Leica Fortress beleaguered? (2004)
Photography is clearly enjoying a new revival thanks to its absorption by the consumer electronics firms and the new life-style of casual, real-time distributed imaging . The phenomenon in itself is not new. The first predator action occurred in the sixties when 8- and 16mm movie cameras were wiped out overnight by the video camera, that in its turn was pushed aside by the digital camcorder. The currently employed capture technology in all digital imaging products has its origins in television engineering and theory that was developed from 1940 and later. The video camera and its successor, the digicam, became far more popular and widespread than the original movie camera, mainly because of its ease of use and the fact that it fitted in the lifestyle of the consumers. Photography has always been an isolated market, dominated by a few big players and there was only internal competition within the same technology. The photographic industry lost this luxury in1996, the year that digital imagery became the focus of the big electronics companies. The photographic industry had two comparative advantages that made it impossible to assault form outsiders: the complexity of film manufacture and the large scale of the film business and the mechanical complexity of the cameras, that made manufacture in low cost countries impossible: workers in the countries in the developing world have no problem with assembling a chip on a printed circuit board, but has problems when assembling a shutter, as you need intimate knowledge of the mechanics behind the product to properly assemble this.
Digital cameras are amazingly simple devices that can be manufactured everywhere and in large volume. The logistic chain of film development and printing is no longer a hurdle and increasingly users print their own pictures at the home printer.
The low cost digital camera with its 3 or 4 million pixel sensor is in danger of losing the battle with the newest generations of mobile phones that offer the same quality and even greater ease of use. Mobile imaging will replace digital imaging in a few years.
No wonder that the current big players are increasingly looking for new strongholds in the market arena. The digital SLR is the obvious answer. Such a product with good performance and moderate pricing could tempt the amateur with a film based SLR to go to the digital route. The culture and style of picture taking hardly changes and one can even use the lenses bought for the film based camera. To make additional money on lens sales, it is now the rule of the day to announce that only new 'digital' designs are suitable for digital SLRs.
The fight for market share goes on relentlessly as can be seen from the amount of new products with additional features, (image stabilizing is almost standard) and more pixels and lower prices.
Several large manufacturers have complained recently that prices will fall by at least 25% in the near future and that it is hardly impossible to generate a profit. A shakeout seems a fact as happened in the fifties of the previous century when the 100 camera manufacturers were reduced to less than 15 in a few years.
The lure of the rangefinder
Where to go for a niche market that will generate a profit over a longer period and that seems immune to the whims of the market and where competition is very low? More and more companies look at the rangefinder camera for this niche. Leica has been successful here for 50 years as is Cosina in recent years with its Voigtlander brand.
The state of the art at Leica
Current Leica M product development seems to be fully dominated by the wishes of a handful of collectors who are more interested in fulfilling boy dreams than in photographic instruments. As long as this stranglehold continues, Leica will continue to lose market share to more innovative outsiders. The Leica brand image has been changed during the last 50 years from an influential to a respectful name. It is in danger of becoming marginalized as long as the collector community seems to dominate product development, at least in the M area. What has Leica to offer at the Photokina 2004 in the face of the Cosina phalanx: a choice of colour and feature options for the M body, a very expensive commemorative M7 set to celebrate 50 years of M and a replica of the age-old Barnack camera. And the promise of a digital version of the M camera in a few years time.
The Leica M camera and the Harley Davidson motorcycle have much in common. Both have a loyal following, a great historical heritage and a product that has hardly changed in its history. The changes that have been made, have been carefully monitored to the taste of a conservative clientele. The big difference between both marques is the sales record. The sales of the M have been steadily declined since 1956 and have stabilized in the last years. The sales of the HD have climbed exponentially. This is remarkable as both marques have a traditional product that has hardly changed over the last half century. Why HD can seduce a large amount of new customers and Leica cannot, is partly the result of the marketing approach of Leica,
One of the first products to try to rejuvenate the rangefinder model has been the Contax G and G2. Zeiss and Kyocera/Contax threw in a sprinkle of electronics, AF and excellent optics in a very elegant styling. The camera was no success overall, despite raving reviews.
