Article Index

Resolution

I must admit that I have a love and hate relationship with old lenses. Generally speaking, I have a weakness for historabilia and  am always excited to try out an old lens, particularly a lens coming from such an excellent heritage as Summicron.  It is always interesting to compare and contrast old designs agains new ones, handling of flare, resolution, rendering. etc However, with a very rare exception, I do not keep old lenses, almost always opting for newer models. The most common issue with older lenses like Pentacon, Carl Zeiss Jena, Meyer-Gorlitz etc is the often subpar resolution. While resolution obviously is not the only criteria that should be considered in a lens, it is the first place I start when looking at an old lens for my personal use. But because I have had the opportunity to try out so many different lenses over the years, I have developed certain rules (of thumb) I follow when evaluating a lens. For example, with fast lenses (f/2 and faster), I expect some softness at wider apertures. It is unavoidable for most but the cream of the crop lens designs and it should not automatically disqualify any lens. I, for example, rarely, if ever, shoot fast 50mm lenses at its widest apertures from far distances, and almost always try to get  as close to the target as possible, to maximize the out of focus area and isolate the subject. Hence it is much more important to me to have an acceptable center resolution at the widest aperture and close to the MFD. Do I want high resolution around corners at f/2 as well? Sure I do, but it's not critical since at close distances corners are going to be out of focus anyway. Hence when examining the resolution (or any other) part of the report, make sure you apply your own criteria for how you plan to use the lens, instead of just accepting a generic claim that says that the lens produces soft (or sharp) results in the corners of the image.

If you are into the rangefinder stuff, you're probably familiar with the Summicron's descendants. Summar and later Summitar preceded Summicron in 1930s and 1940s. Both lenses can be found today for little money, but are not particularly popular among general masses who perceive these lenses as 'specialist' (I tend to agree). Both lenses offered reasonable performance at smaller apertures but varied in their rendering capabilities and even more so in their resolving capabilities wide open. With every new generation of designs, Leica engineers managed to improve resolution and the Rigid Summicron offered by far the largest jump in performance compared to the previous designs. Hence the popularity of the lens back then and even now - while the modern, Type 4 Summicron is often considered to be the epithomy of 50mm lenses, Type 2 Rigid is now perceived as classic and is often thought after by many users. With such a rather lengthy introduction, let's take a look at how the lens  actually performs in real life.

As a reminder to the readers, the lab tests were conducted in the 1m to 10m range (giving 20x to 200x distance to focal length ratio), which covers almost entire focusing range for the Rigid Summicron, except for the 'close to infinity' area. The field test, on the other hand, was conducted at the infinity mark. As one can observe from the simplified MTF50 chart below, the lens shows pretty good resolving capabilities in the center, while suffering quite noticeably around edges at wider apertures. Basically, center resolution remains fairly consistent throughout the aperture range. Borders are quite soft at f/2 and f/2.8, getting a noticeable boost around f/4 and with further stopping in the aperture. Without having done any direct comparison, I would nevertheless argue that in the f/5.6-f/11 range, the Rigid Summicron provides as good resolution as one might get from much more modern lens versions, Leica's or anyone else's, but those interested in corner to corner resolution at wider apertures would probably be left craving for something else. Let's still take a look at a couple of crops before moving on.

 

Leica M8 (10Mp)

 

The two crops of the imaging target below compare image quality around border at f/2 and f/8. You can observe that the resolution at f/2 is fairly low, with the image smudged and lacking detail, pretty much confirming what the MTF chart claims. f/8 on the other hand produces much better detail and gives a crisp, well defined image. Not quite convincing yet - let's take a look at the infinity test results.

 

Leica M8 (10Mp)

Image borders - f/2 vs f/8 (APS-H)

 

Continuing with the resolution tests, take a look at the series of shots below, which compare image quality that the lens offers at infinity. Shots were taken with M8, wide open and stopped down to f/8. The distance to the target was over 20m, which  for all practical purposes can be considered infinity for the Rigid. Reviewing the crops, you can see noticeable differences in quality between the images taken at f/2 and ones taken at f/8. The most noticeable difference is around borders, which is what we suspected based on the Imatest results. Image quality at f/2 is lacking and the crop areas look outright blurry. Center, while somewhat softer at f/2 than at f/8, is not partiicularly bad - to my eyes result is quite acceptable. On the other hand, f/8 results, center or corners, are as good as one might ever wish, which confirms the overall trend discussed earlier in the lab section.

