On the Modularization of Digital Cameras

Motivation:

1. As we know, the essence of a camera, is the optics, namely the lens or a lineup of lenses that is compatible with it. A couple of top brands in the industry manufacture quality lenses, to name a few, Leica, Zeiss, Nikon and Canon etc.. Each of them have some perfect designs in some particular focal length, but none of them are perfect at all types of lenses. For instance, the Leica Summilux 35mm f/1.4 is the best model in the ultra-fast standard lens class while Nikon is proud of the best 24mm f/1.4 auto focus lens ever made. Unfortunately, these top brands are extremely exclusive — a Leica M Mount lens, in most cases, can hardly be adapted to a Canon EF mount and vice versa. I personally want to use the best lens out there at each of the common focal lengths no matter what its brand is. A standard universal mount would therefore be ideal, which is obviously not going to be discussed by these hostile optical giants.

2. Experienced photographers know how to spend $2000 smartly: one would invest in lenses rather than in the body. For example, a “24-105mm L” on a EOS 600D always yields sharper images than a “18-135mm” on a 7D. In addition, as the sensor technology progresses, both 600D and 7D would be gone in a couple of years, the best example could be the Kodak Pro 14n, which was among the best DSLR body ever made in terms of pixel density, continuous shooting and durability by 2003 and therefore priced at about $5000, but now this body can hardly excel in any essential aspects, while any Nikon F mount lens purchased at that time for Pro 14n would still sell at the same price. However, when we replace a body with a new one, we not only throw away the outdated electronics and sensors but also the mechanics and metal frame which by the time of the retirement may still be in excellent condition, and the new body is meant to be abandoned with its mechanics and metal frame again once the next generation of sensor technology comes. So the rational way to minimize the wasted investment in body provided that we always want the newest technology, is to minimize whatever is replaced each time when we update our system.

Solution:

So my response to the aforementioned issues is to re-define a camera as photographic system, which is broken into modules more than two (lens + body).

Stage 1: Lens + sensor module (including the mount) + body module (including all the mechanics and electronics other than the sensor chip, such as the image processor, LCD, etc.). I thought this is a new idea, but the market has actually seen one example of this stage of modularization — the Ricoh GXR system. Although this novelty originally only comes with some “lens-and-sensor” modules and a “body module” without any sensor, recently Ricoh add a Leica M Mount module including just the CMOS and the mount ring , thus with which users can mount any Leica M mount – compatible lenses.

There is still much space left to improve the system concept according to our motivations aforementioned. To name a few, the current GXR M mount module only works for M mount lenses, and the robust body, which itself alone is priced as high as a high-end point-and-shoot camera, already appears outdated to some extent and therefore the expected system update will guarantee some waste.

Stage 2: Lens + mount adapter module + sensor module + body module (including all the mechanics and electronics other than the sensor chip such as the image processor, LCD, etc.). At this stage, we should be able to mount whatever lens we want by just purchasing the corresponding mount adapter module. So we should have already realized “motivation 1″ at this stage: when we want more pixels, we get a new sensor module; when we want to switch from Leica lenses to Nikon optics, we just screw a new mount adapter on. All the modules should be standardized, so that we expect at these stage, the entire CCD/CMOS industry, not only Sony and Canon etc but also some scientific CCD manufacturers, will provide specialized sensor modules for the photographic industry. The advantage of this design is obvious: some CCDs may specialize in extremely high quantum efficiency or extended linear range, or even great dynamical range which essentially make the popular auto HDR not necessary anymore.

Stage 3: similar to Stage 2, except that the body module is further spitted. Here is something really crazy: I expect the electric part and the mechanical part to be completed separated into two individual modules. Aiming at realizing “motivation 2″, we do not want to throw away the entire body when a new generation of image processors come or a larger screen size becomes a standard feature. This is because some body design, I mean the mechanical part, is really very classic: for instance, Leica M3 is still the best 135 camera ever made, and Nikon F60 is still one of the best DSLR body ever designed and copied again and again. So we want to always keep the good part and update just the electronics — the LCD, the image processor, the ROM, the I/Os etc..

The beauty lies in how to separate the electronics from the body. I am enlightened by Apple’s approach. Apple manufactures iPhone, which is already the most popular “camera” on Flickr. Behind this success, is the advantage of iPhone over all the other cameras — its much more superior CPU. I anticipate that an A5 chip for iPhone, or a NVIDIA Tegra 2 chip for Android, or a current TI chip for RIM, will still be faster in terms of image processing than whatever the generation of Digic “X” may be in two years. Simply speaking, an ordinary smartphone processor, once properly tuned by certain image processing algorithms, together with the superior ROM and cache of a smartphone, can be way more better than even a top specialized image processor invented by some optical giant which is not good at developing fast chips. So my way of designing the mechanical module of the body would be to just keep the classic design of some successful bodies except that we get rid of the complicated (sometimes not even compact) electronics inside, and expand the SD card slot into an adjustable smartphone slot, with the ~4-ish inch screen of the cellphone exposed at the back of the body. The slot should be adjustable in dimensions like a slide caliper (游标卡尺) and should be able to hold the cellphone safely and tightly. We can therefore make full use of the big touch screen, some times as large as 4.3 inches (much larger than the 3-inch LCD on most of the high-end interchangeable-lens cameras), to manually choose focus point or to enlarge apertures etc.. The cellphone talks with the other part of the camera system through its standardized fast I/Os (such as the 30-pin port of iPhone), which should already be fast enough to handle ~50MB/s data current nowadays.  The advantages of this design are countless — not to mention the superior processor, the much larger ROM and cache for faster continuous shooting, and the larger and nicer screen — the wifi function and GPS function are built in, so there is no need to buy eye-fi SD cards or to buy optional GPS modules for cameras such as Nikon D7000. Journalists can always directly upload their professional quality images to their news agencies wherever they are under 4G LTE coverage. I happen to find a similar idea in tablet industry: ASUS recently launched its “PadFone”, which is a “system tablet” with a removable self-consistent cellphone that can be put into the slot on the tablet’s screen module to be used as the heart of the tablet.

To summarize, a future cameras, shall be a re-defined photographic system with a series of specialized modules: lenses, mountings, sensors, a mechanical body with an adjustable cellphone slot, and a qualified cellphone. More competitions will thus be introduced into the photo industry; optical giants can focus on improving optics as; the technology on the cutting edge of the CPU industry can be utilized directly in cameras; all the best lenses out there can be used.

This is not crazy fiction. This can be a very good blueprint.

China doesn’t host any optical giant today, but if relevant companies can investigate into this blue print, I bet they can excel and revolutionize the photographic industry.