• Ricoh GR 1:3,5/21mm, Voigtlander Skopar 1:4,0/25mm, Konica Hexanon 1:2.4/50mm

• 28mm Ricoh GR and the Elmarit-M and R versions of the 28mm.

• Voigtlander (Cosina) lenses for LTM mount (15mm to 75mm)

• Voigtlander Nokton 1:1.2/35mm

• Voigtlander (Cosina) lenses Ultron 1.9/28mm and Apo-Lanthar 3.5/90mm

• Voigtlander (Cosina) lens Heliar 1:3.5/50mm & Bessa T

A general introduction and an historical review might be interesting.

In the thirties as we all know, the two firms competing for the optical Olympus were Zeiss and Leitz. They used genius, elaborate manual design procedures, and exquisite mechanical mounts to produce the first golden age of lens design.
In the fifties and sixties, Canon and Nikon joined the competition and the four produced the second golden age and in doing so reached a design and performance platform early in the seventies. Again, superior engineering, optical cunning and daring designs gave us lenses whose performance is still excellent.



Then with the plateau reached two trends combined: optical design became automated and the independent lens manufacturers produced low cost alternatives to the prestigeous firms.
The reactions were different. Canon and Nikon followed a strategy to produce two series of lenses,one lower cost, one high cost. Zeiss and Leica did not want to go that direction, and insisted on building only lenses as good as possibly could be (in their view at least). Contax tried to blend the Yashica and Zeiss lenses as a low cost and high quality pair, but that failed. Contax users wanted simply the best AKA Zeiss. So Zeiss shifted most of the lens production to Asia to lower the cost and keep quality to the required levels..
Leica does not have this alternative, so their strategy is to design and built the best lenses ever at whatever cost. But there is a limit to image quality to be perceived and appreciated, as the discussions on this list amply document.
In a positive way then the new Cosina and Konica and Ricoh lenses can be interpreted as an independent offering of low(er) cost alternatives to the Leica lenses (in some cases niche filling as the 2,5/75 shows). This bodes well for Leica. More people will buy their bodies and eventually will have a mix of Leica and other lenses.
The important fact is that the rangefinder camera (once pronounced dead) gets ever more life injections. And when people start to appreciate the imagery possible with Leica lenses, they will buy into this lens line.
General remarks on the optical quality:
All three lenses to be reviewed offer surprisingly good image quality, some even outstandingly so.
The designers cearly had good examples to work from and could stand on the shoulders of many generations of good designers. And the very good opticaldesign programs of today also are a bonus. Generally these lenses perform as second generation Leica lenses from the period 1960 to 1975.
Being of lower cost there are of course minor and sometimes major deficiencies. Decentring is a major one, as is flare, particularly spillover of light at black/white bounderies of very small object details. If you use elaborate testing equipment,you will note overall a higher aberration content than in the equivalent Leica lenses. The performance in the field, that is the outer zones of the lens, is the weak point here. Also the mechanical mounts are not machined or assembled as well as the Leica ones, but then here Leica is worldleader.
For many picture taking situations and moderate demands these Japanese lenses are very good indeed and are worthy of close scrutinity by interested buyers/users.
Ricoh GR 3,5/21 mm
Closely modelled on the historical Super-Angulons, this lens at full (3,5) aperture gives a low to medium contrast image with a quite strong light falloff, starting half way the middle of the diamater. There is a slight distortion in the corners.



This lens exhibits a discernable bit of decentring and at this aperture also is flare prone. Especially at the edges of strong pointlight sources you can see halo-like degradation. On axis very fine detail is clearly recorded, with extremely fine detail detectable. In the field the image details become progressively less sharply delineated when going to the edges where fine detail is very softly rendered , but still visible.
Over the whole inage area very fine detail is rendered with good visibility. Stooping down a fraction to f/4 crispens the outlines a bit and at f/5,6 we have excellent imagery over the whole image field with very fine detail crisply rendered and a modicum of exceedingly fine detail becoming detectable.
At f/11 we note a visible reduction in contrast. The very good performance holds till 1 meter.
This lens must be rated above the (older) Super-Angulons for the M series and in itself is an excellent lens.
In comparison the Elmarit-M 2.8/21 ASPH at full aperture presents a high contrast image with less light falloff, but a fraction more distortion. Decentring cannot be detected and the overall aberration content is much lower. On axis we see exceedingly fine detail, that naturally is less well recorded in the field (zonal) part of the lens area. Still very fine detail is very crisply recorded and extremely fine detail clearly detaectable.
At f/3.5 the image crispens noticeably and now exceedingly fine detail is very well resolved over an image area with a diameter of 9mm.At f/4 this performance extends over the whole field, but the corners are still soft.Comparing this performance to the GR, it is clear that the Leica lens is in a different legue. For several purposes the GR might be a serious alternative. If you only need sometimes a very wide angle, the Gr might be the answer.
Skopar 4/25.
This lens does not couple to the rangefinder and has a few fixed distance tabs in the distance scale.