The second serious attempt was made by Konica with the Hexar RF. It was closely modelled on the M, one could even speak of a faithful copy, added an electronic shutter, motordrive and AE and very good optics. It could have been a strong contender, but the famous back focal length issues and a lacklustre support from the company killed this model.
The Voigtlander Bessa range from Cosina used a smart strategy. The first models were based on the cheap and tried Cosina SLR body, emulated the style of the famous model 3 of Leitz and offered a large range of excellent and very cheap lenses. The lens range filled all kinds of holes in the Leica range and could appeal to old Leica users and newcomers. With expanding tastes, Cosina offered more and more models, classic and modern and could cover all tastes and fantasies.
In 2002 Rollei introduced their branded version of the Cosina model R2 with a suite of three lenses. It was horribly expensive given its hardly disguised copy of the Cosina model. There is no news here and it may already be dead.
What is new in 2004.
Cosina introduced the Bessa R2A, the long awaited semi-automatic R2 to scratch at the M7 status and a new R3A with a 1:1 finder to cash on the famed M3 finder that is now celebrating its 50th anniversary.
The Epson RD-1 is also based on the Cosina model and offers a 6 Megapixel sensor. The camera tries to be modern and classical at the same time. The Epson management may be insecure about its success, as only 10.000 units seem to be in the making. It is very expensive with Euro 3000.00, but cash rich Leica users will buy them to see how the famous Leica lenses can perform with a digital sensor.
The introduction of the Zeiss Ikon (ZM) is a rehearsal of the famous battle between Zeiss and Leitz in the thirties of the previous century when Contax and Leica dominated the world market for rangefinder cameras. This body again is made by Cosina and under the shell is yet another version of the Bessa model. The Cosina phalanx seems to be unstoppable! On pictutes the Zeiis Ikon looks very good and certainly has more style than the basic Bessa. It is a fine blend of Contax G and Leica M styling elements. Its new lens line is partly new and partly copied from the G2 series. Most of these lenses are made at Cosina. Now the bayonet is the M-version.
The Zeiss lens line consists of
More lenses are promised and Zeiss is keen to stress the fact that the lenses can be used on future digital versions. Zeiss can simply wait and see how the Epson RD-1 fares and decide accordingly! Zeiss claims that all lenses have superior resolution, high contrast and no distortion and they state that focus shift is minimal. The topic of focus shift has been introduced by me, when discussing the Noctlilux behaviour and with other Leica lenses. It seems that Zeiss is a good listener! It is remarkable that Zeiss does not use aspherical surfaces or do not find it useful to mention it.
Normally Zeiss is realistic and even modest when making claims about optical performance of their lenses. Lately however, they seem to go for hyperbolic claims. The statement that all the new ZM lenses will resolve easily 400 lp/mm in normal photographic practice seems to be a bridge to far. I will return to this topic shortly. When a company like Zeiss issues and supports these statements, we are walking a thin rope.
The Leica M digital: is it Leica's biggest gamble? (april 4, 2004)
At this moment the only fact we know for sure is the intention of Leica to design and manufacture an M style camera with an integrated digital image sensor. There is a large amount of speculation about this enigmatic camera and the Photoshop artists have their day.
Let us try to cut through the fog and the marketing hype and discuss some basic facts. An digital image sensor has a certain thickness (sensor surface, microlenses and filter array) that necessitates a larger distance between the film gate and the shutter curtain, For the M camera the current shutter is too close to the film gate and a new shutter unit must be installed. Presumably (because Leica cannot build one themselves) they need the unit that is available for the Contax G and the Hexar RF). That is a vertically running metal blade electronically controlled shutter with integrated motor drive. As the proposal in Leica Fotografie International indicates, the advance lever will disappear. Functionally then the proposed M digital is a very close replica of the Konica Hexar RF (and will inherit all is characteristics?).