It is worth mentioning here that Summicron 50mm f/2 Rigid shows very minor signs of field curvature - focus bracketed shots revealed very minor variance in quality in series where sharpest corners did not coincide with series that had sharpest center. The difference is pretty minimal at f/2 and close to the MFD, but basically non-existent beyod f/4 or at infinity.


Center

f/2 vs f/8 (Leica M8)

L.L.Corner

f/2 vs f/8 (Leica M8)

U.R.Corner

f/2 vs f/8 (Leica M8)

 

Finally, before wrapping up with the resolution section of the test, let's take a look at the simplified graph that shows the dependency of Summicron's resolution to the focusing distance to the target (as measured by the Imatest). The graph plots function which is an average of of center and borders at all measured aperture levels and specific focusing distance. Note a fairly uniform performance across most of the focusing range, with a slight dip at the MFD.

 

 

Color & Rendering

Like many older lenses from 50s and 60s, Summicron Rigid is a lens of contrasts, pardon unintentional calambur. What I mean is that many users are likely going to be polarized by this lens, some will love how it renders and others will hate it. After playing with the lens for about a month, I am still undecided whether I like the Rigid Summicron or not. What makes me pause here is the fairly low contrast that the lens generates at wide apertures. Now, for some of you this is going to be a deal breaker, in which case you should pick up Carl Zeiss Planar T* 50mm f/2 ZM and never look back. Low contrast is not a deal breaker for me personally, as long as I can boost contrast in post processing without sacrificing overall image detail too much - since contrast boost means that the color gamut gets compressed as an after-effect, the raw files produced with this lens should contain enough data to mitigate the compression. Shooting in Adobe RGB  instead of sRGB is the first necessary step to expand the gamut coverage, but the lens glass should also be able to transmit the entire color gamut without compressing or distorting any significant part of it. That is where I have been having some challenges with the Rigid Summicron - the data files produced by M8 at f/2 and to a slightly lesser degree at f/2.8, had less information in the shadows, which in turn did not allow me to easily boost contrast while also preserving good shadow tonality. Let me pause here for a second and also mention that I am not talking only about global contrast levels, but also about local contrast (how accurately minor changes in gradation are represented in a region). Summicron Rigid suffers in both areas at wide apertures, which is what is making it hard to recover contrast without loosing detail. Boosting global contrast is not much of a problem and any of the standard post-processing technique like resetting black/white points, curve or contrast boost, even unsharp mask would give fairly decent results, but unfortunately at the expense of some detail loss in shadows. So whenever I wanted to retain a little more detail in shadows while also increasing overall contrast, I had to layer masks with different transparency gradients over areas where I wanted to preserve detail - not a particularly fun excercise for me personally, but doable.

ISO 160, 1/2000, f/2.8, 50mm (Leica M8)

However, one should not despair, particularly if one is predominantly shooting in B&W or prefers a low contrast photography, for example for portraits. If that is the case,  than Summicron might actually delight you. Seriously, a combination of high center resolution and low contrast could be used effectively in portraiture photography to lighten the feel, even make the subject look whimsical (is that actually a word?).

But, if you're like me and own a couple of other 50mm lenses, then the question you would probably ask is whether Summicron brings anything special to the table, anything that the other lenses don't. If I had to rank Summicron 50mm f/2 Rigid, I'd place it somewhere between Summitar 50mm f/2 LTM and C Summicron 40mm f/2 in its contrast and overall color reproduction capabilities (I have not tried other versions of 50mm Summicron extensively and so cannot comment about their rendering capabilities at this moment). Summitar is a pretty archaic lens in my opinion, interesting in its own right for certain type situations and effects, but certainly is not an every-day lens like Summicron. And unlike Summitar, the Rigid brings up detail much faster - by f/4, contrast is back to pretty good levels, with enough detail to rival anything modern.

Besides feeling somewhat challenged in the contrast area, Summicron also delivers somewhat mixed results in handling chromatic aberration. Lateral CA, as measured by Imatest, remains fairly negligible, hovering around 0.4px across the frame and all aperture levels. On the other hand, the lens also showed a little bit of spherical aberration in image periphery. It is not as bad as with Summitar or even modern super-speed Nokton, but it is visible in some cases.