It shows decentring and the same type of flare as noticed with the GR 21. Distortion is very low, but light fall off is visible, but not disturbingly so. At full aperture the overall contrast is high and very fine detail is crisply rendered over an image area with a diameter of about 6mm. Still very fine detail stays visible over most fof the image area and becomes barely detectable in te corners. Fine detail over the whole image area is crisply rendered. This excellent performance holds till f/8, and stopping down further reduces contrast. Close-up performance is as good as the infinity setting
In comparison the Elmarit 2,8/24 shows a high contrast image with exceedingly fine detail rendered very crisply over an image area with a diameter of 12 mm. Corners are soft with a slight flare prone characteristic. After f/4 no improvements.
The Snapshot-Skopar is an excellent lens and a tempting proposition. For its depth of field the issue of not being coupled to the rangefinder is not as bad as it may sound.
Hexanon 2,4/50.
This collapsible lens couples with the M rangefinder. Collimator checks show an excellent performance with accurate coupling. 
This lens has a bit loose tube, which flaws its performance. My comments are based on the correct positioning of the tube. At 2,4 a high contrast overall image combines with a crisp rendition of quite fine detail over most of the image area. At f/2,8 the image improves very much an now exceedingly fine detail is rendered clearly on axis and in the field. Corners stay weak as is natural. Stopping down improves a bit, but the excellent performance holds till about f/8.
In comparison the Elmar-M 2.8/50 at full aperture is a shade softer and extremely fine detail is clearly rendered.The outer zones however are visibly softer. Stopping down the Elmar becomes the equal of the Hexanon, but at 2.8 the Hexanon is indisputably the winner. If its mechanical parts were improved. Here we have a lens that challenges the classical and redesigned Tessar design.
Conclusion 1
I shot comparison images on the same film (lens by lens) (my bayonet is the first item to exchange because of usage) and noted that the Ricoh, Skopar lenses exhibit second generation characteristic. Overall the images are duller and a bit muddier than the current Leica lenses. Still the Ricoh is better than the S-Angulons and the Skopar not having a predecessor can be compared only to the Elmarit-ASPH 24.
A most interesting phenomenon became visible with these side-by-side-shots. The Ricoh and Skopar give images with a grainier pattern and with grain clumps that are rougher than when the Leica lenses are used. This is caused by the lower aberration content of the Ricoh/Skopar lenses. When aberrations are abundant the light rays emanating from a light point source do not converge to one point in the image but have a more random pattern around the central core. These more widely spread rays energise more silver grainaround the center spot and randomly so. A rough clumping is the result.
Modern Leica lenses produce a smooth pattern of very tighly contained grainclumps, which helps preserve the rendition of very fine detail and the smooth gradation of fine light modulations.
Leica lenses exhibit a crisp clarity of the finest possible detail, that the Japanese contenders can not match.
Conclusion 2
On its own the Ricoh and Skopar lenses are excellent value and especially the Skopar at f/4 is on axis a strong performer. The outer zones are no match for the Leica (at f/2,8!!) but still it is a tempting lens.
The Hexanon 2,4/50 is a 6 element lens and a complicated proposition. Any lens with a collapsible mount is difficult to manufacture and the mechanical play in the lens I used may not be typical. I made my comments based on the optimum position and assuming that quality control will tighten up, it has the potential to be one of the first contemporary lenses to be a threat to a Leica lens. Its optimum performance at f/2,8 is simply better than the Elmar-M 2.8/50 and even at its worst setting would kill the older version of the Elmar.
The excellent performance of the Ricoh and Skopar are partly the result of the modest aperture. The higher aberration content will not be visibly buried as this will be among other things in the depth of field.
The generally weaker performance in the field is also a characteristic that distinguishes these lenses from the Leica lenses.

28mm Ricoh GR and the Elmarit-M and R versions of the 28mm.

