The magazine of the Melbourne PC User Group A View from the Crow’s Nest The Future of Digital Photography Mike Chambers

In 2002, we saw another year of incredible growth in digital photography. The number of Internet connected households with at least one digital camera grew to from 21 percent to over 31 percent by year's end, in total 50 percent growth from December to December. With the Christmas season past and the northern hemisphere holiday season approaching, digital camera manufacturers now have their minds turned toward the next wave of cameras. Already their marketers are telling us which features we just can't live without this year. So this is a good point to stop and consider impartially where digital photography is heading and the types of capability we'd like to see manufacturers bringing to market.

Megapixel Madness: 
When is Enough Enough?
 
For the past four years we've seen a one Megapixel (Mp) per year increase in the prevailing Mp standard. 2001 was the year of the 4 Mp camera; 2002 saw 5 Mp emerge as the prosumer standard. Clearly most manufacturers have been working on the premise that "more is better." Until now, I think they've mostly been right; however that may be changing.

Digital cameras long ago passed the capability of most monitors to display their images at full size. More Megapixels today means larger print size. But how are most digital photos being printed? They aren't. Most shots happily live their nearly immortal digital lives never leaving the confines of digital media (hard disk, CD-ROM, or Web). Remember that this is a selling point of digital photography: you don't have to print everything, just the shots you really want to print. So the intuitive yet somewhat surprising fact is that more Megapixels won't help most of our shots at all. But they may help our "best shots".

Second, we need to consider the destiny of the shots that do get printed. Not surprisingly, most are printed at 4x6 or 5x7. A much smaller number find printed life at 8x10 and the number printed larger is minuscule. Although estimations vary and have changed over time, I hold that a 4x6 print requires a 1.3 Mp image; a 5x7 print needs 2.1 Mp and the venerable 8x10 print requires 3.3 to 4 Mp. Five Mp and beyond enable us to print at sizes that few of us ever will.

This isn't new. Few photographers would argue the supremacy of a medium format film over 35 mm. Why then does 35 mm dominate the consumer market? For the vast majority of uses and situations, 35 mm film is sufficient, providing adequate resolution, convenient size, widely available processing and lower cost. The same facts hold in our emerging world of digital photography.
 
Sometimes less is more.

What is the downside to more Megapixels? Storage and management. Recent years have seen a 500 percent increase in digital photo sizes. Today's 5 Mp cameras are pushing the 2 MB per JPG file boundary. High-end 6 Mp cameras are nearing 3 MB per JPG. These file sizes have implications at every stage of the digital photography process. A 256 MB memory card holds about 250 shots from a 3.3 Mp camera. The same card holds about eighty-five 6 Mp shots. While the Megapixel growth has not quite doubled, the space required has more than tripled. An afternoon's outing (200 shots for me) has grown on my hard disk from 200 MB to 600 MB - which is beginning to nudge the capacity of CD-Rs to hold a single afternoon's work.

Processor capacity is also affected. The digital darkroom is one of the great benefits of digital photography. Larger photographs, however, require exponentially larger amounts of RAM, I/O speed, and processor power to perform image transformations. For example, one macro I use took about 10 minutes on a 2.1 Mp Nikon D1 image. It takes over an hour for 6 Mp Nikon D100 image. Time to upgrade, again!

Do more Megapixels have value? Yes, I think they do. However, much of the need for higher resolution has already been accomplished by our journey to the 5 Mp level. For most people and most uses, resolution beyond 5 Mp is costly and may be unnecessary. What will continue to drive our Megapixel lust is the hope of capturing that once in a lifetime image that we'll print 4 metres wide and hang in the dining room. Hope springs eternal!

Better Megapixels 

The past five years have seen the digicam market walk a path of product convergence in which technologies developed for the high-end professional market crept their way into the high-end consumer market, creating the prosumer market segment. I think 2003 will see a fork in that road. We may see professional digital photography push toward the 20 Mp boundary and beyond. In the consumer world, I think that we should, and perhaps will see a shift toward better Megapixels, rather than just more of them.

The foreshadowing of this trend is already with us. In the last two years we've seen movement toward much improved lens systems (like the Carl Zeiss lenses appearing on a number of Sony's cameras). No matter how many Mp you have, they can only record what the lens is capable of capturing. Today, there is a clear differentiation at a given resolution between higher and lower quality lens systems.

