The 28mm M-Hexanon, like the its focal length, occupies a strange space that is neither here nor there. I have never had good luck with 28mm lenses, if only because the angle is a little wide to be comfortable for close shots of people and a little narrow for some of the landscapes I shoot.
Only on the verge of selling mine (for lack of use since way back when I had an M8) did I shoot a bunch of tests with an M typ 240. This particular lens had been recollimated to be at exactly Leica spec (most lenses made before the M8 were not set up to hit the center of a flat sensor).
This piece will not editorialize much but instead show it like it is. Which is quite good, far better than I had remembered.
First, the obligatory “how sharp at a meter” exercise. This is f/2.8.
Next: does it shoot good pictures of children? Yes.
E poi – how is the bokeh? Strangely, it’s actually really good, especially for a wide lens. Here is the sequence f/2.8, 4, 5.6, 8.
Sunstars? Got ’em too. Here is f/2.8-8 (clockwise):
Gross resolving power (again, f/2.8-8):
And now, we laugh at your Elmarit-M!
Flare resistance, same range:
Another test; can’t remember why. Seemed like a good idea at the time.
As you contemplate modern lenses, It is difficult to associate these optics with the proud civilizations that created them: Tokyo, Rochester, Sendai, Jena. They worshipped image quality, because it is strength that makes all other values possible. No picture survives without it. Who knows what delicate objective lenses have died out of the world, for want of the strength to continue.
It’s a long way from the Canon 7sz to Andre Agassi, from the Serenar to the EOS Kiss Merkur XR4ti (…or whatever the amateur model of the week is). Canon used to be a hard-core producer of Leica knockoffs, many of which were more functional, convenient, and reliable than the rickety prewar designs that Leica kept elaborating. The Canon P was a standout, as were the VI-L and the 7 series.
The Canon 50mm f/1.2 (September 1956… can you believe that it is now almost 60 years old?) was the company’s penultimate halo lens. It came out with the VT (not Deluxe), and it would not be surpassed until 1961 with the Canon 7 and its 50/0.95 Dream Lens (from a size, weight, and cost standpoint, dream is clearly defined broadly to include nightmare). Ironically, the 50/1.2 cost more than the 50/0.95. Modern Canon lenses are quite good, but they don’t have that certain fun factor to them.
Synergies. This is a new thesis on my part, but I am starting to suspect that the M typ 240 and 246 tend to interact with lenses in ways that might not be immediately intuitive. Some lenses seem to work unexpectedly well with the color sensor (possibly the demosaic-ing algorithm accidentally boosting sharpness). Many lenses work better than expected with the Monochrom camera — likely because the color-bind sensor is not bothered as much by chromatic aberration. The “why” these things happen is probably insoluble; it is merely interesting that they do. And yes, some dog lenses do stay dogs, and some lenses have bad corners to f/8. Nulla regula sine exceptione.
Look and feel. Perhaps “Carré Otis” is the best way to express the relationship between the size of this lens and Canon’s more common 50mm lenses (2.8, 2.2, 1.9, 1.8, 1.5): generously proportioned, slightly flaky, and exhibiting certain, ahem, virtues.
The finish of the 50/1.2 is typical mid-to-late Canon that came into vogue in February 1956 with the 50mm f/1.8 mark II: a black enamel focusing ring and sandblasted chrome aperture ring and distance/DOF scale. This is much more pleasing than Canon’s contemporaneous all-black lenses (like the 35/1.5, 35/2, 100/3.5, 100/2, and 135/3.5), whose finish looks good in smaller lenses and feels chintzy in larger ones. But the affect with the 50/1.2 is solid. It will keep your M typ 240 on balance.
The focusing effort is heavy, and even the aperture turning is heavy. Unless, of course, lubricant has leaked out onto the aperture blades. The focusing pitch is very slow. Even the interesting push-button infinity lever takes a deliberate effort to dislodge. Is there a problem with any of this? No. Will you have issues tracking fast motion or switching from near-to-far subjects (or vice-versa) as if you were a D700? Yes.
Flare and use with filters. The first thing you have to remember with this lens is that you are dealing with an SLR-sized front element that sits very close to the front of the filter ring, seven elements, and single lens coatings from the mid-1950s. This isn’t going to flare, right?! Really, cleaning marks are the least of your problems here.