The Ricoh (#02818) is a diminuative lens. The chrome version with the focusing aid is very Leica like. Mechanically it showed a slight axial play in the distance ring mechanism and a very slight decentring.
At full aperture (2.8) the lens shows a medium to high contrast, with some light fall off in the corners. In the center very fine detail is recorded with medium microcontrast and slightly soft edges. Extremely fine detail is just recorded with very soft outlines. This type of performance holds on axis over a circle with a 4 mm radius (an area of 8 to 10mm diameter). In the field (the outer zones) fine detail is recorded with high contrast. The edges however show colour finges and there is also a trace of astigmatism. Fine and very fine detail is fuzzy, but just visible. The corners are very soft with outlines of large subject detail clearly visible.
At f/4,0 the image crispens a bit and the overall contrast becomes quite high. On axis the edges of fine detail clearly sharpen,. In the field the improvement is slight. At f/5.6 the overall image is still a bit soft in the field. At f/8.0 we get an excellent image with extremely fine detail clearly resolved, with a faint trace of softness. At f/16.0 the performance drops.
The close-up performance ( 1 meter) is identical to the infinity setting.
The Elmarit-M 2.8/28mm (#3793205) at full aperture shows a high contrast image with very fine detail crisply rendered. On axis over an image area 0f 10 to 12mm diameter) the extremely fine datail is very clearly visible with some softness at the edges. In the field (outer zones) the contrast drops, and the outlines of large objects show a very faint colour fringe. Extremely fine detail is recorded with good clarity, with a slight fuzzyness at the edges.
This performance holds into the corners where extremely fine detail is still visible, but with fuzzy edges.
At f/4,0 the image crispens in the detail rendition. Extremely fine detail is now clearly resolved with great clarity and sharp edges and exceedingly fine detail is now crisply resolved on axis, with some fuzzyness in the far corners.
At 5.6 the outer zones improve a bit, but the center is already past its optimum.
Close-up performance is equal to the infinity setting. No decentring could be detected.
The Elmarit-R 2,8/28mm (#3701134) is almost identical in performance to the M version. The R version exhibits slightly more chromatic aberrations and the performance in the field is a bit lower. Generally however this lens is equal to the M version. Close-up performance till 30cm gives excellent imagery comparable to the infinity setting, thanks to a floating element Again no decentring could be detected.
Conclusion:
The use of aspherics (two surfaces ) in the Ricoh is not automatically a ‘free’ ticket to best perfomances. We have been evaluating here absolute performance and on this quite lofty Olympic heights the Ricoh is clearly not as good as both Elmarits, which deliver superior imagery. To stay for a while in this comparative field. The Ricoh is better (by a good margin) than the first and second generation of the Elmarit 28mm M lenses. The third generation of the Elmarit however shows better performance in all areas, as of course does the current (fourth ) generation. Absolutely speaking: the Ricoh at 5.6 and 8.0 is almost as good (but not quite) as the Elmarit ‘s at 2,8.
(There is one area where the Ricoh has an advantage: the mechanics of the separate viewfinder. It is a solid chrome metal piece of work. Optically the finder distorts quite a bit: that mars the joy a bit. )
But we can also look at the performance from a slightly more casual perspective. If we assume that the Ricoh and Elmarit pictures will be taken on colourneg films or higher speed B&W films and enlarged to let us say the ubiquitous 20x25cm format, then the performance differences shrink. Of course on close scrutinization the overall softer rendering of the Ricoh will show. In the center area however the margin between the Ricoh and the Leica lenses is under these conditions is not that big.Stopped down to 5,6 or 8.0 you would be hard pressed to notice any difference. As said before, the Ricoh lens is visibly softer and that will be discernable at all times. You need a comparison picture to notice it though.
In close comparison the clarity of the Leica lenses is obvious as is the much higher micro contarst at edges of fine detail. The rendition of the fine textural details which are etched out in the case of the Elmarits and hardly visible in the case of the Ricoh, gives the Leica lenses their reputation.

Voigtlander (Cosina) lenses for LTM mount (15 to 75)

Super-Wide-Heliar 4.5/15.




At full aperture this lens has a medium overall contrast. Distortion is surprisingly low, only a bit barrel distortion, but some severe light fall off. But for a 15mm that is acceptable. On axis we see a crisp definition of very fine detail over an image circle of 6mm diameter. Beyond that image quality drops fairly rapidly. At an image location, 6 mm from center, very fine detail has quite soft edges, but they be detected without difficulty. Going from here to the corners the very fine detail becomes quite fuzzy, but stays within visible range. Fine detail is rendered with good visibility. The detail rendition is a bit dull. There is some serious flare that lowers contrast. Astigmatism and curvature of field are well corrected. In the field however the chromatic errors (longitudinal) are quite visible as color fringes around edges. There is a trace of decentring. Corner and edge performance is good.
Stopping down to 5.6 crispens the whole image a bit and fine detail is recorded now with good edge contrast. The overall improvement is slight but noticeable and now very fine detail is recorded over most of the image field. This aperture is the optimum one. Close-up performance (1 meter) is excellent and about equal to the infinity setting. At 1:8 image quality drops, as contrast takes a dip. Vignetting is now almost invisible. At 1:11 the contrast drops rapidly, reducing the image quality to barely useable. There is some reason why Zeiss stops at f/8!!
This is a very good, if not excellent lens, with a high level recording capacity on axis and a good one in the field. The details are a bit muddy and miss the clarity that characterizes the current Leica lenses.
Snapshot-Skopar 4/25.



This lens does not couple to the rangefinder and has a few fixed distance tabs in the distance scale.
It shows a fair amount of decentring. In the outer zonal areas and at the edges strong pointlight sources produce halo-like degradation. Distortion is very low, but light fall off is visible, but not disturbingly so. At full aperture the overall contrast is high and very fine detail is crisply rendered over an image area with a diameter of about 6mm. Fine detail stays visible over most of the image area and becomes barely detectable in te corners. This excellent performance holds till f/8, but with reduced contrast. on axis. Close-up performance is as good as the infinity setting
In comparison the Elmarit 2,8/24 shows a high contrast image with exceedingly fine detail rendered very crisply over an image area with a diameter of 12 mm. Corners are soft with a slight flare prone characteristic. After 1:4.0 no improvements can be detected.
I shot comparison images on the same film (lens by lens) (my bayonet is the first item to exchange because of usage) and noted that the Skopar lens exhibits most characteristics of the Leica lenses from the earlier generations. Overall the images are duller and a bit muddier than the current Leica lenses.
A most interesting phenomenon became visible with these side-by-side-shots. The Skopar gives images with a grainier pattern and with grain clumps that are rougher than when the Leica lenses are used. This is caused by the higher aberration content of the Skopar lens. When aberrations are abundant the light rays emanating from a light point source do not converge to one point in the image but have a more random pattern around the central core. These more widely spread rays energise more silver grainaround the center spot and randomly so. A rough clumping is the result.
Modern Leica lenses produce a smooth pattern of very tighly contained grainclumps, which helps preserve the rendition of very fine detail and the smooth gradation of fine light modulations.
Leica lenses exhibit a crisp clarity of the finest possible detail, that the third party designs can not match.
On its own the Skopar lens is excellent value and the Skopar at 1:4 is on axis a capable performer. The outer zones are no match for the Leica (at 1:2,8!) though.
The fine performance of the Skopar is partly the result of the modest aperture. The higher aberration content will not be detectible in many picture taking situations, hiding as it were behind among other things the depth of field.
The generally weaker performance in the field is also a characteristic that distinguishes these lenses from the Leica lenses.
The Nokton 1.5/50mm



