Codecs: Compression versus RAW
Another practical matter in discussing the ultimate camera is image compression. Just about all current digital cameras use some sort of codec (short for compression/decompression) to convert the very high data rates of raw video into a more compact form for storage. Arguably, film is compression-free because it’s a photochemical process. But this advantage is diminished when you make prints and inter-positives from camera original negatives because quality suffers with each generation.
The quality of the final image can be compromised by too much compression. For example, standard definition DV can look pretty good to the raw eye even though it’s compressed 5:1, because standard definition video doesn’t have a lot of detail to begin with. Once you get into HD (and up) the raw data rates jump dramatically. For example, raw 1080/24p video at 10-bits would require around 120 megabytes per second to capture. That exceeds the capabilities of most currently available solid-state memory technologies and would require a small, mini-bar sized rack of hard drives to capture.
In my experience, cameras that use codecs like HDV, H.264, and AVCHD can be a bit harsh in their compression. You get artifacts like blocks and noise in the image, and color fidelity is flattened. I prefer codecs like DVCPROHD, ProRes, HDCAM, XDCAM EX, and REDCODE, which do a much cleaner job of compressing HD and up into manageable file sizes while retaining high quality.
Figure 5. A Canon EOS 7D DSLR camera offers the smallest professional form factor currently available.
REDCODE, and codecs like it, have a nice additional feature: RAW capture. RAW records the raw data captured by the image sensor without any specific gamma or color bias. As with a raw film negative, you have a much wider degree of control over the final image during post-production. That’s also why all professional DSLR still cameras offer a RAW format in addition to the typical JPEG-compressed modes. You’re always trading off efficiency and storage requirements against image quality.
The ultimate camera would require a few more quantum leaps in memory technology. Let’s say we get a solid-state memory card that’s as small as a coin with 5TB of space at 600 MB/second read/write speed. That should be adequate to capture 4K and up uncompressed in RAW with lots of recording time. That technology might sound a bit far-fetched, but compare the capacities and speeds of Compact Flash cards of a few years ago to those of today. The ultimate compression codec (or, really, the ability to avoid compression altogether) could be right around the corner.
Form factor: Compact versus Shoulder Mount
Form factor is an important consideration. How much camera do you want to lug around?
Figure 6. A Pace Fusion3D camera used by filmmakers like James Cameron.
Even a few years ago, camera rig size contributed significantly to the perceived professionalism of a camera crew: the bigger, the better. But HD prosumer cameras and new HD DSLRs have thrown out those rules. It’s hard to gauge camera quality from size alone, whether it’s a shoulder-mounted behemoth or a device you can easily hold with one hand. I like the convenience of a really small camera when I’m on a family trip, but there is something to be said for a really ”big gun” when you need it.
Electronics and memory cards are so miniaturized now there’s no practical disadvantage to a tiny camera. The real limiting factor is optics. There’s a reason why DSLRs and film cameras have huge lenses: They provide the image quality and light sensitivity critical to professional cinematography. A compact point and shoot camera (like the Flip HD) or even a camera phone can compress and store high-quality HD video but not photograph it. Until a major breakthrough is made in lens design technology, the ultimate camera needs to be at least as large as a DSLR to permit mounting high-quality cinema lenses.
3D versus 2D
3D movies first appeared in theaters in the 1950s with the red and blue glasses but they never really caught on beyond the initial curiosity factor. 2009 is supposed to be a big watershed year for digital 3D, with major 3D movies like Pixar’s UP and James Cameron’s upcoming Avatar.
3D production is difficult for filmmakers because they have to decide up front to capture every scene with two optically interlocked cameras. That adds a significant layer of expense and technical complexity over 2D cinematography. CG animated movies are more straightforward to repurpose for 3D because everything is already laid out in a 3D animation program.
I checked out the IMAX 3D version of Harry Potter and the Half-Blood Prince. It looked amazing even though it wasn’t shot in 3D. The 3D effect was created in post-production by carefully slicing up the frame into foreground and background elements to artificially create depth. The movie switched back to 2D after about 10 minutes and never returned to 3D, which was kind of a disappointment.
Figure 7. 35mm film audio tracks, from left to right: Sony SDDS, Dolby Digital, analog Optical, and DTS time code.
The 2D to 3D process is very expensive, although I didn’t mind leaving the glasses off. Even George Lucas has shied away from 3D-icizing and re-releasing the original Star Wars trilogy because of the expense involved. 3D may well come into the home too but it will probably be years before there is sufficient content to justify it as a complete replacement for 2D.
3D probably won’t catch on for all mainstream movie production until the technology exists to make it viewable without glasses. Even the best modern 3D technologies can fatigue your eyes, especially over the course of a 2 hour feature. Someday, perhaps we’ll get a holographic display that doesn’t require glasses to view, but I think this is still quite a few years off. So for the moment, I’ll say our ultimate camera doesn’t need to be 3D.