An Open Letter to Leica
A Modest Proposal For Reinventing the M Series
Leica M9 with Tri-Elmar @ 50mm. ISO 160
Cards on The Table
Long time readers will be aware from both this site and earlier magazine articles that I've written, that I am a big fan of Leica M series cameras and lenses. A search on the word "Leica" using the panel to the left of this screen will show literally dozens of reviews and articles on the subject. I've used M series Leicas professionally since I worked as a photojournalist in the mid-60s, and an M Leica was always tucked away in my briefcase during years of subsequent international business travel.
Though many other camera brands and models have passed under my bridge, so to speak, an M Leica always seemed to be at hand when I wanted to use a system that offered the highest image quality possible combined with the smallest size and lightest weight.
So – plainly put – I am a fan of Leica cameras and lenses. But being a fan of something doesn't mean accepting it without comment or complaint. Far from it. In fact fans are often a team or product's most vocal critics. The reason being – because we care.
With that up front and out of the way, here is what I am describing as an Open Letter to Leica. I could of course send it directly to the senior managers that I know there, but I think that the ideas that I'm putting forward are worth a public debate, and that hopefully a lively and open discussion has a better chance of accomplishing something than a closed one.
What We Have
The Leica M3, the first of the breed, was introduced in 1953 and was an instant success. Indeed it was so well designed and executed that now, more than half a century later, it serves as the basis for the current digital M9, which in most fundamental respects is little different in its core approach to executing the taking of a photograph than its progenitor, the M3.
Of course that it is a rangefinder camera defines not only how it focuses, but also much of its core gestalt as a photographic tool. The RF design means there is no need for a reflex mirror, and as discussed below regarding lenses, this allows for the design of optics which are considerably smaller and faster than those which an SLR design requires.
The shorter back focus distance of an RF body also means that the chassis can be made slimmer and thus the body smaller, and potential at least, lighter.
Leica M9 with Tri-Elmar @ 28mm. ISO 160
Of course the cornerstone of the M Leica's success has always been its lenses. It's fair to say that there have been few lines of optics in the history of photography that have enjoyed such admiration or commanded such high prices, both new and used. And for good reason. There is something about the "look" of images produced with Leica glass that is hard to quantify, but which one learns to appreciate and even recognize more often than not. Of course Leica lenses almost always also test exceptionally well, but that's to be expected, especially given their cost.
Their size is also an important consideration. Because an M Leica has no reflex mirror M series lenses are considerably smaller than those designed for SLRs and DSLRs. This is especially true for wide angle and large aperture lenses.
The M series camera has now evolved into the M9, which has received considerable press here as well as elsewhere on the net and in print, and so needs no further introduction.
Not to put too fine a point on it, but with the M9 Leica has finally realized the dream of a digital M. In other words; full frame 35mm, no IR filters needed, and most if not all of the M8 and M8.2's foibles put to pasture.
There are many things that the M9 does well, and some not so well. Image quality is superb, and the use of a well-refined Kodak CCD sensor without an AA filter is largely responsible for this. As a company Leica has also now developed the in-house digital expertise needed to play with the big boys, and the company is consequently no longer reliant on others for assistance outside of their traditional core competencies.
But – and this is what lies at the root of the following Open Letter – whereas the M9 is the latest of the M series, it desperately needs to be reinvented. Not because it isn't a good camera, or even a great camera, but, because at its core it is a camera based on an antiquated approach to composing and focusing.
Leica M9 with Tri-Elmar @ 50mm. ISO 160
Let me explain. Traditionally there have been two ways to focus a camera's lens – either on a groundglass or with a rangefinder. A reflex camera simply interposes a prism and mirror to make the traditional groundglass of larger cameras more compact and accessible. The rangefinder uses instead an optical viewfinder window and a series of prisms and semi-silvered mirrors, as well as cams to link to the lens, so that two images of the subject are superimposed in the viewfinder allowing the user to manually achieve visual focus confirmation.
