There are approximately 65,000 movies available in the world today. Each movie, if compressed using the state-of-the-art, MPEG-2 format, would require about 1.5 GB (gigabyte) of storage. That means the 65,000 movies will require 95 TB (1024 GB, or 1 terabyte). The largest databases in existence today are about 1 TB. This theoretical movie database also doesn't take into consideration all the news, documentary, and television that might be made available.
Current storage devices are largely mechanical. Video servers that are used for VOD are essentially large hard disks. When configured into server arrays (linked hard drives), the technicians responsible will be doing constant maintenance. Running a large video server could be a return to the late 40s when engineers were constantly replacing blown vacuum tubes in ENIAC, a gargantuan beast of a computer that the military used to calculate the ballistic paths of artillery shells. ENIAC had no memory, could handle only 20, 10-digit numbers at a time, yet needed 19,000 vacuum tubes and 175 kilowatts of power to operate. It averaged 5.6 hours of use between repairs. A server array will have similar problems with at least one failure per 1,000 hard disks, per day.
Figure #8:
Arnold Schwarzenegger as he might appear on VOD with the virtual VCR controls displayed on a teleputer. (5K .gif)
At the receiving end there is a digital set-top box which decodes the compressed video signal and sends it to the television or teleputer (telephone/television/computer) that is connected to it. Designing set-top boxes is a highly competitive business right now, and also a highly technical one. Set- top boxes contain powerful CPUs, such as a 486s with 3 MB of RAM, MPEG decoders, Dolby decoders, audio convertors, RGB convertors, RF modulators, and infrared interfaces for remote control. All of this makes a set-top box more complicated than a PC, and currently more expensive. Today's set-top boxes are being sold for about $2,000. This makes for a marketing nightmare. Industry forecasters have determined that set-top boxes have to be available in the $150- $200 range before VOD becomes economically practical. In most current VOD trials in Canada, the set-top box is supplied free to the consumer. Eventually set- top boxes will disappear into the television, the same way cable convertors originally sat on top of TV sets, before being integrated into the TV tuner.
A subset of VOD is Broadcast on Demand (BOD). This concept involves delivery of traditional television broadcasting programs in an On Demand mode. Instead of time-shifting Seinfeld or CBC's latest Witness documentary with your VCR, BOD would allow the consumer to simply call up any program within maybe a two-week release period. The Time Warner Full Service Network trials in Orlando, Florida offers a feature called News Exchange, which allows viewers to call up the news whenever they wish to see it. Also being tested in the U.S. is Your Choice TV, an interactive "best sellers rack" of television programs. Partners in this venture include ABC, NBC, CBS, and PBS. The progamming is provided after the initial network or cable release.
THE PROBLEMS
"Any new technology must be ten times better than an old technology in order to supplant it." - Eric Tsang, BADLAB.
The VOD trials that are taking place right now, are using MPEG-1 compression, mostly for economic reasons. MPEG-2 encoders are currently worth about $250,000 US. MPEG-1 theoretically delivers S-VHS quality. MPEG-2 will deliver a broadcast-quality signal equivalent to D2.
Figure #9:
MPEG-1 is only capable of S-VHS quality while MPEG-2 can deliver D2 quality. (3K .gif)
How that signal gets delivered to the home is also an issue. The cablecos have a fibre optic, backbone network with coaxial cable doing the final curb-to- home link. The telcos have a major stumbling block to overcome. Though they have a more complex fibre backbone than the cablecos, the curb-to-home delivery medium is invariably twisted-pair copper wire which has an extremely limited bandwidth. A digital, network protocol called ADSL (Asynchronous Digital Subscriber Loop) has been developed to overcome the limitations of twisted-pair copper, but it requires yet another black box decoder, along with the set-top box. ADSL also has severe limitations in terms of bandwidth. It will handle the limited interactivity of VOD, but is not robust enough to handle the sophisticated interactive, multimedia applications that will be available along with VOD. That is why at the recent CRTC hearings Stentor was arguing in favour of being allowed to piggyback onto Rogers's fibre/coax network for the final curb-to-home loop. Whether the CRTC will allow that, or whether Rogers would be interested in helping their competition, is one of the more interesting questions to be resolved in the future.
The technical problems seem daunting, even to the engineers. The major VOD/Interactive TV trials in the U.S. are months, if not years behind schedule. The much vaunted Time Warner trial in Orlando, Florida is a good example. The Full Service Network, as it's called, was built in collaboration with US West, Silicon Graphics, AT&T and Scientific-Atlanta. The trial was supposed to involve 4,000 homes in suburban Orlando. However, when it was unveiled to the press in December, 1994, it was eight months behind schedule and had only five homes hooked up to it, and they belonged to Time Warner employees. This is two years after the $5 billion project's announcement. Viacom's Castro Valley, California trial has also run into trouble and is on hold.
