above.