The viewfinder has its advantages. It allows one to see the subject without the dimensional flattening of a ground glass, and also (at least with longer lenses) with a view of what lies outside the lens' field of coverage. It also has significant disadvantages, including a small field of view (especially with longer lenses), a limit on the use of long lenses (above 135mm), and of wide lenses (below 28mm) without the use of awkward accessory viewfinders.
Rangefinder focusing is also problematic, in large part because of its slowness, and also because there are entire classes of subjects which simply don't have the simple clean "edges" required to focus on. Even subjects with lots of edges can be problematic, such as vegetation, where too many similar edges lead to confusion as to what is being seen.
Indeed, I could argue (and will below) that though there is a generation (or two) of photographers who cut there teeth on rangefinders and who are therefore used to their quirks and limitations (and even sometimes enjoy them), few though, at least those under the age of thirty, have much interest and patience in learning to cope with them.
Of course autofocus is no panacea, and as we'll discuss is not without its limitations and annoyances. Call me an equal opportunity complainer.
Leica M9 with 24mm f/1.4 Summilux @ ISO 160
What We Need
So – my friends at Leica, and the rest of you who care to read this – we're at a crossroads. The M9 is the best damn rangefinder camera that you or arguably anyone else has ever made. But, is being the best at doing something that a diminishing number of people are interested in a winning proposition?
I would argue not. Yes, I'm sure that there will be those, like me, who would eventually buy an M10 and even an M11, which continued to be based on the fundamental M3 design. But I somehow think that as those of us who were already photographers in the pre-autofocus and pre-digital era go to that big darkroom in the sky your market will shrink rather than grow.
The S2 is too new to be considered the next standard bearer. It may turn out to be a well accepted product, but it lacks the one thing that would be needed to make it a mainstay of Leica's financial future, and that's a wide selection of lenses. There are literally millions of lenses out there, worldwide, that fit the M series Leica, both Leica's own as well as those from numerous other lens makers.
Now that Leica has shown that it can build a full-frame 35mm digital camera system that can take any and all of these lenses, and use them to full advantage, there is a vast potential market for a camera body that can take best advantage of them.
I am arguing that the current M9, and any other camera built further on the viewfinder / rangefinder paradigm, has become an evolutionary dead-end. It serves to support the superb range of current and previous Leica M lenses, but it does so with a design brief that is now at least several decades out of date.
If one accepts this premise, even in part, then what should the next M series Leica be like?
I realize that considerable revenue still flows from the film traditionalists as well as the collector market. So, please continue with the M7 and MP, and if you have to do the occasional model for Japanese collectors, which are covered in reticulated ostrich testicle leather, then by all means continue to do so.
The first order of business is to stick with what you do best, and that is your lenses. Yes, they're expensive (and presumably you make very healthy margins on them). But they are just about the best optically that they can be for any given aperture and focal length, and their small size and build quality are welcome and admired even while we gulp at the prices.
I would argue against a move to autofocus. To add AF to M series lenses would require a complete redesign, which would be immensely expensive and which would of necessity charge the fundamental character of what an M lens is. No doubt you could do it, since the S series lenses retain traditional Leica optical quality in a modern AF design. But, we would lose the build quality and small size which have become the hallmark of Leica M glass. I also expect that the financial burden of such a broad redesign would also be beyond the company's current capabilities.
Leica M9 with Tri-Elmar @ 28mm. ISO 160
Instead, I argue that the current manual focus lens design be retained. If a live view screen or EVF is used (more on this in a moment) then "peaking", as used in pro-level video cameras can provide an ideal form of visual electronic focus confirmation. This causes the edges or areas in an image with the greatest contrast (ie: greatest sharpness) to shimmer or be outlined. Phase One uses this as a form of post capture focus confirmation in Capture One 5.x, and it is very effective.
What I like about this approach is that it would allow the user to choose the point of focus across the entire field of view, not just where AF points are located. So even the extreme corners can become focus areas. Also, because an M lens is used with the aperture at its shooting stop, rather that with a diaphragm pre-set, one will be able to exactly see the depth of field, because the entire part of the image that is in acceptable focus will be indicated with a confirmation overlay.