The financial investment in these trials is phenomenal, but the early returns have been less than inspiring. 200 homes in the Denver, Colorado area were involved in a pseudo-VOD service that used VCR's, human attendants, and a library of VHS cassettes to deliver movies in response to client's phone calls. The system was a clunky, crude, "sneaker net" version of VOD. The partners included US West, AT&T and TCI (TeleCommunications Inc). The results: an average of 2.5 films were accessed per month, per household. That's a figure that sent shivers down the spines of a lot of telecommunications executives.
In terms of content, VOD will be very similar to the corner video store model. It will be box office driven. Tens of thousands of viewers will be wanting to see the latest Arnold Schwarzenegger flick, so you can count on it being compressed and added to the library, but what about less popular material? How many of those 65,000 movies will actually be compressed for VOD delivery? You can bet It's a Wonderful Life will make the cut-off -- colourized of course. What about the rest? What about Canadian material? Canadian film critic Geoff Pevere recently pointed out in Take One that Don Shebib's 1969 breakthrough film Goin' Down the Road, which is "arguably the most important English-Canadian film of its era", is not available on videocassette. Will VOD make a difference? It will all boil down to economics. Are enough people interested in a particular film to warrant digitization? Robert Forget, the NFB executive in charge of CinéRobothèque and the NFB's Web site plans, estimates that only about ten percent of the NFB's collection is popular enough to warrant the cost of digitization. And what happens to material that was independently produced, outside of the major studio complexes? What chance does it have for VOD?
THE VIDEO DIALTONE
"Over the next five years, the gates will disappear. Through fibre optic cable and direct broadcast satellite, the number of channels doubles, triples, quadruples until they become virtually infinite." - Mark Starowicz, CBC-TV, 1993.
We are working in an era when the means of television production is becoming cheaper and cheaper. The three-chip, Hi-8 camera and the non-linear, digital editing suite are causing a revolution in production economies. These two technological developments have democratized the production process. It is now relatively easy and cheap to produce so-called "broadcast-quality" programming. What we may be facing is a similar revolution in the means of distribution.
The process by which programming gets to "air" has not changed in decades. Expensive, multi-million dollar facilities are used. These facilities are controlled by networks that spend millions of dollars just to get their licences from the CRTC. Once licensed they succumb to the "common-denominator" type of programming, where the number of viewers attracted determines the success or failure of the programming.
However, if the InfoBahn develops along the model of the Internet/WWW, then there will be ample opportunity for both commercial servers offering the equivalent of Blockbuster VOD, and alternative servers offering foreign films, multimedia artist work, or documentaries. There are gatekeepers on the Internet, but the cost of becoming a gatekeeper is minuscule compared with becoming one in the world of television. The people who control the file servers connected to the Internet are the Internet gatekeepers. There are tens of thousands of file servers connected to the Internet, which means there are tens of thousands of gatekeepers. Compare that to television, where a handful of professional bureaucrats ultimately determine what gets made in this country, and what gets shown. But if the InfoBahn is designed in such a way as to provide universal access, not just to the reception of information, but to the dissemination of information, then anybody with the price of a video/file server and the hookup fee could become a broadcaster.
In an address at the Canadian Embassy in Washington on March 29, 1993, CBC Producer Mark Starowicz speculated on the shape of this new "broadcast" environment. First noting that "data storage systems of the future make the word channel somewhat antiquated," Starowicz said:
"In fact, the future is visible in the present by staring at a large newsstand. Channels will parallel the magazine covers.... If there are enough people to sustain a reasonable national circulation for a magazine today, these people will be serviced by a channel within a decade. A hundred thousand people will soon be a commercially viable community for a service. Just stare at the largest periodical newsstand you can find and see the future."
A video dialtone also opens up the possibility for low-cost virtual studios. Imagine being able to send a non-linear, rough cut of your latest documentary from the edit suite in Toronto to a music composer's studio in Vancouver, at something approaching long distance rates instead of the tariffs now charged by Stentor for their VideoRoute service. Or better yet, screening that rough cut in real time on a videoconference line, stopping and starting while you discuss scenes with that same Vancouver-based composer. Video dialtone brings all this within the realm of economic feasibility. Off-the-shelf, Windows-based, ATM-compatible software already exists that will allow you to do that. All that's needed is a fibre optic ATM, switchable network, at reasonable rates.
