Leica M9: the final frontier of the CRF? (2010)
The M9 is now six months on the market (and in my hands too). Enough experience and knowledge has been accumulated to try to give a final verdict. Recall that in the classical age of precision miniature cameras a full test of the capabilities of a camera system took at least a year before one could sum up the results. Recall too that these cameras, impressive as they were and are, offered much less sophistication and complexity than comparable current digital slr models. A Canon F1 is simply not in the same league as a Canon 1Ds Mark x. And a camera like the Nikon F2, once sitting on the Olympic heights of 35mm engineering, is a very simple tool compared to the Nikon D3s. This is not to imply that pictures made with the F1 or F2 are less good than what we see today. Progress in vision and creativity has not advanced as much as progress in digital technology.
To digital camera is no longer that stand-alone product that the mechanical precision models were. In the digital world the interaction and interdependency between the camera, lens, the in-camera and in-lens software and the post-processing software is quite intense, quite complicated and partly a black box. I have referred to the case of the Zeiss 1.4/85mm lens where the results are influenced by the interaction of the in-camera software, the Eprom in the lens and the layout of camera functions and optical construction of the light paths for AF and exposure. While it is easy to see the defects in the results, it is hardly possible to offer a good explanation of the causes and by implication it is not possible to say whether the camera, the lens and/or the software and/or the operator/photographer are the villains in the plot.
To grasp these complexities asks for an extended period of use and reflection and that old rule that at least a year of concentrated usage was required before a meaningful report could emerge is certainly true today. But the current mood is to ask for real time discussions and reports. Such reports (and the forums and sites on the internet are teeming with this type of discussions) have limited value.
Only after a prolonged period of use, some persistent patterns are visible.
In my email correspondence I have encountered two extremes in the assessment of the capabilities of the Leica M9 camera. On the one extreme we have the owners who stare at the M9 in admiration and are convinced that the M9 is the best digital camera in the world, beating the redoubtable competition of C/N with ease. The other extreme consists of owners who are obsessed with purple color fringes, focus shifts when stopping down, red edges when using extreme wide angle lenses, artefacts in the neighborhood of the Nyquist limiting frequency and so on.
The truth, as usual, is in the middle.
Take the case of the color fringes. In the agX days, color fringing was present and caused by the lens defect known as chromatic aberrations (CA). By careful design and the use of special types of glass the amount of CA could be minimized, such that in all but the most extreme magnifications, the effect could not be or only with difficulty discerned. And when it was detected, it was accepted as a fact of life: perfection is not possible in this world. This state of affairs still exists: lens design has not changed, but the capture medium has and the way people deal with current technology is different too. Most users now are not focused on the final printed image, but use their computer and software as a personal optical and technical lab, searching for defects at magnifications that are unrealistic in normal photographic practice. Who is making prints of billboard format? In the past you had this approach: when the practical limits of 35mm photography were encountered, you switched to a next larger format where you could work with smaller magnifications and bigger sizes. The current equivalent would be this: you are struggling with the limits of the M9? Buy an S2!
If you study the MTF graphs of older and highly acclaimed Leica lenses, you will see the effects of the focus shift when stopping down. But there were no complaints in those days. The effect was there to be sure, but again no one would notice it, as the focus was on image content, not technical lab testing. I only mentioned the focus difference once, when reviewing the Noctilux 1/50, because with this lens the effects could become noticeable.
The real cause of all the troubles in the era of digital capture is the sensor with zero depth. The agX emulsions had a layer of 10 to 30 microns thickness and this layer could accommodate most of the focus difference and some other mechanical and optical defects. The digital sensor on the other hand has zero thickness and on top of this there is that layer of micro-lenses which should not be interpreted as lenses in the classical sense, only smaller, but as a large blanket of bubbles overlaying the photodiodes. The shape of these micro-lenses is not exactly identical, element for element. In the case of the Leica M camera, the shape of the micro-lenses is such that some color fringing is exaggerated and even caused by that shape. If we note color fringing in a Leica M picture, there are several causes: the property of the lens, the shape of the micro-lens and the interpretation algorithm of the in-camera processing system.