Because this overlay, or pattern, or colour, can be user selectable as to appearance so too can it be for sensitivity and for circle of confusion. A custom setting will be all that's needed to allow the user to be as tight or loose in these parameters as their needs may require. Indeed, being able to accurately see not just the point of focus, but also depth of field at shooting aperture, and with the ability to control COF, would be an outstanding contribution to photographic control and productivity.
So, now the controversial part. Viewing. Since this camera will use M series lenses it can not be a DSLR. Some pundits are arguing that this design's days are numbered in any event. That means either a live view LCD screen and / or an electronic viewfinder, and Live View pretty much means a change to CMOS from CCD.
In a recent discussion with a leading digital camera designer I asked about the future of CMOS vs CCD for the highest end cameras, such as medium format digital backs. While not given a timetable I was told that a move to CMOS is almost inevitable in high-end cameras and backs so as to be able to provide advanced features which are being planned, and which CCDs are not amenable to, at least not at anything resembling reasonable prices.
Source one of the one million pixel screens that are found on top-of-the-line cameras. I would even suggest that a size somewhat larger than the now almost ubiquitous 3" be considered. Given the size of the camera body needed to house a full frame sensor, and given that there is no longer a viewfinder port, a larger screen should be easily feasible.
I would argue against an articulated screen so as to reduce size and complexity. But, just as we had pop-up hoods on early reflex cameras, who not have such a design on an LCD? It can be made removable, if desired, but for tripod shooting outdoors would help even the brightest screen when in direct sunlight.
A detachable optional electronic viewfinder is a must. The best EVF available today is found on the Panasonic GH1. Surely something even better is possible, and with price less of a concern than on a consumer-grade product like the Panasonic, this shouldn't be that difficult. A high resolution OLED would be fantastic for this application. Also, as this is an area of technology that is changing quickly, since it is a detachable device if the interface is made flexible enough then future upgrades become possible.
Of course there's no reason why an optical viewfinder can't be attached. For those that prefer this I'd suggest that Leica consider developing a 28 – 90mm accessory viewfinder that mimics the view though an M9's viewfinder. I would also suggest that the top-rear panel of the camera have a small LED that lights when focus is confirmed. This way the user will be able to look through the accessory viewfinder, but then by turning the focusing ring be able to know when focus has been achieved.
Of course you won't know exactly what the camera will be focusing on, but there is a way around this. Read on.
Advanced Hyperfocal Prompting
For decades many rangefinder camera photographers have gotten by with hyperfocal focusing. In other words, pre-set the camera to a distance and aperture that provides depth of field sufficient for the type of shooting that one is doing. That's what the DOF lines on the lens are for – right? Then, the camera can be used almost as a point and shoot, with no need to be concerned about focusing.
But with the advent of digital this has become problematic. High resolution sensors require a considerably smaller circle of confusion than does film, especially when larger prints are made. In my experience one needs to stop down at least two stops beyond what is indicated on the engraved lines of a Leica M lens for optimum sharpness if hyperfocal focusing is used with an M9.
But – why not permit the LCD or EVF to show what hyperfocal setting to use? An M8.2 or M9 is able to determine the aperture set, even though there is no mechanical coupling of the aperture ring. This is achieved through comparing the light transmitted to the TTL metering sensor and the light measured through the separate exposure sensor window on the body, and is recorded in the EXIF data of the camera's raw file. I've found it to be accurate within a half stop most of the time. (Lightroom now knows how to read this data).
It then would be simple for the camera to produce a graphic display on the rear LCD or in the EVF of the preferred hyperfocal distance for the user to set the lens to, and of course a custom setting could accommodate the user's COF preference.
All major digital camera makers seem to be stuck in the film era when it comes to exposure metering and setting. Part of the problem is that consumers want the image on the rear LCD or in the viewfinder to "look right", but looking right and being optimum from a raw image quality perspective are not the same thing.