If you look at the development and evolution of high speed 50mm lenses, you will note that lenses with an aperture of 1:1.4 leaped forward optically in the mid-sixties and then reached a plateau. Almost every 1.4 design is based on the Double Gauss formula and generally has 7 lens elements. One of the last ‘new’ 1.4 lenses was the Planar 1.4 for the Zeiss Contax in the seventies and its performance was not breathtaking to put it mildly. It is very difficult to design a 1.4 lens with a performance equal to the best 2/50mm lenses. Oblique rays in the sagital plane play havock with all good intentions of the designer. Do not forget that a 1.4 design admits twice the amount of light energy and aberrations grow disproportionally. Many of the specific 1.4 aberrations errors do not improve when stopping down, making a 1.4 more of a compromise than a 2/50, which is a much more evolved type of lens design. So for decades designers had no serious option to improve on a 1.4 design (not possible or to expensive). Recently Leica introduced a redesigned Summilux-R 1:1.4/50 which is a big step forward to produce f/2 quality a 1,4 design. At 1,4 the Summilux-R offers a high contrast image with outstanding imagery on axis and a very good quality in the outer zones. The design utilizes the classical Double Gauss formula stretched to 8 lens elements.
The Voigtlander Nokton 1.5/50 is the most recent addition to the 1,4 club.
Its design is closely modelled to a ‘normal’ 2/50 design with 6 elements, but with the last element a double aspherical one, that is both surfaces are aspherical. It is a rule that one aspherical surface replaces a full lens element in a full spherical design, so this lens can be compared to the 8 elements of the Summilux-R.
The use of an aspherical surface is not restricted to the correction of spherical aberration, as it is sometimes stated. Many more optical issues can be addressed with an asphere. In fact the Nokton has more residual spherical aberration than could be expected.
My version had some decentring which became visible in testing of course (one outer side of the image gave reduced contrast and a drop in performance. It could also be seen in real life pictures when large billboards were photographed. One side visibly gave a lower definition and loss of detail rendition. Not that much but still.
On test the full aperture gives a medium contrast overall image with a trace of veiling glare. On axis very fine detail was rendered with very good edge definition. This definition extended to the outer zones with only a slight loss of contrast and edge definition. This lens excells at a very fine even coverage over the bigger part of the picture area (neglecting the decentring for a moment, which might be atypical). Astigmatism is well but not fully controlled and lateral chromatic aberrations are very well reduced. You will see these effects in the outer zones as a reduced rendition of very fine and extremely fine detail and a high noise ratio which makes it difficult to detect separate details in the finer structures. There is visible vignetting in the corners.
At f/2 the image becomes more contrasty and especially the edge definition improves in the field.
But it is not up to the quality you expect from a topclass f/2 design. In itself the image quality is very good. Very fine detail is now detectable with slightly soft edge definition. At f/2.8 the contrast again improves and now extremely fine detail can be recorded. At f/4 we find an excellent quality over the whole picture area. Extremely fine detail is now rendered with good clarity. A very critical look will reveal that the overall contrast and micro contrast are lower than the contemporary Leica designs and some veiling glare also reduces the recording ability of the finest details. There is a focus shift, reducing the contrast on axis. Tangentially the aberrations grow quite a bit in the field.
From an aperture of f/5.6 quality drops, especially in the field where detail gets a bit fuzzy.
Close-up performance is not as good as at infinity, Th whole image is a bit softer now and you need to stop down to 2,8 or 4 to get optimum imagery at this distance.
The Nokton is quite sensitive to flare and here you need some caution.
Conclusion.
This is a difficult lens to evaluate and it shows the classical dilemma of any tester. Figures and MTF graphs are not enough. Starratings wil not show the subtle differences and a cursorary report will fail to draw attention to tye character.
The basic optical design of this lens is outstanding and it will certainly be studied by several optical departments over the world.
Mechanically and from an engineering standpoint it gives mixed feelings.The decentring is an indication of mounting tolerances and the economics of manufacturing. You get what you pay for.
The image quality in most practical situations is impressive. Flare is quite pronounced in backlightning and when recording specular highlights or small lightpoints.
The Summilux-M as comparison has at full aperture higher contrast on axis but its performance in the outer zones is not as good and the recording ability of very fine detail over the picture area is also not as good. But its engineering is superb and its flare reduction is also better. So which lens is best?
My list would be. Number 1 is the Summilux-R new, the number 2 with a fair gap is the Nokton and the number 3 is the Summilux-M, which is better engineered but optically not as good.
Do you really see these optical differences? Thats the 10.000 dollar question,
Whatever the vanishing breed of people dedicated to resolution numbers may declare, most picture taking situations can be covered with 10 to 20 lp/mm for exhibition quality and in this area the Nokton performs admirably well. It does not have the clarity of fine detail of the better Leica lenses, nor the clean edge definition of larger subject outlines. In fact it is a bit dull in the shadows and fine detail has a coarser image and a fuzzier edge. But face to face with the Summilux-M the Nokton wins on points, not by knock out.
The Ultron 1.7/35 mm aspherical