 

Leica M8 (10Mp)

 

Earlier I mentioned that I am undecided whether I like the Rigid Summiccron or not. Let me rephrase this in a slightly different way - - to me, the question of whether to keep this lens or not is all about whether it closes a gap between the lenses that I already decided to keep. These include Carl Zeiss Planar T* 50mm f/2, which is my every-day lens, Voigtlander Nokton 50mm f/1.1, which is my 'low light' lens, and the above-mentioned Summitar 50mm f/2, which is my 'I am feeling lucky today' lens. I will most likely ditch Summitar in favor of Carl Zeiss C Sonnar T* 50mm f/1.5, which I find  capable of doing what Summitar does plus more.

ISO 640, 1/60, f/2.8, 50mm (Leica M8)

So then where does the Rigid fit? When you think about it, it does not fit anywhere - two of the above mentioned lenses, Nokton and C Sonnar give low contrast at wide apertures, C Sonnar also offers excellent center resolution, even at f/1.5, all three lenses produce good contrast and good resolution at smaller apertures, and Planar is super-contrasty and  sharp throughout the aperture range. So basically Summicron becomes redundant. And that's the cause of my personal dilemma. Would things change if I did not own other 50s? Most certainly, and the Rigid Summicron would certainly be an interesting lens to consider, but I would have still tried to find a lens capable of delivering good contrast at fast speeds to supplement the Rigid - a lens like Planar or perhaps the now discontinued Nokton 50mm f/1.5 ASPH.


 

DOF & Bokeh

Ahh, bokeh... The most elusive and fuzzy subject one can ever think about when evaluating a lens. If you read comments about the Rigid Summicron peppered throughout the web posts and occasional review, you'd build an impression that this is the lens that produces the best bokeh. Ever. Until you discover that there is also the 'bokeh king' Summicron (which is actually a 4th generation Summicron 35mm f/2) as well as 'bokeh deity' and 'bokeh monster'. Wait, what? Bokeh king? Seriously? How can poor Rigid compete with that? Well, let's take one step at a time - turns out that Rigid is not a slouch in this area either...

We will start with the traditional shot of a ruler, done at the MFD of 1m and at ~45 degrees to the surface. There is not much to see in this shot except for the simulated DOF scale which can give you a sense of how thin the DOF will really be. If you plug the numbers into a DOF calculator, you'd get 3.5cm, which is while decent, is not as shallow as with Nokton, because of its larger max aperture or even Planar, because of its closer MFD (by the way, for M9 users, DOF will be ~4.6cm because of the full frame sensor).

 

ISO 320, 1/180, f/2, 50mm (Leica M8)

 

But the main influencing factor with the Rigid actually seems to be the minimum focusing distance of 1m rather than anything else. If you examine the next series of shots, which compare how the lens renders out of focus areas at f/2 and f/8 at MFD of 1m and at ~2.5m, you will understand what I mean. At f/2 the lens produces a pretty nicely blurred background, with all but the largest objects rendered in fairly smooth, wide brush-like strokes. You can notice also some badly blown out OOF highlights - typical for this lens in bright conditions where I discovered I had to underexpose the shots by 0.5EV to achieve a more even luminocity distribution. Lower tonality and resolution around corners helps make the background feel more subtle, but unfortunately same low contrast hurts us in the center, making isolation between the foreground and background less distinct. Stepping further away brings a bit more background into focus, but general characteristics remain same. On the other hand, lower apertures boost up both contrast as well as resolution, making images look more crisp. Contrast however the results at f/8 here vs the results shown in the resolution section (both sets shot on the same day by the way). To my eyes, it looks like the lens produces better contrast as well as resolution at infinity rather than closeup (Imatest results do show the difference but a much more subtle one).

DOF @ 1m

ISO 320, 1/1500, f/2, 50mm

ISO 320, 1/125, f/8, 50mm

DOF @ 2.5m

ISO 320, 1/1500, f/2, 50mm

ISO 320, 1/125, f/8, 50mm

 

If you examine images at 100% magnification, you will notice some signs of spherical aberration in the out of focus areas, which I briefly discussed earlier - not a biggie really, the artifact is rather subtle, but noticeable if you know what you're looking for. The overall feel of bokeh at wide apertures is decent in my opinion (assuming you're shooting at distances close to MFD, like I do), but I can't stop wishing that Summicron had better local contrast in the center  at f/2 - just a little bit, enough to bring up more detail to make the subject look a bit more lively.


ISO 320, 1/180, f/2, 50mm (Leica M8)