We might even follow this scenario, that the new KonicaMinolta company waits and sees if Leica succeeds technologically and then choose their own solution. The Hexar RF equipment is still there, one could expect. The M digital needs a new main body to accommodate the new shutter mechanism and the sensor device. But a new outer shell is required too. The Photoshop artist leave out all the slots that are required to inset the battery, the memory cards and the cables. One might assume that Leica could try to use the bottom of the camera for all these slots, but a glance at the Epson/Cosina body indicates that there might be engineering problems.
It is safe to assume that a new main casting and outer shell will be required. This will cost a lot of money. The main reason why Leica in the past has refused to make structural changes in the M body (hinged back cover as example) has always been the prohibitive cost of the new tooling for the casting. There is a hint in the LFI interview with Mr Cohn that Leica needed additional money to finance the investments in the new M digital body.
The ramifications are far reaching. A new body implies that the economy of scale that might be possible if the analogue and digital bodies could be produced and assembled in one assembly line are lost. We also may assume that the arrival of the digital M (given the current trend to buy and use digital cameras in great numbers) will cause the sales of the current MP and M7 to drop. Lower sales in this area (Leicas core market) will automatically imply a higher price or even, when sales fall below the break-even point a halt of production. The higher investment for the design and production of the digital M also implies a higher price and most certainly substantially above the current M price, which is already at the top of the market. The alleged 10 Million pixels sensor for the digital M will not be cheap and if the price tag of the digital back for the R8/9 is a guideline, the digital M could be selling at a price between 3000 and 4000 Euro.
Will anybody buy a digital M at this price level and with the functionality of a Hexar RF? This is a big gamble in my view. And Leica is a small company with limited financial muscle. If the market does now swallow the digital M at a very elevated price level in substantial units and if the presence of the digital M will erode the sales of the analogue M bodies, Leica could be in big trouble.
In the past Leitz did not pay much attention to the demands from the market, and they paid the price. Now Leica is maybe listening too attentively to the voices from all over the world.
The second gamble in this digital M proposal are the fixed focal length Leica lenses. The main trend in the digital market is the overwhelming swing to zoomlenses. Nobody nowadays wants to haul a bag full of heavy lenses to make pictures. And the zoomlens has very good quality as the test of the Digilux-2 lens indicates. Leica assumes that the digital M will be bought by people who now use a filmbased M and want to switch to the digital world while keeping their trusted Leica lenses. If this assumption is true, they are in fact saying that the digital M is a replacement for the analogue M in the traditional Leica market. But Leica needs desperately new customers, who must be persuaded to buy a digital camera that is very high priced, classically designed and equipped with traditional features as manual focusing. Yes I know there might be a retro trend in the digital photography world, but will this be wide enough for Leica to sell their products in substantial numbers?
And what to say about the lenses?. Fixed focal length lenses are not very popular and even the Olympus E1 is offering more zoomlenses than fixed focal lengths. The reduction of the viewing angle with the assumed 1.3 to 1.4 factor might be a problem too. Your 28mm will become a 36mm and your 75 will become a 100mm. The rangefinder accuracy might suffer. The frame lines are not the problem: you 'see' the 50mm frame lines and must 'think' that you have the angle of view of a 65mm lens. This is the same with the Epson/Cosina body. The finder is a 1:1 finder and many people were at a loss how one could put a 28mm frame in a 1:1 finder. But with a factor of 1.5, the 28mm lens becomes a 42mm (gets the angle of a 42mm lens). So the 28mm frame lines are in fact for a viewing angle of 42mm and that is easy to do as the classical M3 finder shows. In this respect one should try not to think in terms of focal length but in terms of viewing angle and then the switch is easy to make.
With a 10 million pixel array and an estimated 6 micron pixel size the limiting frequency (Nyquist) is 83 cycles (linepairs) per mm. Most Leica lenses are optimized for a lower frequency, but the limiting frequency is not a problem. With most M lenses I can get resolutions of 70 cycles/mm and more.