A complex design with 8 elements, a negative front curvature and one aspherical curvature.
One half stop in between a 1,4 and a 2 design, this lens has the outward appearance of a older Leitz lens, the Tele-Elmarit from the sixties. At full aperure it gives a medium contrast image with a tendancy to flare. Very fine detail is recorded on axis with clean edges and in the field performance drops a bit and now fine detail is recorded with fuzzier edges. In the outer zones some astigmatism and color fringes lower the rendition of the finer detail and outlines are fuzzy. The far corners are very soft and of low contrast. Outlines are visible and some detail detectable.
Some decentring is noticeable in the outer zones only. At f/2 the contrast naturally improves but not very visible, At f/2.8 we see a noticeable improvement in contrast and especially the aberrations in the field are reduced to a minimum. Extremely fine detail is now clearly detectable over most of the image field (corners still in worse shape) and at f/4 this level of detail rendition now is recorded with clear visibility. Edges stay on the soft side though. At f/5.6 the edges crispen and now we have the optimum with excellent recording of the critical 10 to 20 lp/mm . Close up at full aperture gives a faintly soft image and here stopping down to 4 is enough for really critical work.
This lens is clearly not ont he same level as the Summicron /35 asph. It is a match for the non-asph Summicron 2/35 which has however a small edge in contrast and field performance.
The Color-Heliar 2.5/75mm.



A classical 6 element Double Gauss design with 6 elements in 5 groups. It looks like an older Tele-Elmarit 90.
At full aperture we have a high contrast image with outstanding rendition of extremely fine detail on axis and a visible drop of quality in the outer zones where contrast drops, and outlines of fine detail become fuzier. Corners are very soft and of low contrast. At 2.8 you do not note any improvement and at 4 contrast improves. As with the Ultron the color fringes at the edges of outlines and finer detail structures do not disappear until 5.6, which is the optimum with an overall imagery of very high order; extremely fine detail is rendered crisply over the whole image field excepting the outermost corners.
Close up performance is Ok and on the same level as at infinity.
This general performance is comparable to the older Tele-Elmarit 2,8/90 and a class behind the current 2.8/90.
I also compared the Summilux 1.4/75, stopped down to 2,4 . Here the Summilux performs at ist personal optimum and we see exceedingly fine detail crisply rendered with high edge definition and contrast ove rthe whole image field. The Color-Heliar is a strong performer in itself, but the 2.5 performance is below the level defined by the Summilux-M.
Perspective.
I have now tested many of the Voigtlander lenses and all show some interesting family characteristics. All have decentring, sometimes degrading the image quality more than should be tolerable for high quality imagery. All have flare, pointing to a coating technique that could be improved. And all have a somewhat dull, flat rendition of details and outlines. This points to glass selection and a certain choice of aberration correction.
On the positive side I note an excellent sometimes outstanding behavior on axis, and a very good to excellent performance in the field. (when negelcting the decentring problem).
All give better imagery than first class lenses 10 years ago and clearly show the direction of the Cosina designers: Astounding value for the money. I personally dislike the borrowing of the outward design of the Zeiss Contax G and the older Leica lenses. But if you have to choose between optical design and industrial design when you are on a tight budget, this is a sound choice.