The big (physical) hazard with this lens is filters – as in don’t tighten down a filter without checking the clearance between the filter and the front lens element. If you want use conventional filters, you will likely need to use an empty ring as a spacer or fit a rubber o-ring around the threads on the filter. The only alternative is to use Canon RF filters, which are expensive and present a flush glass surface on the front of the lens. Not only does cabin your choices for lens hoods, it also presents ample opportunity for sidelight flare. Like the picture below (and let’s be fair – this is shot outdoors with floodlights everywhere). But be sure to use some kind of hood if you can.
Canon filter ghosting (at f/1.2). All hell is breaking loose here (not surprising with Christmas lights a scant two feet from the lens), but it’s not all bad, given the motif:
On balance, it is better to use a filter if you don’t want to deal with the rickety Canon metal lens caps, which never want to stay on. Just watch how you use it, and if you can space it correctly, consider an MRC.
General optical performance. On an M, at 24mp, the 50/1.2 really acts like Beauty and the Beast. At its largest aperture, it has microscopic depth of field, though with practice, you can tell where it is going to land. It does what most other super-fast spherical lenses do: it front focuses at wider apertures and settles down at middle ones. Every one of the three copies I have owned seems to have had slight differences in the midpoint of depth of field at close ranges and large apertures, which is not surprising. But this is par for the course; all fast 50mm lenses are testy on Leica rangefinders. As you start stopping down, the 50/1.2 becomes exponentially sharper. F/1.4 is light years better than 1.2; f/2.8 is orders of magnitude better yet.
Rather than dragging through every optical trait one at a time, let’s take it at f/1.2. Soft contrast, heavy vignetting, reasonable resolution of details, with a field that seems to curve at the edges toward the camera. This vignetting persists, even when the lens is coded as a Leica Noctilux. To be fair, this was more than enough for most 35mm film use. A lens like the Canon 50/1.2 would be used at night, where contrast would be high. It was not designed for thin depth-of-field fetishism. You can click on the picture below for a full-sized image.
The picture at the top of this article is also shot at f/1.2. In terms of controlling bokeh, if you are into that thing, this is yet another lens where your best bokeh is achieved by (a) getting the subject as close as possible and (b) getting the background as far away as possible. Not to belabor a point from the previous article (on the MS-Sonnetar), but an easy and almost unavoidable rule of thumb is that the better a lens performs wide-open, the worse the bokeh. The list of these suppressed lenses is long and distinguished: this Canon, almost every Noctilux, the Hexanon Limited, the Nokton 1.1, the Nikkor 1.1, and basically any lens faster than f/1.4.
The next click, f/1.4, is a touch more contrasty, and is still a good setting to use in harsh nighttime conditions. Take a look at the field curvature here – there is no way the Gummi Bear wrapper should be so close to being in focus like the boys. Actually, it seems unlikely that both boys should be in focus, but there you have it. And here you also have the bokeh vibe. As in vibration.
Just for fun, below is one shot wide-open with an M8 (you can click on it to see it at full size):
Stopping down to f/2 gives better results yet – and the focusing point is now exactly where you would expect it to be. And here is your bokeh test outdoors. This is essentially minimum focusing distance with a background that is 50m away.
At f/2.8, the lens is starting to hit its stride. Especially with the M set for emulation of the Leica 50/1.2, the lighting evens up, the sharpness goes up significantly, and the field flattens a bit. Although this begs the question of getting a slower lens and shooting more wide-open, with these old lenses, you are often better off with a fast lens stopped down than a slow lens wide-open. Also, the big old glass looks cooler and makes child ward nurses less likely to think you are using some kind of super-high-tech digital camera to document medical procedures (when you’re actually just bored).
At f/4, performance starts to max out (like the 1.2 shot, you can click on the one below to get a full-sized image). You almost go into double-take mode because a good example of the lens looks sharper than lot of modern glass. I’m thinking particularly of the 50/1.4 AF Nikkor, which the Canon crushes – and likely because the Nikon is usually tied to a somewhat limited AF system (phase detect systems seem pretty much incapable of compensating for the focus shift that occurs when a lens stops down).
F/8 is where performance starts to degrade a little bit (see the full-sized image). Sharpness starts to decline. People take on cartoonish, ascetic, or vampiric features. Bring your gloves, lightsabers, and garlic.