The topic of the microlenses deserves some closer attention. Conventional opinion (and that is partly driven by marketing people) wants you to believe that light rays should fall perpendicular on the pixel area. The image sensor should convert light into photoelectrons efficiently. The measure of this efficiency is called the quantum efficiency (QE). Light rays that fall on the pixel in an oblique reduce the QE and the result is a loss of photoelectrons or in photographic words: vignetting: the darkening of the image. The use of microlenses will increase the QE but has one bad side effect: the angle roll off: light entering the micro lens at higher angles is directed away form the photodiode and gets lost. This phenomenon has given rise to the proposal that it is best to use lens designs with a telecentric construction: that is all light rays from the lens will strike the imaging sensor at right angles and all rays are parallel to each other. This is the theory. But in reality here are no optical designs with a true telecentric construction in the current photographic market. Sorry for Olympus and others! If Olympus had telecentric designs in the new E1, why would they need such a large bayonet, compared to the diameter of the sensor area?
It may be possible that the angle of the rays may be somewhat reduced compared to previous designs, but that is all. And we do not need these designs. In a Kodak article (Photography with an 11-megapixel, 35mm format CCD) you find the arguments. There you find a graph where the angle roll of efficiency is diagrammed. This diagram tells you that at angles between +20 degrees and ?20 degrees the angle roll off is hardly important. Translated in normal parlance: light rays from a lens striking the microlens at angle at less than 20 degrees will be recorded with good QE! Given the usual 1.3 to 1.5 reduction factor in viewing angle most lenses will behave properly and their light rays will strike at angels within the indicated +/- 20 degrees. Another myth is being born and transmitted through the world! The same with the so-called D-lenses: lenses designed specifically for digital capture.
The idea is that these lenses have the optimized cut-off frequencies dictated by the Nyquist calculation and also have the more telecentric design 'needed'; for the microlens angle roll off. Both assumptions are not true (I need a full article to explain it in detail). What does happen is that the lens designer, knowing that his actual focal length of say 16mm will get an angle of view of the 21mm, can optimize the design for that viewing angle: so he can correct the lens as if it were a true 21mm lens (which is easier to do than to correct a true 16mm lens). Again a marketing trick!
The real fact is that we should not look at microlenses angles and Nyquist frequencies, but at the exit pupil location. Optically we are not interested in the last lens surface of a lens as the limiting factor for the angle at which rays strike the sensitive surface, but the exit pupil: here the rays are actually leaving the optical system. If the exit pupil is located very close to the image surface then the rays must be bend substantially to reach the edges of the image surface. And then, only then a moderate amount of telecentricity might help. The Leica R lenses, due to the relatively large mirror box, have their exit pupils located far in front of the image plane. These lenses can be used with confidence on the proposed digital back and the assumed limitation of microlenses and their angles will not pop up.
M lenses have their exit pupil locations closer to the film gate, but not all of them at the same location. How the M lenses will behave on a digital M body, cannot be guessed at this stage of the story.
">Is this the end..? asked Jim Morrison many years ago. (February 20, 2004)
If we have to believe the current hype about the proposed 'digital-M', one has to assume that one of the last bastions of high quality analogue photography is about to disappear. For some time it has been predicted that a rangefinder camera with an electronic capture module and interchangeable lenses based on the M-bayonet could be a real possibility. This type of predictions is of course very superficial and easy to make. And Photoshop manipulations where the current Leica M camera has been mated with some LCD display at the back were 'proof of concept'. At PMA 2004 Epson unveiled a prototype camera, based on the Cosina/Voigtlander Bessa R with M-bayonet. The number of persons who proudly announced that they had correctly predicted the 'digital-M' is proliferating daily. Let us try to take a balanced view. In today's world of instant information overload, it seems almost inappropriate to take some time for reflections on current events and trends.
The broader view of disruptive technologies.
Technologically speaking, digital photography is a subset of digital signal processing (DSP), more specifically of the area of image processing, that is signal processing where the parameters are measured over space and not over time as with most topics of DSP.