Leica is of course fully aware of these conditions. The engineering precision and manufacturing tolerances are at least a factor two better than what is possible with the film-loading M cameras. The rangefinder assembly and adjustment of the M8/M9 is twice as accurate as with the current M7/MP cameras.
When holding the M9, you have the best built and best assembled M camera ever in your hands. It is still a CRF camera which means that there is an opto-mechanical linkage between the point of focus of the lens and the sensor surface. And the sensor surface of the M lacks the lowpass filter of other camera types. Most users of digital cameras do not grasp the fundamental difference between these types of cameras. The lowpass filter in combination with clever in-camera software can mask many defects and enhance the image sharpness over a wider range of frequencies. The software can also shift the visual occurrence of defects beyond the perception threshold of most users. The lab results of high-end digital slr-cameras are often shockingly lousy, but are concealed from most critical analyses, thanks to he cloaking effects of the lowpass filter.
The Leica M9 camera is quite honest in this respect: it shows you was is possible and where the limits are. The good point is that you can understand the limits and develop a working procedure to cope with these imaging borders and mold the results such that they satisfy the imagination. This is the same approach as Ansel Adams used when he found the limits of silver halide photography. He devised the Zone System to cope with the inherent capabilities of silver capture and to enhance the possibilities.
There was a period when the Leica M had natural advantages in engineering quality and optical capabilities. But now half a century later, these advantages have been minimized by the relentless improvements made by the competition and the different technology of digital capture and processing. Now the Leica M9 delivers results as good (or only slightly less good) as the best of the digital reflex cameras, but now as then you need to adapt to the characteristics of the camera to get the most out of it. If you approach the M9 as you would do with a 1Ds or a D3, you miss the point. The C/N approach is to optimize the camera software to improve the results and to use a lowpass filter to mask the defects. The Leica approach is to show the physical/optical limits of the camera and let the user decide. I find this a more sympathetic way of practicing photography.
For most situations and in most hands the C/N philosophy delivers the best results and without too much interference of the user. The Leica philosophy asks for a more balanced approach and a more dedicated and experienced use.
The Leica M9 system has its shortcomings: the color artefacts at the Nyquist limit, the color fringing due to the shape of the micro-lenses, the white balance is not always correct, the exposure often needs some adjustments and the rangefinder mechanism does show more flare in the rangefinder patch than the M7 as example, especially when you use the camera in vertical position. Some of these points can be corrected during assembly or during a service. Some others are inherently coupled to the Leica concept. The focus shift will gradually disappear when new lenses are introduced. All new designs are specifically trimmed to reduce this effect. But older lenses can not be adapted.
There is a story about Martin Luther, the man who initiated the Protestant Reformation. When he was asked to withdraw his statements, he stated: here I stand and that is how it is.
For the Leica M9 the same applies.
The Leica M9 is the only camera that allows you to print directly from camera image to inkjet printer without elaborate image manipulation. The raw image is already of excellent quality with high definition that directly reflects the quality of the lenses. In many cases the dcraw algorithms suffice to extract the best quality. If you want the shortest route from camera to printer, then the M9 will provide you with this.
The great value of the M9 is the concept of compactness, simplicity, direct vision, lack of lowpass filter and the rangefinder/manual focus mechanism, coupled with excellent engineering and accurate assembly. In our basically imperfect world, we have to accept that every advantage implies a certain disadvantage.
To cope with the disadvantages is a better approach than to complain about them. Mastering the M9 is a route not without difficulties, but inherently a pleasant quest and the results are most satisfying. Exaggerating the complications is as wrong as denying them. The M9 should be approached as a competent instrument with a strong character and philosophy, but do not neglect the limits. The camera deserves it!