Conclusion? This lens was sold as the premium lens on a lot of Canon cameras – and in some ways a “bragging rights” lens vis-à-vis Leica and Nikon. The performance at f/1.2 is reasonable but not world-beating; but given limits seen even in digital M shutters (1/4000 sec), you would rarely be trying to shoot this lens wide-open in daylight. At least absent a neutral-density filter.
The pricing of this lens is all over the place; super-clean examples (from a cosmetic standpoint) seem to command a hefty premium, but almost all will exhibit microscopic scratches (cleaning marks), and oil haze is a recurring theme (and it is fairly destructive to these lenses’ coatings). But as with a lot of things, cosmetics are not indicative of performance; you never know whether a lens has been abused over the years until you actually try it.
Be ready to poke around through multiple examples to get one that works really well; keep an open mind about the condition of the coatings; what makes these lenses flare badly has little to do with the coatings but rather with the hazards of having that really big front element.
Unfrozen Cave-Man Design
The comparisons are inevitable (if you were born before, say, 1985). They are unnoticeable to Fujifilm’s obsequious band of pre-release “reviewers” (more on this later). But the similarity is undeniable. Fuji has, for its sixth camera based on the X-Trans II sensor and its eighth based on the 2011 Sony 16Mp base sensor, copied the design of a camera given away with magazine subscriptions. Hopefully unconsciously. That said, let’s not denigrate the Time-Life unit too much; it has a 50mm f/5.6 glass meniscus lens that at a small enough aperture will be competitive with multi-element lenses. It also contains so much lead in a ballast plate in the base that the scrap metal content outweighs (literally) the purchase price. Operators are standing by.
The only thing that makes the X-T10’s design really egregious coming from Fuji is that the Fuji X line is supposed to be a better-thought-out alternative to DSLRs. Yet here we are, in 2015, and the most recent two models have aped DSLR designs. Are we as a market that gullible? Do they think this will somehow make it easier for us to swallow giving up heavy SLR gear? Whatever it is, it does not say good things about the market or the manufacturer.
The silly game of making one thing look like another goes back a while. Consider the Horsey Horseless Carriage. Whether it was serious or a parody perpetrated by a rich gentleman, you get the point:
One is left to wonder whether the head was to be sourced from taxidermy or upholstery, but whatever the intent, it was not going to end well for horses.
Mimicry in camera design is not new, but it is a relatively recent phenomenon. In most cameras, form has to follow function; a camera is a box with a lens on one end and an imaging surface (film or digital) on the other. In the old days, there were no twin-lens reflexes that looked like rangefinders and no SLRs that looked like anything else. It is probably also fair to say that with a few exceptions (like the Zeiss Tenax or a couple of Raymond Loewy specials), no one actually cared whether a camera was ugly or not. After all, a Rolleiflex is only attractive in the context of twin-lens reflexes. You wouldn’t put it on a coffee table.
For some time, the proportions of digital SLRs were tied in to the film cameras that spawned them. Some of this was understandable; makers were in many cases recycling the chassis castings/moldings of existing cameras – or reusing key components like mirror/shutter boxes and viewfinder assemblies. When DSLRs started to feature their own purpose-built main castings, there was some carryover that were hard to explain – such as why grip surfaces retained proportions originally designed to house 35mm cartridges. But then again, the Space Shuttle’s engines’ dimensions are ultimately traceable to the size of the rump of a Roman soldier’s horse.
Fuji, for its part, stuck to function in designing its early X-series cameras. The X100 looked like a baby Leica M3, but any combination of an integrated optical finder is going to force a certain layout – the window either goes on the left of the right of the lens, and most people are right-eye-dominant. Yes, there was a little window-frame embellishment, but that has evaporated in the X100T. The X-Pro1 carried very subtle call-backs to the G/GL690-series cameras, but it too stuck to the function-defines-form script for the most part (it is clear given subsequent cameras that Fuji made this camera much thicker than it needed to be, given that it had a non-articulating screen). The XE, XM, and XA cameras looked like other finder-equipped or finder-less bodies – various Panasonic G, Sony NEX, and Olympus EP cameras.
The industry turning point (for the worse) came with the Olympus OMD-E5 in 2012, an unabashed visual clone of any of a number of Olympus OM-series SLRs. There was no reason to stick a pentaprism-looking housing atop a mirrorless camera. Pentax was also right there with its K-3. As if it had passed through a mirrorless camera development stage, the K-1’s top bump suddenly blossomed into a full-figured faux prism.