Commercially speaking, digital photography is an adaptation of the digital camcorder to the world of stills picture taking. We may recall that the first crisis in photography occurred around 1970, when the VCR, based on analogue tape recording, killed not only the 8mm movie industry, but also photography as a major leisure activity. Photography as a serious hobby for the masses was transformed into the TV based home entertainment industry. The main advantages (ease of use, integrated in the household electronics environment, reusability of the tapes and therefore reduction of cost) were obvious and are in fact the same ones, now being used for the digital photography. The current digital cameras are very closely related to the digital camcorder in features and even body shape and ergonomics. Many features of the digital camera (image stabilization, very wide zoom range, adjustments of color balance, electronic viewfinder, image processing with software) are common practice in camcorders. Several of the main players in the digital photography market are major manufacturers of digital camcorders: Sony, Canon, Matsushita/Panasonic and that is no accident.
The natural link of the moving camcorder image to the TV screen and the computer screen is the result of the new role of the PC as the hub of home electronics and entertainment. The inkjet printer has evolved into the home equivalent of the photolab/minilab. Printing paper and ink are now the main consumables in the market of digital photography as film and chemicals were in the analogue world. It is logical that the manufacturers of PC's and printing equipment will capitalize on this trend and so we see the likes of HP, Canon and Epson introducing printers, printing paper and related camera equipment almost monthly, trying to create a tight link between equipment and consumables.
The second crisis of photography is occurring at this very moment as the onslaught of the VCR is followed by the attack of the camcorder in stills format. Like it or not, the majority of digital camera products is simply a derivative of the camcorder technology. When the original VCR arrived on the market in the early seventies, photography received a staggering blow and there was a massive migration of camera users to vcr-users. The same migration will happen now in the transition from film-based to electronically based imagery. It is estimated that in the western world now about 30% of households have one or more digital cameras. The sales of analogue cameras are dropping rapidly and prices for second hand analogue equipment are in a free fall.
The third crisis is imminent, and will introduce the cellular phone with integrated camera. At the moment of writing, the number of sales of cellular camphones is already higher than that of digital cameras and by 2008 the camphone with a 5 Megapixel ccd is predicted. Companies like Nokia and Sony/Ericsson will be among the major players. Culturally we have made a full circle: the basic drive to take pictures has always been the need to record and fix permanently our visual memories of scenes (interesting landscapes, buildings and ) and events (marriage, babies, holidays). Kodak used this drive to create the famous slogan: you take the picture, we do the rest. Now the ubiquitous slogan is: you take the picture, the computer program does the rest. So we have in fact re-invented the wheel: 120 years ago Kodak noted that people wanted to fix their visual memories as fuss-free as possible and he invented the appropriate technology. Now we have digital technology that makes picture taking even more accessible and in real time. We may question the content and quality of the pictures being generated by the millions every day, but the fact remains that a new generation of users likes to take unpretentious pictures and see it as fun. And that is all there is.
Hybrid systems are a solution? Forget it!
Anyone who assumes that it is possible to hold on to a hybrid approach (Leica's stated goal for their M and R systems) misunderstands the basics of the disruptive technology that is being rolled out now. Once in the digital area, there is no reason to cling to analogue media or to allow one self to have the option of choosing the best medium for the job. The image quality of the current digital technology is as good as that what you get get from analogue media. There are hundreds of sites in the internet who try to prove with examples and/or calculations and/or technical reasoning that digital is superior/inferior to analogue (pick your choice).
The plain fact remains that users vote with their feet and that is digital! And for 95% of demands and expectations a 5 Megapixel camera with a decent lens will satisfy even professional demands. Remember that a few years ago the first Canon D1 digital camera had 1.2 Megapixels and most professionals were impressed. Even Berek calculated that 1 million pixels would be enough for serious small format images! And the current Nikon H2 with 3 million pixels is accepted as a tool for professional results. Filmbased users like myself, can argue and prove that the immaculate barita print with slow speed film and Leica lenses will have the edge over digital results, but that 95% of photographers will shrug their shoulders and ask: who needs this quality and who wants to input this amount of effort? And they will point to the ease of use, the immediacy of the results and the simple way of distributing the image. In filmbased photography the result is a unique product (the print) that needs to be seen in a museum or exhibition. In digital photography the result needs to distributed instantly as it will loose its impact when there is a long time lag. Filmbased photography will not co-exist with digital photography, but will be forced into a niche as vinyl (LP) records compared to CD or DVD music records.
