June 01, 2011 — By the dawn of the new millennium it was clear to all but a handful of diehards that the upside potential of digital imaging was far greater than that of chemically-based image capture, also known as film. One could, and did, argue about when, not if, the digital revolution would be upon us in full force. And as it turned out, the great transition occurred far more rapidly and completely than many experts, including yours truly, predicted at the time. It was undoubtedly a seismic event that completely transformed the way we shoot, enhance, store, share and print pictures.
Nevertheless, the first iteration of digital cameras was pretty much digital versions of existing film camera designs with CCD and CMOS image sensors instead of film inside, and LCDs on the back for composing the picture and/or viewing the captured images. The DSLR in particular was, and still is, largely based on the optical-mechanical SLR paradigm—a body housing a mirror box with a flipping reflex mirror mounted in front of a focal-plane shutter and surmounted by a pentaprism (or mirror prism) that presents the image formed by the lens and reflect upward as a right-side-up, laterally correct viewing image in the eyepiece.
In many ways, today’s DSLRs, among the most highly developed digital cameras on the planet, have been an unqualified success—technologically, functionally and in their salability. But we are now entering an era where the unending quest to improve their versatility and performance parameters is running into challenging mechanical and optical limitations that are inherent in their basic design. In other words, going forward, devising cost-effective mechanical and optical solutions aimed at improving the breed is likely to be increasingly difficult, and achieving significant advances in everything from burst rate to HD video performance is much more likely to be accomplished electronically. Indeed, this scenario probably holds true for all classes of digital cameras. To give you a clearer picture of the electronic future in camera design, let’s step back a bit and take a look at some interesting cameras in current production.
The Sony SLT A55: Harbinger of Future SLR Tech?
Frankly, I was skeptical when I first perused the specs of the Sony A55. Its signature feature is a fixed non-flipping, semi-transparent mirror placed directly in the light path in between the lens and the CMOS sensor. Sony calls this Translucent Mirror Technology and it’s functionally similar to the non-flipping pellicle mirrors Canon installed in the landmark Pellix in 1972 and later, the much-improved Canon EOS RT autofocus 35mm SLR produced from 1989–1992. However, in the Canons the semi-transparent (two-way) mirror reflected part of the incoming light up to a conventional optical prism SLR viewfinder and allowed the rest to pass straight through to the film. In the Sony A55 the light reflected upward by the Translucent mirror is directed to the AF sensors, and the camera uses an eyelevel electronic viewfinder similar to that found in mirrorless SLR-like cameras such as the Panasonic Lumix GH2.
Why use a Translucent mirror in a camera without an optical viewfinder? Is losing about 1/3 of a stop of light and a small but measurable amount of resolution and contrast worth the trouble when the main advantage is directing more light to the AF sensors? Well, after field-testing the A55 I have come to the conclusion that the overall design of the camera not only makes sense, it also points the way to the future—a future where camera operations are simultaneous rather than sequential, as they must be in a conventional DSLR with a flipping mirror.
By eliminating the bulky mirror mechanism in the A55, Sony has achieved a compact, quiet, lightweight, mainstream- priced, interchangeable-lens camera that accepts Sony Alpha-mount DSLR lenses, functions much like a DSLR, has a maximum burst rate of 10fps (top in its class) and is capable of shooting 1080p HD video complete with phase-detection AF and focus tracking. The big advantage of phase-detection AF, the system commonly used for shooting still pictures with DSLRs, is that it is a true electronic rangefinding system that swiftly locks focus on the subject. The contrast-comparator system used to achieve video AF in current DSLRs focuses via the live view image, moving the AF motor in the lens back and forth until optimum contrast, and thus proper focus is attained. This system is precise all right, but fairly slow, often taking about one to two seconds to achieve proper focus, as you can see in your videos. At certain shooting distances, the clicking of the AF motor as it moves back and forth is also audible on the soundtrack. None of these defects is present in videos shot with the A55.
The Sony A55 and its lower-priced sibling, the A33, represent Sony’s entry into and demonstration of the benefits of Translucent Mirror Technology, but they won’t stop there. At Photokina 2010, the company showed a prototype successor to the Alpha 700, an enthusiast class camera based on Translucent Mirror Technology that is slated to sell for over $1000. Befitting its price class, we expect this as yet unnamed model to feature more rugged cast alloy construction, higher resolution than even the 16.2MP A55, plus more sophisticated AF and auto-exposure systems. The camera, which is expected within a year, will be based on “developing key devices on which Translucent Mirror Technology Depends,” an intriguing statement that points in several possible directions.