Voigtlander Nokton 1:1.2/35mm

Introduction
In the sixties and seventies of the previous century, the great camera companies competed with each other in the area of the fastest lenses. The aperture of 1:1.4 was common and several focal lengths could be bought with a 1.2 aperture. The 55mm and the 85mm were often chosen as the relatively narrow field of view assisted the optical designer in creating a lens with a good performance. With the advent of high speed-high quality films and in addition the ubiquitous zoomlens, that limited apertures to a common 1:2.8, the race for the fastest lens alive died.
We may rationally ask why a 1.2 lens has any added value when compared to a 1.4 deign, apart from the emotional arguments.
The jump form 1.4 to 1.2 is at best a half stop. I say ‘at best’ because of the tolerances involved. A lens with a designation of 1.4 or 1.2 (or any other aperture) may depart 5% from the nominal designation. Quite often a 2.8 lens is in reality a 2.9 (a 3% change) and if we add the normal production tolerance of 5 to 10% from the nominal value, we may end with a lens that has a practical aperture of 3.0. Current film material is very good and can handle some under exposure and in any case our eyes will not detect a difference in density that is below 10%. To be precise: our eyes can detect changes in density as low as 2%, but only when we compare adjacent areas with different densities. Consider this: the maximum black in a high quality print has a reflection density of D-2.25. But most people have already trouble seeing a difference in shadow density from D-1.95 to D-2.25. This is a difference of more than 10%. It is a not-unreasonable assumption that the need for a 1.2 aperture is justified by the argument that at the limits of low light recording, every additional photon could make the day. In this situation a 1.4 lens would generate underexposure in the critical shadow densities. And the extra half stop would allow the film to build up enough density to a level that is useful for recording differences in shadows.
This borderline of photographic recording is the stuff dreams are made of.
And the Leitz Noctilux 1:1.2/50mm fits in very well with this approach. My review of this lens can be found on my site too.
The Voigtlander Nokton 1:1.2/35mm Aspherical
Mr. Kobayashi is a person who relentlessly pushes his designers to accomplish the impossible in terms of price-performance-emotional value. The new Nokton 1:1.2/35mm is a perfect example. Borrowing the prestigious word ‘Aspherical’ from the first Leica Summilux 1.4/35mm with two aspherical surfaces and adding one aspherical surface and a half stop in speed, the Nokton 1.2/35mm is targeted at that lens. The Nokton 1.2/35mm has the large scalloped focusing ring, we know from the first Noctilux 1.2/50mm (also with two aspherical surfaces). With 10 elements and three aspherical surfaces, this lens must be one of the most complex designs in moderate wide angle domain.
Compared to the Voigtlander 1.7/35mm ( a full stop slower) and the Leica Summilux 1.4/35mm ( a half stop slower) the Nokton is a very large and heavy lens. The physical dimensions indicate the price you have to pay for a half stop additional speed. The lens itself is very well made, with smooth moving components, which could be the envy of the Leica manufacture. On the plus side we note the absence of a screw in the focusing ring, an indication of clever assembly. On the minus side we note a damaged screw in the bayonet mount ring.
To stay with manufacturing quality, I could detect only a small decentring of 0.01mm.
Optical performance
This is what it is about. Let me first analyse the lens with my customary method. At full aperture (1:1.2) overall contrast is quite low, in fact lower than I have seen in the last ten years. Some people see this as an advantage in high contrast situations, but in my view this is not the case, as the critical shadow densities will be pushed below the visual detection level. There is a substantial amount of spherical aberration and a high level of chromatic errors. The edges of small detail are very soft, as there is a very visible amount of colour fringing, even at the 10 lp/mm frequency. The resolving power is quite high with above 100 lp/mm in the centre area of the image, going to 20 lp/mm in the corners. But the micro contrast is very low and we can detect a substantial amount of non-imaging forming light that reduces contrast visibly. On the plus side we note a low level of coma and a low level of vignetting (actually below the amount we know from the Noctilux 1.2 and 1.0). Pin cushion distortion is quite visible too. Astigmatism is well controlled. Field curvature is acceptable with 0.08mm. The lens has at this aperture a remarkably low propensity to glaring flare and secondary reflections.
Stopping down to 1.4 hardly brings any improvement and even at 1:2 we see the same performance as at 1.2.
At 1:2.8 the overall image starts to crispen and the colour fringes are reduced, improving micro-contrast. From 1:4 we get really food performance with 150 lp/mm in the centre area, 100 lp/mm in the outer zones and the edges still on the low side.
From 1:4 we see the usual improvements in contrast and covering power over the image area. At 1:8 the optimum is reached, and at 1:22 the overall definition is soft again. Compared to the Voigtlander 1.7/35 and the Summicron 2/35 asph, the performance at all apertures is not at the same level.
If we would buy the Nokton 1.2/35mm it is for the performance at the wider apertures, and not for the middle apertures. We need to concentrate on the performance at stops from 1:1.2 to 1:4 to get an impression of the lens quality.
From the description one gets the impression that the lens wide open is not good and that is indeed the case. Compared to a good 1.4/35 (Canon or Leica) the drop in quality is quite big, related to the speed gain of at most a half stop and in practice close to a third stop. You pay a substantial price at the wider apertures for this speed gain. Especially the 1:1.4 and 1:2 apertures suffered from the need to balance the aberrations at the maximum aperture. Remarkably good are the vignetting and the flare control. But overall and micro contrast are very low, delivering a rendition that lacks body and sparkle.
Another view
We use the Nokton mostly in handheld situations with high speed film and low level ambient light. Here we need only to record the main outlines of the subjects and sufficient detail to add depth to the picture. At lower shutter speeds it is illusory to record fine structures. So we did another test to look specifically at “street” values of 10 and 20 lp/mm. From 1:1.2 to 1:2 the image quality is almost the same: At 10lp/mm the overall image in the centre area brings a good edge contrast and a low overall contrast; in the outer zones the lines are rendered quite soft and at the edges are just visible. At 20 lp/mm the centre performance is identical, but the outer zones are now very soft, but clearly visible. At 40 lp/mm the outer zones are just visible.
At 1:4 and 1:5.6 the critical 20 to 30 lp/mm are very well resolved and rendered with good contrast over most of the image area.
From the standpoint of a practical photographer, working in the situation described above, the performance at the wider apertures is quite acceptable and quite uniform over most of the image area. When using films like Tri-X, its high acutance and good edge sharpness can compensate the softness of the Nokton edge detail, With slide film and slower speed bw-film however, the definition of the Nokton is not up to the current state of the art. In most high speed lenses, we see a substantial improvement when stopping down a bit. With the Nokton this improvement is postponed to 1:2.8. We may conclude that the demands to get acceptable quality at the widest aperture, demanded a compromise that lowered the performance at 1:1.4 and 1:2 to a high degree.
Conclusion
The lens is a complex design with ten elements and three aspherical surfaces and from these specifications we may infer that the image quality might be at least as good as the performance we expect from more normal designs. The additional half stop to 1:1.2 and the wide angle of view seem to impose such limits to the designers that the compromise they searched for is a bit elusive. While very good 1.2 designs were made in the seventies in the 50mm focal length (see my Canon report), the step to a 35mm focal length is, even today, a bridge too far. The Nokton 1.2/35mm Aspherical is a good performer at smaller apertures, but at the wider apertures, it is just acceptable.