The move toward better Megapixels can be seen in two other developments. Year 2002 saw the long awaited release of the Foveon X3 sensor. Although it is a mere 3.5 Mp sensor, dwarfed in the land of 5 Mp and 6 Mp giants, its revolutionary technology gives it the effective resolution at least a 7 Mp image, and perhaps that of a 9 or 12 Mp camera.

Traditional CCD sensors capture Red, Green, and Blue values images in a mosaic grid (a colour filter array). Often four sensor locations are used to generate a single pixel value. These values must then be interpolated to make up for the detail lost at each sensor point in the colour not sampled. (see Figure 1)

Foveon's X3 technology captures Red, Green, and Blue measurements at differing depths of the same point, eliminating the need to interpolate the missing detail. The resulting image captures greater detail at greater fidelity. (see Figure 2)

Today, the Foveon X3 technology is found only in the elusive Sigma SD-9.

A second sign of better Megapixels to come is Fujifilm's announcement of its SuperCCD SR technology. For some time now digital photographers like my friend, Max Lyons, have been stunning the world with digital still photographs of incredible dynamic range. They achieve this by careful use of a tripod to take multiple photographs of the same scene at different exposures and then digitally combine the photographs in the darkroom. (Michael Reichmann's "Luminous Landscape" Web site describes this "digital blending" technique in greater detail.)

Fuji has taken this technique into its technology down to the sensor level. Their SR sensors use primary and secondary photodiodes at each photosite on the sensor to measure bright highlights and dim shadows and produce a composite image that captures detail out of both extremes in the same shot. Fuji estimates that its SR sensors will deliver a 400 percent improvement in dynamic range compared to conventional sensors. (See Figures 3 & 4)

This approach has an analogy in the audio world, where crossover networks separate low and high frequencies and produce them in woofers and tweeters respectively. Thus Fuji is touting their 4th Generation CCD technology as "High Fidelity Photography."

The SuperCCD SR technology holds out the hope that high dynamic range photography can move beyond still-life digital photos to capture friends and family in more natural surroundings and lighting.

Historically, Fuji has taken a lot of criticism about its SuperCCD marketing, particularly around its resolution doubling. (Actually I've been quite impressed by the results of using that technology in the Fuji S2 Pro, particularly for large size portraiture.) In their SR technology, I think they've made an advance to be proud of. In both cases, Fuji's focus has been on delivering better image data at a given level of physical resolution.
 
Both Foveon and Fujifilm seem to be addressing the challenge of better Megapixels with the same concept, but vastly different approaches (Figures 5 & 6). Both technologies aim to improve image fidelity by making the digital sensor more like its film ancestor.

Technologies like these really are significant developments in the evolution of digital photography. They improve the experience by increasing the percentage of shots that are worth printing, enlarging, and manipulating. Whether these specific Fuji and Foveon technologies will take the market by storm remains to be seen. But developments like these are confirmation that the future of consumer digital photography will follow a path of better Megapixels, not just more of them.

The Dream Camera

For several years now I've prodded our members on the Coolpix990 group to envision the ideal digital camera. Over time, I think we've had an effect, as many of our wild dreams in 2000 began appearing in new products in 2001 and 2002. As anyone who has shopped seriously for a digital camera will agree, the ideal camera hasn't hit the market yet, and likely never will. There are so many variables, so many differences in the way we use the cameras, that my ideal camera may be a little less perfect than your dream camera. So in Figure 7 I've tried to focus on just a few of the major technological frontiers that should challenge the digital camera engineers in 2003.

Beyond the dream is the fantasy: Wireless network integration would mean direct to desktop/laptop/palmtop shooting without a memory card. Pseudo-optical viewfinders would finally overcome the issue of LCD visibility in bright light. The next frontier in digital photography may be completely beyond the camera: software and processes to manage thousands of photographs. But these challenges we will leave to another issue!

About the Author
Mike Chambers has been working with digital imaging technology for over 25 years. He founded and moderates the Coolpix990 group, the largest digital photography group of its kind on the Internet. When not travelling the world, he makes his home near Houston, TX, USA. Write to him at photos@mikechambers.net.

Reprinted from the March 2003 issue of PC Update, the magazine of Melbourne PC User Group, Australia

[ About Melbourne PC User Group ]