Fuji was always late to the party, and it took Fuji until 2014 to imitate SLR design in the X-T1, the pretext being that the big EVF required a pentaprism “hump.” Fuji dropped that pretext with the 2015 release of the blocky X-T10, stating now that it did this to recall Fuji’s (forgettable) AX line of SLRs. But the X-T10 does not look like an AX at all; it looks like a rinky-dink plastic camera. And its design appears driven neither by function nor aesthetics. It’s an ugly little box.
Why should anyone care?
On one hand, one would be tempted to ask, who cares? Fuji owners (and potential Fuji owners) should. Like a photographic version of roles written for Jason Statham, Fuji has for three years pumped out camera after camera based on the same sensor and incremental inclusions of off-the-shelf technology. Fuji’s three big additions since the X-Pro1 – namely, high-quality EVF technology, on-chip phase-detect focus, and face-detection – were set up for consumer products before the X-Pro1 came out (check out the timing of the NEX-5R and its patents). By the time the X-E2 came out, all the pieces were in place for a serious update to the X-Pro, the “flagship” camera. Between then and now, Fuji has instead pumped millions into design, tooling, and software for multiple minimally differentiated cameras – far more than it would have taken to put an X-Trans II chip, EXR II processor, and better EVF into an X-Pro2. This points to one of two possibilities: (1) the X-Pro1 was such a dog for sales that management required the engineering team to start doing what other mirrorless makers were doing or (2) Fuji has turned to avidly churning the market to keep up market share in the declining interchangeable-lens market, and an updated X-Pro1 was not anticipated to do the job.
1. Looking like what sells. On the first point, it is of some note that the X-E2 resembled the Olympus and Panasonic Micro 4:3 cameras, as well as the Sony NEX-6 and -7 APS-C Cameras. The X-T1 and -T10 have followed other manufacturers’ quasi-SLR digital designs. The lens selection in compacts of both formats (APS-C and M43) also reflects a more into competing with entry-level DSLRs: zooms, big zooms, and big primes.
This direction (physical bloating) undermines what APS-C (and Micro 4:3) were supposed to be about: smaller, lighter cameras. This has never really happened: Fuji’s and others’ lenses are not as much smaller than FX lenses as one might have been led to believe. Part of this may be that it’s cheaper to design big telecentric lenses than smaller, more symmetrical ones that require offset micro lenses. And autofocus probably exerts its own size expansion.
But for people who liked the idea of the X-Pro1, this translates into a camera that is somehow bulkier than a 24x36mm Leica M. That does not seem to be the right direction in an era where camera phones (that everyone is already carrying) are eating into compact camera sales. If aside from a camera phone, we are going to haul around another box with its own lithium-ion battery, one that is not plugged directly into social networking, do we want it to be bulky?
2. Churning and burning. The second possibility is more sinister-sounding – but it is supportable. Fuji’s product releases have occurred twice yearly since the X-Pro1. That is very often considering that the underlying technology has moved very little since fall 2013. Fuji’s marketing strategy for the XF has been simple: use shills to build up excitement, release products at high prices, slash prices when sales start to flag a couple months in, and then build excitement for the next big thing.
Fuji is not alone here, but it seems more visible in its use of “reviewers” to promote the process. The practice began with with some Fuji employees — but at least they disclosed who they worked for. But then it moved on to “reviews” started coming rom (a) semi-pros; (b) Fuji-sponsored photographers; and (c) a few easy marks who believe that whatever just came out – from whatever manufacturer – is the greatest thing ever (we all know who they are). Throw into the mix some hyperventilating Fuji-oriented sites that get revenue when people click through to retailers, and you get the perfect storm of non-objective reporting. After all, whether it trips FTC guidelines or not, who would bite the hand that feeds him? And in a world where people pay good money for SEO work, catapulting your photo business to the top of any search has value.
Then comes the product. It’s great. It takes great pictures. I know this first-hand.