A live view image means that the camera's electronics are aware of the dynamic range of the subject and that means the ability to set the exposure "to the right", so that the data is placed optimally. Also, why can't there be a custom function that allows us to set the percentage of clipping that we want to allow, taking into consideration specular highlights, for example? This way the camera can constantly provide the optimum exposure to the raw file while the photographer concentrates on the preferred combination of aperture, shutter speed, and ISO sensitivity for any particular situation.
As for the viewfinder image, how hard can it be to normalize for viewing? Not very, I'm sure. Of course such an image file will come into raw post-processing potentially looking overexposed. But experienced photographers understand this and will work on the files accordingly. Or, how about taking the Live View compensation factor used to normalize the viewfinder image and embedding it in the raw file? I'm sure that there is somewhere in the DNG format that this can be placed, and then used by Lightroom to provide some form of auto-compensation on import.
In other words – let's leave the film exposure paradigm behind. Digital exposure is different than film exposure, and basing 21st Century cameras on 19th Century exposure rules has to end.
Leica M9 with Tri-Elmar @ 50mm. ISO 160
In its day the M3 was a brilliant design, and it still is esthetically pleasing in a way that few other cameras are today. But your slavish adherence to tradition has, in my opinion, hindered the M8 and M9 in their role as actual tools for doing photography.
For example – the lack of a film wind lever. Of course a digital camera doesn't need one, but the wind lever was also a critical component of the film M's ergonomics. It allowed a place for ones thumb to brace, which was intrinsic to proper handling.
Just about every M8 and M9 owner that I know has purchased the Thumbs-Up device. Why? Because it makes the camera easier to hold and shoot with. That should be the point of camera design after all – usability.
Of course since the 1980's cameras have become much more ergonomic, with molded grips now being de-rigeur. I'm not asking for the next generation M to become a jelly bean or a bar of Dove soap, but I do believe that it's time for Leica to bring some contemporary industrial design and haptics to this camera line. Your new S2 shows what's possible. It's an exemplary design. Why not take what was learned from that exercise and apply it to an M?
Oh yes – then there's the matter of the base plate. M series cameras have always featured a removable base plate because designing the camera with a fixed pressure plate, rather than a hinged back, made for greater body rigidity and thus, theoretically at least, a more accurately positioned film plane.
But of course with the M8 and M9 there is now a fixed sensor, not film, and therefore this whole design is simply an esthetic conceit rather than a meaningful design implementation. It means that to get to the battery and SD card the base needs to be removed, which raises the question of where to put it, the risk of dropping it, and the general all-around dumbness of this 60 year-old design carry-over.
So, please, let us have separate battery and memory card hatches on the M10, with proper weather-tight covers, and leave the removable base plate to models for collectors and the nostalgic.
It seems to me that the Leica M platform is at a crossroads. With the M9 we have reached the end-game in the process of translating the M series gestalt into the digital era. Naturally, there are endless tweaks and enhancements possible. But, will they attract new users to Leica and the M family and also motivate current M8 and M9 owners to upgrade?
I don't believe so. I think that as the installed base of photographers who grew up owning (or at least wanting to own) an M2, M3, M4, M5 (well – maybe not an M5), M6, or M7 begin to drop out of photography, or at least lose their inclination to purchase new gear, the potential market for yet another iteration of a rangefinder Leica will decline.
To attract new customers to the company's mainstream product line (and I'm not convinced that the S2 can take on that role), and to create an on-going market for Leica lenses, which are what you've always done best, requires some fresh thinking and a new approach to a 35mm format digital camera.
Maybe you won't care for my suggestions on exposure, focusing, or even viewing options. But regardless of the merits of these, I urge you to consider that the days of the traditional optical rangefinder camera may well have come to an end. Certainly there will be some potential digital rangefinder M series buyers for years to come, but will there be enough of them to sustain a profitable business?
With my best wishes for your next generation of products.
Toronto, January, 2010