Part of the problem in improving the performance of optical-mechanical DSLRs is overcoming the cost-benefit-ratio dilemma. It’s easy enough to make a high-end DSLR that can shoot 8fps, but a camera that can do that (the Canon 7D for example) sells at a street price of about $1800, body only! On the other hand, achieving high-performance full HD video capability with pro-caliber AF and focus tracking requires either a design concept breakthrough or a vastly upgraded contrast AF system. The latter can be achieved with a much faster operating sensor and image-processing software plus an upgraded AF motor, but the solution is largely electronic rather than mechanical. The problem with mechanical systems is that they follow a much slower development track while electronic systems, for the most part, follow Moore’s Law, which states that capacity doubles about every 18 months.
Yes, there will be modest mechanical improvements in optical-mechanical DSLRs over the next two to three years, but probably no major breakthroughs at each price point due to the aforementioned mechanical limitations. Today’s mainstream DSLRs can shoot bursts at around 3–4fps (the Canon T3i gets up to 3.7fps) and to get twice the firepower you have to double the price. This doesn’t sound too bad until you consider that any number or $200 point-and-shoots can attain 10fps. One of the fastest-firing under-$1000 DSLRs is the Sony A580, which gets up to 5fps, a figure that will probably be emulated at the $850 level by its competitors. As for contrast AF, we will probably see an incremental increase in its performance parameters over time, which will have advantages not only for shooting HD video, but also in live view focusing and viewing, a technique that is becoming increasingly popular among DSLR shooters. Right now, focusing using this method can take two to three seconds—hardly ideal for capturing action subjects. Another advantage of EVF and SLT cameras when shooting video is eye-level viewing—you can’t use the eyelevel finder when shooting video with a DSLR.
In Search of a Big, Bright, Low-cost Viewfinder
Another virtually insoluble cost-benefit problem encountered with the optical DSLR viewfinder is achieving a big, bright, high-magnification view with good eye relief at a reasonable price. Middle tier-cameras like the Nikon D7000, Canon 7D and Pentax K-5 all use solid glass pentaprisms, and their viewfinders are excellent in every respect. Mainstream DSLRs in the $600–800 class all use mirror prisms which are lighter and far less expensive but not as bright, and their coverage and magnification (O.8X compared to 0.95X or better) is typically much lower, giving you that dreaded “looking down a tunnel” effect. It’s just another example of how the inherent limitations of the classic SLR design are extremely difficult to overcome.
One obvious solution is to improve electronic viewfinders, and since EVFs are (with the exception of condenser lenses) based largely on the silicon, semiconductor consumer electronics paradigm, at some point in time we are likely to see EVFs that rival what can be done optically. It will be quite some time before we put our prized DSLRs on the shelf next to our Underwood typewriters, but we are beginning to see how the fundamental limitations of moving-mirror cameras will be an impediment to further development, and it is equally clear that most of these barriers don’t apply to the electronic alternatives. It is noteworthy that when DSLRs are used in HD video mode, all the things that make them SLRs are basically overridden, and the camera functions like a mirrorless device. If that’s true, what does it say about moving-mirror technology having reached its technical limits?
At this juncture, it seems inevitable that mirrorless interchangeable lens digital cameras are ripe for development and this is already underway in such cameras as the Olympus E-PL2 and Sony NEX-5. Free from the limitations of electronic-mechanical designs, they will advance rapidly as the speed of semiconductors increase, and they will feature simultaneous rather than sequential functions including viewing, capture, AF and much more. With the dramatically increased clock rates of future sensors, many things will be possible, and some of these advantages will trickle down to DSLRs and point-and-shoot cameras as the throughput speeds of image sensors and image processors increase dramatically over the next few years. Without the “gateway factor” imposed by the flipping mirror, the speed of contrast-detection AF will improve noticeably as the refresh rate of sensor and the speed of image processors improves, though the back-and-forth speed of present AF motors may impose some limitations. Today, no consumer priced large-sensor camera can shoot 1080p HD video continuously at 60fps like some camcorders, but this capability is coming soon. As the silicon chips get faster many things are possible, but they can’t happen if the mechanism gets in the way.
As for point-and-shoot cameras, cost is a limiting factor in the most popular models, but less so in lower production models aimed at enthusiasts. The general perception is that interchangeable-lens cameras are the best choice for serious photography. But as performance improves, ultra-zoom point-and-shoots with such features as Auto HDR and layering for low-light capture will certainly provide excellent real world performance. In the meantime, enhanced sensor capture speed and performance is the rising tide that will lift all boats, so the digital future surely looks bright.
Jason Schneider is best known as a prolific writer and editor on all aspects of photography. He began his career at Modern Photography in the late 60s and in 1987 signed on as editor-in-chief of Popular Photography, a position he held for nearly 16 years. Considered an authority on the history of camera design and technology, he has written three books on camera collecting, is an active contributor to leading photo magazines and Web sites, and is Senior Editor of Photo Industry Reporter, an industry trade magazine.