And a few months later comes the burn. Left with a run that it can’t sell, and even absent any fundamental spec change or replacement model, Fuji will usually slash prices 20-30% within six months. This gives an impression that every Fuji model is overpriced to begin with – and in slashing new prices, Fuji puts its own new sales directly in competition with the secondary market. This in turn hurts middle-class amateurs trying to unload old Fuji equipment to upgrade within the line. This is a great strategy for fixing a one-time inventory problem, and certainly no budget shopper in the used market will object. But especially where forced depreciation occurs without some compelling improvement (or even the oddly missing “camera body” roadmap), existing users start to feel burned, and smart shoppers learn to hang back. Why would you ever buy new? Look at completed sale prices on Ebay. Buying an XF body or lens new costs you 30-40% the day you open it. Put another way, Fuji’s pricing practices violate a fundamental rule of luxury goods sales (and let’s face it, a $1,300 camera body is a luxury good for most people): never slash MSRPs. You can have occasional rebates, bundles, or “demo” units. But once you start slashing prices, you begin degrading your brand equity. Or has that happened already?
3. Rewarding risk? Fuji should never lose track of the risks that one takes on a proprietary camera system. XF lenses do not fit anything else. There is no repurposing the same lenses on old film bodies (such as with Canon, Nikon, Pentax, and Leica) – or even repurposing them on different types of digital bodies (you can stick the same Nikkor on an APS-C D7500, FX D4, and 36Mp D810, for example). In a closed digital system. people invest in a collection of lenses in part on the premise that the line is going to continue – and that the line will remain viable compared to other systems. In a sense, everyone knows that they will be replacing camera bodies in 3-4 years. But when real upgrades never come, it causes justifiable questioning. And it’s not just sensor resolution. It goes to functionality:
- Will battery life ever improve?
- Will there ever be a good TTL flash?
- Is there something about X-Trans decoding that makes it too processor-intensive for a 24Mp sensor?
- Is the “organic sensor” thing a dodge for never upgrading the X-Pro?
- Will the video function get less “aliasy?”
These are questions that Fuji should be in a position to answer.
Fuji presents a strange case. Its X100 line is fantastic (and its marketing low-key). Even in the XF line, there is little to complain about in image quality. But the reaction to Fuji’s marketing strategy? Maybe the best strategy is to wait out new Fuji XF product releases and just buy used. History, after all, tells us that most of the the prices are inflated anyway.
There is a tasteless joke whose punchline is, “well, we’ve established what kind of girl you are; now we’re just trying to establish the price.” It goes back to a newspaper column by the Hereditary Peer and reformed Canadian Lord Beaverbrook, it is probably fictional in origin, and it has been twisted around in a number of ways. Nevertheless, the quip is a great counterpoint to people who make a point of maintaining their photographic “integrity” by using some “less automated” form of digital.
Every technical aspect of digital photography (or as film snobs would call it, digital imaging) is nontraditional and somewhat automated. Light does not write an image on anything (we have the φωτός part; we have no γραφή). Instead, light hits an electronic sensing surface that translates light into analog measurements automatically, those measurements are converted to numbers automatically, and a computer in the camera bakes those numbers into a RAW image file automatically. That file is in turn transformed into something visible to humans, either in the camera or on a computer – and it is only in this final stage that human control returns, and it is a totally different type of control than chemical development and optical printing. The physics and chemistry of film photography are actually simple compared to the computational power required for digital photography. Put it this way: the oxidation-reduction reactions used in film photography are taught in high-school chemistry; the mathematical transformations needed to convert Bayer sensor measurements into recognizable images are almost graduate-school math. Or to put it bluntly: men went to the moon in vehicles with computers less sophisticated than what we now use to replicate the 1960s Hasselblad film cameras they took.
Functionally, digital imaging is like film photography in that you ultimately get an image on paper — but only similar in the way that a Selectric typewriter and a laser printer can both put crisp Courier text on a piece of white office bond. In both instances, you start with a keyboard and end with clean text, but the intervening operations are completely different. And with photography, both film and digital begin with a using camera and end with a physical image. But nothing in the middle is the same. That makes two things immediately suspect: (1) claims by manufacturers that their digital cameras build on their film competencies; and (2) claims by photographers that people should avoid using some of the possibilities that digital technologies provide. Leica culture is guilty on both counts. The easy part to identify is the design ethos of the digital M line: a digital M is designed to look like a film camera and not like a ground-up digital camera. This is understandable in light of the other part: the hard core of Leica culture thinks like Hesiod: there was a golden age (the M3), a silver age (the M2), and a progression of lesser ages that run up to and include the current product line (iron age is especially appropriate given Leica’s late penchant for stainless steel). Even among apostates who keep buying new Leicas (scribe, prepare the interdict!), technological resistance has historically expressed itself in apologetics. Leica zealots denounced autofocus — or autoexposure, or auto-advance, or digital, or whatever at the time of the denouncement Leica’s R&D budget had not yet allowed Solms/Wetzlar to implement. With autofocus, it was not entirely Pharisaic; even today, the only truly competent AF seems to come from larger, heavier DSLRs. But just as Paleo diets have captured the imagination of some, there is a set of rangefinder users who would like to go back to the days of the Kodak DCS line, when men were men and “chimping” referred to primates at play. Or better yet, they would like to return to the metaphor of the M3.