In a recent article in The Economist (03-01-2004) the magazine analyzed the future of Kodak and concluded that this is not a bright one. As long as Kodak is still hooked on film, there is no way to adapt to the new business model, required to adapt to the rapid technological change. What is true for Kodak, is also true for Leica. The highest growth rates in digital cameras can be found in companies with a nontraditional background in photography. HP grew by 113% in Q3-2003. But in market share it is Sony, Canon and Kodak who are at the top three positions. The Economist proposes that Kodak be divided in two separate companies, one chemical and film based and one digital and electronically based, as the technologies and markets are too different to be covered by one single company . This might be true for Leica too!
What is Epson up too?
Looking at the figures, we see that Canon and HP, both making printers and cameras are big in market share and growth rate. Epson tried loose partnerships with Olympus, but the 4/3 format is still no winner. So Epson must have thought that it needs a higher profile in the camera market. Epson is Seiko and Seiko has experience with camera shutters and is familiar with the Cosina cameras. They need a cheap camera to capture a slice of the market and knowing the popularity of the Leica M and screw mount lenses/cameras in Japan and USA, it is logical that they want to profile themselves as a maker of digital cameras to cater for that market. Leica has no product to compete in this specific digital area and if Seiko/Cosina can succeed to seduce Leica/Voigtlander users to go the full digital route, they may be inclined to buy the Epson printers. This is a big gamble of course: the digital market is mostly focused on cameras with fixed zoomlenses with a wide zoomratio as this is the most convenient option. And changing lenses is not a popular act today. Japanese firms however are famous for capturing easy markets first before going into the tough markets. Canon started by copying Leica cameras when that market was booming. It may be that Seiko assumes that competing head on with Canon and Nikon and Sony is not the best option to start with. But gaining market profile and experience in a not hotly contested area like the rangefinder market, may be the best starting point for a more aggressive strategy over the next years. And if they fail, there is no big investment to write off, as the Cosina cameras are already available and a chip can be bought from several sources.
Is the proposed Seiko/Cosina camera the digital M some people dream about?
Of course not! A camera body with an M-bayonet is just that: a body that accepts Leica M lenses. As far as I can read the specs the proposed S/C body has a 1:1 rangefinder without frame lines for the several focal lengths. You focus with the rangefinder and frame the picture with the digital screen at the back. There is no AF or AF assist as this would render the camera incompatible with the M-lenses. If the S/C camera hits the market at the end of year, we can at least assess one important issue: can Leica lenses offer superior results when mated with a 6 megapixel back? I personally doubt it, but let us wait and see. I am sure a S/C camera will deliver quite good results, that will satisfy most users of the M-system and the V-system. This is easy to predict: any 5 or 6 Megapixel digital camera delivers very good results at this moment. And the Leica lenses will perform commendably, but the ultimate quality depends not on the optical quality, but the software in the camera and in the printer drivers. Leica cannot influence these parameters.
What about the Leica statement announcing their own digital M version?
The announcement by Leica must have been motivated by the unveiling of the S/C prototype camera. Some months ago, the company stated that a digital M was not feasible. Technologically nothing has changed in the last months and now the company makes a 180 degrees turn in philosophy and approach. The S/C prototype is not a proof of concept, just a mock up. It is too early to speculate, but we may assume that there are many technological hurdles to take to develop a digital body that accepts M-lenses and can deliver exceptonally good results. Whether Cosina or Leica is in the best position to solve the problems remains a secret at this moment. We may refer to another Economist statement: most predictions are based on current knowledge and so will be false as we do not know what the future will bring.