Voigtlander (Cosina) lenses Ultron 1.9/28 and Apo-Lanthar 3.5/90

Apo-Lanthar 3.5/90mm.
Optically it is an easy lens to describe. At full aperture we have a high contrast image with clear definition of very fine detail over the whole image area from center to corner. Stopping down improves microcontrast a bit by cutting off internal reflections, That is all there is to say. Stops are used for depth of field control and not for image quality control. From 3,5 to 11 this lens delivers equal quality. There is no astigmatism, and a slight curvature of field. The apo designation is a bit optimistic, there is colour fringing in the outer zones. Some slight vignetting, but worth to think about it.
Mechanically this lens is excellent. There is no decentring and the collimator check gave true infinity at the intended postion. Collimation deviation is zero. The high quality of assembly is not accomapnied by the choice of materials: there is some play in the mount and the collimator changes when you press on the lens. Here we see the compromises.
The current Elmarit-M 90 at 2.8 is as good as the 3.5/90 at 3,5 and the Elmarit at 4 is ahead of the 3.5/90 in brilliance and clarity of rendition. This said the Voigtlander is much better than all but one previous Elmarit and Tele-Elmarit designs.
Compared to the Apo-Summicron 90 ASPH we note that the Voigtlander at 3.5 is not as good as the AA-Summicron at f/2. An amazing performance, when we remember that the Summicron has only 5 elements. Wide open the Summicron can be described as delivering very high contrast imagery with very crisp definition of exceedingly fine detail over the whole image area.
Given the price of the Apo-Lanthar 90, this lens is a better choice than most second hand 90;s from Leica on the market and is very close to the performance of the current 2.8/90, which is some act: the apo-lanthar is unbeatable in price/performance relation. Its performance can be improved upon and mechanically it may be of limited durability. As it stands it is optically a first class design.
The Ultron 1.9/28.
The difference between 1.9 and 2 can be neglected so I limit myself to the f/2 aperture. At full aperture the lens exhibits low to medium contrast and definition of very fine detail is soft but very even over the image field extending till the corners. The extreme corners are very soft. Even finer details are resolved in the center, but rapidly become softer when going to the zonal area. There is no astimatism and only a very small amount of curvature of field.
There is some barrel distortion, but quite small. At 2.8 the overall contrast improves and at f/4 the optimum is reached. After f/8 contrast lowers again and from there we see a softening of image details with every additional stop. This is usual behaviour. At f/4 we note a resolution over the whole image area of above 100 lp/mm.
As a design philosphy this lens is a bit tuned to high resolution versus high contrast.
Mechanically this lens is excellent: no decentring at all, collimation at zero. As with the 3.5/90 there is a definite amount of play in the focusing mount. Also the amount of dust in the lens is above what we are see in Leica lenses.
Still at wider apertures the lens is a shade soft and lacks the clarity and crispness of the Summicron-M 2/28. I shot pictures with Kodak E100VS and in low light and contre-jour situations and noted the low contrast compared to the Summicron. Specifically the tightness and sparkle of high light and shadow rendition was excellent with the Summicron and a bit dull and washed out with the Ultron 28.
These differences in rendition are quite visible in critical inspection and may or mey be not be not interesting for some users. Fact is they exist.
Where the Summicron 28 excels is the high contrast and the very crisp rendition of extremely fine detail, given the Summicron images a brillance and clarity that can be compared to the Ultron's duller representation.
Stopped down the differences are quite small.
The Leica Summicron 28 has the advantage of superior mechanical engineering and materials. It also sports a visibly better imagery at wider apertures. Wide open there is no contest: Leica is still the one to beat. If you do not see the difference, you are the cause! It as simple as that! And while the Ultron comes close, it just is not able to challenge at wide apertures. Stopped down it is a draw.
The Ultron 28 is one third of the price of the Leica and delivers excellent, and close to outstanding imagery. For heavy duty use the choice is simple: Leica. For somewhat less demanding situations the Ultron is a tempting alternative, if you can live with its lack lustre wide open performance. To get some perspective: an older 28mm lens from Leica is blown to pieces by the Ultron.
MTF measurements show that the Ultron at wide apertures delivers quality that is one step behind the Summicron: The 20 lp/mm of the Summicron have a contrast transfer of 90%. The Ultron has 76%. At 10 lp/mm the Ultron has 88%. So Voigtlander at 10 lp/mm is as good as Leica at 20 lp. That does show!
While the Summicron is still the one to beat, the Ultron delivers quality, that was unbelievable some years ago. Leica designers are challenged!!
Some general remarks may be inserted here.
Where lenses today differ is in the mechanics, and the engineering quality and the performance at the wide apertures. But lenses are at such a high level that the user and the choice of film and the handling of the imaging chain become more impartant than ever. The most important factor now is the user expertise and expectancy. The Ultron is an amazingly good lens, that will satisfy the demands and expectations of many a photographer. If you wish to work at the cutting edge or need to deliver simply the best, the Summicron is the answer.
The differences are clear. The choice is yours.