The Leica „M Edition 60” is simultaneously the fantasy and horror of Leica traditionalists. One group seeks continuity: an ersatz film camera suggests an unbroken line. Where that is not compelling, another craves “simplicity.” And yet others believe that omitting things like a screen would make a camera less expensive. A $20,000 camera package that is no lighter or smaller than a Typ 240 is going to sorely disappoint two out of these three groups. The acrimony is understandable. The remaining group might find suspension of disbelief easier. After all, Byzantine emperors still thought of themselves as Romans.
It is fair to guess that a camera made in an edition of 600 and packaged with white handling gloves will never sully its sensor with photons nor flush it with electrons. If it did, there would be legitimate questions of whether a digital camera, particularly a Leica one, is viable without a LCD screen and shooting only RAW:
- Shooting in DNG (i.e., RAW) is a poor substitute for proper exposure – and the Leica M meter has a tendency to produce results outside an easy adjustment range under a variety of circumstances: sunrise, sunset, flash. If the metering were more sophisticated on this camera, it might provoke less concern. But it’s fair to say that in tricky light, shooting the M architecture blind is not unlike exposing Kodachrome by guess. That, one assumes, is why the Typ 240 has auto-bracketing available.
- Lack of JPG capability can severely cabin on-the-road productivity and completely inhibits the use of Eye-Fi.
- Certain mixed lighting conditions that are relatively invisible to the eye (such as incandescent and daylight in the same frame) are detectable with an LCD, are correctable on-site at the time of shooting, and are extremely difficult to fix afterward.
- It would be a bitter pill to have a malfunction throughout a shoot that ruined the shots and was not detected until it was too late to make corrections. Think: rangefinder misalignment, a spot on the sensor, travel use.
In addition, some normal digital camera functions are completely dependent on the use of an LCD:
- Sensor cleaning is a stab-in-the-dark exercise without being able to look at stopped-down exposures quickly. And in any event, one would lose the dust detection capability of the camera.
- Lens profile selection becomes entirely dependent on Leica 6-bit coding.
- Filename/folder arrangements, formatting SD cards, and other “disk maintenance” functions make it hard to clear space if needed.
- Firmware updates would be difficult to implement.
And then there are some other things (normal features of digital and even many film cameras) that go away with the M Edition 60:
- Strap lugs
- Video (this is explicit)
- Self-timer settings
- Exposure bracketing
- Slow sync controls
- Auto ISO
- Frameline color
- Focus peaking
- Clipping detection
- USB mode controls
- Date/time setting
- User (settings profiles)
- Anything that has to do with JPEG generation (white balance, resolution, compression, film modes, color space)
There is no EVF workaround because the camera lacks an EVF port. So yes, as a digital camera, it is quite limited. These limitations may not have much effect on individuals shooting for pleasure. Theirs is no worse than the experience of shooting film, though the foibles of electronics inject a new element of risk. Photographers working in high-pressure contexts will not use something like this for the same reason they do not use medium format digital cameras: it is not the absolute disadvantage; it is the competitive disadvantage.
Functionality is a non-issue. Though a few perverse people will actually use the M Edition 60 to take pictures (just as one could use a silver dollar as currency), it is far from likely to be common. Leica’s replacement for the M Typ 240/M-P will undoubtedly have more technology, not less, and the superb industrial design of the „M Edition 60” will become a footnote like the M9 Titanium designed by Porsche or the M6J. Features of these special models may reappear (just as the high-magnification finder of the M6J and the LED-lit framelines of the M9 Titanium), but the whole package will not. The terrorists have not won; we can go back to screens and JPGs and video.
Leica, ultimately, wins here. It does not win on profit – a product in this low of a run barely pays for its own tooling. It wins in media exposure. Google “M Edition 60” and you will see that this device has put Leica on Engadget, DPReview, Wired, Forbes, CNET, and Petapixel. This puts the Leica line in front of a lot of people who previously did not know what Leica is – and more importantly, it puts the Leica brand in front of many people with disposable income. Not only does this represent a lot of free advertising for a niche brand, it is also likely aimed at selling more $7,000 M Typ 240s to people who don’t have $20,000 to drop on an M Edition 60 package.
Well played, Leica.
Above: 35mm Summilux ASPH (non-FLE), wide-open
***Note: click on pictures to enlarge
You’re going to lose your religious beliefs. There is a lot of talk about “rendering” and something special with the M9 and its CCD. When these people use an M long enough (and it does not take that long), they will wonder why they tolerated the Atari-age eletronics or the low-light performance (or lack thereof).
Challenges to image quality
The fact is that the Leica M’s 24mp is verging on the usable resolution of a 6×4.5 negative at ISO 400 and above. With no antialiasing filter and a new pixel well solution, it actually presents fairly stunning results. The catch, sadly is that three things hold back the camera.
a. Limits in the lenses. One limitation is older lenses (i.e., those before the late 1990s). Though charming in contrast and focus characteristics, they resolve less than the sensor can. And where they have uncorrected spherical aberration, they exhibit very obvious focus shift. This makes the extra resolution over the M9 (15% linear in each dimension) seem a fairly marginal proposition. The concept that comes to mind, though not completely apposite, is empty magnification in microscopy. You can multiply the pixels as much as you want – but in the end, the lens is not delivering enough information to make use of them. This is part of the problem with the D800E (and soon-to-be Sony A7r)
b. The human factor. The other problem is poor technique. With high pixel density, errors like poor focusing and camera shake are more apparent. Lenses of 90mm and up are difficult to use with the M’s rangefinder. They are already challenged because they are in the range where an SLR (or an EVF) does better with focusing. But the killer is camera shake. A good rule of thumb with the M is to set a shutter speed that is at least double the reciprocal of the lens length. A low-speed threshold of 1/125 second for Auto ISO is advisable.
c. Moiré. The new system uses considerably less aggressive moiré removal than the M8 and M9. As a result, the smallest resolvable details in shots (like the texture of bricks, for example) may pick up a little moiré sparkle. This is the downside of the clamoring to eliminate antialiasing filters.
The thing about having a very high resolution sensor is that it allows you to throw away data to get a better result. And it is fairly well documented that a large image downsampled to a smaller size is typically better than a small image that started that size. This is why you lose nothing with old lenses when going to 24mp. It also enables pretty spectacular high-ISO peformance.
The short of it is that the M crushes the M8 and M9 in high-ISO performance. There is plenty of quantification available on the interwebs, but here are the key things I have noted are:
- High-ISO (≤2500) is almost exactly the same as the D700 – except that you have twice as many pixels in play.
- ISO 3200 is completely usable, but dynamic range starts taking a beating. At a pixel level, it is still comparable to a Fuji X-Pro 1.
- ISO 6400 has regular pattern noise but is no worse than Tri-X pan film in b/w; no worse than scanned 400 color film.
If you want some concrete information on the last point, here is an example from where I accidentally shot a picture in broad daylight at ISO 6400 with an SF-58 flash. It’s a small miracle this worked at all, but here is the Tri-X detail at 4 stops more sensitive (and this is straight out of LR, with no attempt to fix the noise):
As with all other cameras, the lower the color temperature (i.e., incandescent light), the more noisy high ISOs become. This is due in part to the fact that auto white balance is trying to raise a blue channel that is at the noise limit.
Auto ISO weirdness
The fact that Auto-ISO doesn’t work for manual shutter speeds is well documented. As is the fact that your choices are either to go with the reciprocal of the focal length (if a lens is coded or selected) or an arbitrary shutter speed limit. One wrinkle that is not documented is what happens with flash. When the flash (whether Leica or Nikon-style) tells the camera it is ready, the ISO changes to the last value set in manual mode. To get to the picture above, I had previously used the 6400 speed to shoot at night. Taking a fill-flash shot during the day caused the ISO to change from auto to 6400. If you plan to use fill, I would recommend taking a couple of shots at ISO 800 in manual shutter speed mode to use that speed.