Voigtlander (Cosina) lens Heliar 1:3.5/50mm & Bessa T

 The year of celebration.
This year the Tessar lens by Rudolph is 100 years old and the Heliar by Harting is 101 years old. Both lenses are still current, a tribute to the master designers behind them. The Heliar should be the better lens, as it is more symmetrical than the Tessar. The Heliar lens has also been designed by Tronnier, Dallmeyer and Deser, to name a few. The Heliar in the past has better flatness of field and the Tessar better colour correction
The Bessa-T
I had the possibility to test the current Voigtlander Bessa-T with the Heliar 3.5/50 and compared it to the Elmar-M 2.8/50.
But first the camera. The register is off by +0.03. The previous version was spot on so this is an indication of the tolerance bandwidth. The + sign means that the film plane is 0.03mm deeper into the body relative to the bayonet flange.
The shutter (running from bottom to top!)test revealed this: The shutter (running from bottom to top!)test revealed this:

1	+12%
2	+ 6%
4	+10%
8	+3%
15	+10%
30	+15%
60	+20%
125	+15%
250	-1%
500	0%
1000	+20%
2000	+20%

 
This is good but not outstanding result. To be honest, most Leica shutters are in the same legue, with the exception of the M7, which is more accurate.
The lens Heliar 1:3.5/50mm
On the collimator (infinity setting) there was zero tolerance. Also decentring was zero: a first for Cosina. There is a trace of distortion.
Wide open at 3.5 the lens has medium to high contrast and delivers really outstanding performance. From center to the extreme corners the lens resolves around 100 linepairs/mm with good microcontrast.
There is no astigmatism to speak of and the corner performance is simply stunning for this type of lens. There is a trace of colour fringing, but far less than with higher speed designs. At 1 meter this quality is the same. Stopping down does not improve the performance, bar a trace of higher contrast, and from 4 to 16 we have a very even quality. The excellent performance from 3.5, not only in the center but especially in the outer zones, makes this lens a very strong competitor to higher speed designs and may evoke the question why we really want high speed lenses.
Mechanically there is something to complain about: the focusing movement was a bit rough, and the collapsible tube has some play. when not secured, but this is nitpicking. Optically the lens is oustandingly good in the class of standard focal lengths.
The Elmar-M 2.8/50mm
Again no decentring and high class mechanical design. At 2.8 the center performance is a trace better than that of the Heliar, but in the outer zones and corners the Heliar is better, not by much however. I doubt if many users would detect it, but it is real and measurable. At 4 the Elmar is at its top, and delivers the 100 linepairs/mm from center to the outer zones (excluding the corners where we read 77 lp/mm). Stopping down does not improve the image quality.
Based on resolution figures both lenses are comparable. The Heliar however scores with a better flatness of field and less astigmatism, which give the pictures a bit more sparkle. The Elmar has somewhat less colour fringing, but the edges of very fine detail are a bit softer.
For most applications both lenses would appear to be equal. But optically speaking the Heliar has the edge and delivers on all counts the better image quality, not by much, but it is there.
In the past it has often been said that a triplet design is best at 3.5 or 4 and the stretching of the lens to 2.8 is not the best option. We see this piece of wisdom demonstrated here. Leica would be well advised to use the lens design in the Null-series to counter the Heliar.
The upshot.
The choice of Berek and Barnack to select a 3.5/50mm lens for the Leica was a very wise one. And in those dys, one could take fantastic pictures with an aperture of 1:4, even with filmspeeds of ISO12 or ISO25. And today with more glass and better optimization techniques, the modest aperture standard lens may be in a revival, at least performance wise. We should be careful not to overhype the results. A topclass six element 50mm lens with an aperture of 1:2 and stopped down to 1:4 will perform very close to what the Heliar delivers, at least at the lower frequencies. Where the Heliar (and for that matter lenses like the Elmar and the Tri-Elmar) really can score is in the very fine image quality at the higher spatial frequencies. Many users may not be able to exploit this potential. And comparisons based on the low frequencies (outlines of subject details or the bo-ke), may miss the finer points of the differences that do exist between lenses.
All considered, the Heliar ranks in that sparse category of an outstanding design, that is realy challenging to the user.

.
 

Links

my Blog is now active: see Tab