As against the M8/M9
The M is thoroughly modern in its sensor performance, whether measured as a function of resolving power or low-light performance. The move to CMOS, though lamented by the hard core of M9 users, has produced no measurable change in “rendering” and has increased low-light performance at least a couple of stops (and the collateral effect is to allow video). Not bad.
# # # # #
The Leica M viewfinder situation is a mixed bag.
Optical. On the one hand, it’s at least as good as an M9 finder in terms of pure optical clarity. This is no surprise – and it is difficult to imagine why it would not be. That said, there are several things one might want to note before dropping big money on this camera:
- The LED-lit framelines show the 2m view with a number of lenses. This means that for most shots, the lines will be a better approximation of the subject. The problem, though is that in a 0.68x finder, the 28mm frameline is essentially invisible, even before you add eyeglasses to the equation.
- 35mm-frame lovers will rejoice – because the 35mm frame takes up almost one’s entire field of view.
- The simple LED display now shows exposure compensation as you dial it in.
The bugbear here is that 0.68x does not seem to be high enough magnification to consistently focus a 90mm lens at a big aperture and the maximum resolution of the camera. This is frustrating, but at least it can be rationalized: most 35mm film has about 12mp of usable data; doubling the megapixels requires a 1.41x increase in resolution in each dimension; and if 90mm lenses were at the limit of the Leica effective baselength even in the film days, putting 1.41x the demand on the whole lens-cam-lever-prism system may be unreasonable. Of course, reducing the file size to 70% and downsampling returns apparent accuracy just like an M9 or M8. The fairly obvious solution is to move to the accessory EVF for critical work (under 2m, wider than f/4).
EVF. With the Olympus VF-2 (and why would you buy the more expensive Leica-branded one?), the camera takes on a new life. Contrary to many reports, the accessory VF is perfectly usable with the M. The high points:
- Works with all lenses and shows 100% of the FOV – as well as the distortion.
- Focusing accuracy increases with focal length because the system uses the lens’ magnification.
- Because it focuses through the lens (TTL), it takes into account focus shift and field curvature in a way that an RF cannot.
- Has an auto-magnification figure (5x or 10x) that is triggered by turning the lens focusing ring (and yes, it senses movement of the RF roller).
- Shows focus peaking, which is especially useful in the 5x view.
- Can show a live histogram that changes with your exposure choices.
- Costs less than any Leica glass viewfinder (a used VF-2 is about $170).
- Is difficult to activate by accident (see discussion of on-camera Live View, below).
The EVF has its somewhat clumsy points too. It requires you to lead the shot a little more. You also get a 2 second pause whether or not you do auto-playback (so you might as well use it). The refresh rate is not phenomenal. And it generally auto-dims to simulate the selected exposure when you half-press. But all in all, this is a very useful feature that enables the use of almost any lens with an M. It certainly provides a more accurate way to focus a 50/1.0 or a 75/1.4.
Live View (back display). Last (and least) is the back panel live view (LV). With the M, live view on the back panel is like the proverbial teat on a bull. This is not because live view violates some law of Leica conceptual purity, as some would claim. No, it’s because this feature is actually fairly useless. Start with the LV button, which is placed exactly where the play button should be (top of the button bank). It is very easy to activate by accident, and to tell the truth, for much shooting, you might not even notice it is on. Even when it is activated intentionally, the mode gives you a sight picture that is only usable in the one position where your hold is weakest – i.e., where the camera is being held by extended arms and not pressed up to your face. This means that things are quite shaky when auto-magnification comes on during focusing – and remain just as shaky in shooting with long lenses. Granted, you can shoot up or down at unusual angles, but truth be told, you can’t use it for truly low-angle shots. This is a feature that may have better been left on the cutting room floor. But you could see how it might work for the tripod-and-cable-release set.
What the hell?! moment. The dual viewfinder system of the M is brilliant for the Leica world and in reality is only slightly less usable than the hybrid optical/LED finder of the Fuji X100 and X-Pro 1. In fact, with the EVF, it might be possible to kiss off all Leica optical viewfinders except for one thing: no in-viewfinder level. The camera actually has 3-dimensional leveling, which is accessible by Menu–>Horizon (4th screen of options). But this leveling is not visible in the EVF or in Live View. This makes any optical viewfinder with an integrated bubble level a superior option. Is this to protect sales of the $900 Universal Wideangle Viewfinder M (12011)? Also conspicuously absent is a composition grid – which in the absence of a level, can still help right the shot.
# # # # #