In the last few months, the quality and diversity of Internet video has soared. Sites like YouTube have made it easy to watch user-generated content and dubbed TV clips. The old-line media is going online, too. ABC just began offering episodes of four current shows online as an "experiment," and Fox will soon make 60 percent of its shows available for download the day after they air. As Internet video has grown, so has high-definition television. Big-screen HDTVs have now hit the mainstream: Costco has 42-inch HD plasma panels starting at $1,900.
The next logical step, it seems, would be to stream high-definition video via the Internet. In the long term—10 years or more—perhaps. But for the foreseeable future, that won't happen. That's because the very properties that make the Internet great make it a lousy video-distribution network, especially for the high-def era.
The Net doesn't play favorites. All data packets are handled the same way; none get preferential treatment for faster delivery. That's fine for nearly every type of traffic. A fraction-of-a-second delay due to network congestion has no practical impact on an e-mail or even a 100-megabyte program download. But the same hiccup can mean the difference between premium and intolerable video. The various broadcast systems that handle traditional television distribution, on the other hand, are closed networks designed specifically to ensure smooth video delivery.
It is possible to change how the Internet works. New routing equipment is smart enough to analyze the contents of data packets nearly instantaneously and to assign each one an appropriate priority level. But installing these routers across the entire Net would require a massive infrastructure upgrade, and it would require the companies that run the Net to come to an unlikely consensus on how to hand off data between their systems.
Money is a better motivator than technological idealism. A more likely scenario is that the Internet will eventually prioritize traffic based on the customers who send and receive the data, with those who pay more getting faster service. (Network providers like Verizon may also give preference to their own video services.) This would improve video delivery for big players like the TV networks but would probably shut out low-rent sites that distribute unique, Web-only offerings.
With or without a fast-tracking system, the Net will need a lot more bandwidth to deliver standard-definition quality TV, let alone high-def. Today's Web video looks rather good because we see it on very small screens. In reality, most Web video is lucky to reach VHS quality at a time when your grandma is leaving her old VCR on the curb. ABC's 700-by-394 pixel videos may look decent in a small window on a laptop but will likely be unwatchable on a 1280-by-768 pixel, 50-inch plasma TV.
Streaming standard DVD-quality video requires a data rate of 1.5 to 5 megabits per second (depending on whom you ask). That's far more bandwidth than most homes with broadband receive. Delivering HD quality requires 5 to 7 megabits per second. But even that seems stingy compared with the roughly 19-megabit-per-second rate of network HDTV broadcasts—an amount of bandwidth that you can get for free with a set-top antenna. Keep in mind that most homes have more than one TV—if you wanted to show different programs on different TV's, the bandwidth requirements would be two, three, or four times as high. This amount of bandwidth will be possible some day. Cablevision and Verizon already offer up to 30 megabits per second to a smattering of communities around the country. However, it will be a long time (if ever) before such services reach the same number of people who can already receive HDTV via antenna, cable, or satellite.
While streaming shows the weaknesses of the Internet, downloading plays to its strengths. Bandwidth requirements are far less critical with a download-and-play model than with real-time streaming, and downloading allows you to watch what you want when you want. A company called Akimbo recently started offering high-definition TV downloads from a set-top box. In April, AT&T announced that it would bundle the Akimbo service and movie downloads from MovieLink into its upcoming Homezone package, which will offer high-speed Internet service, video downloads, and satellite TV from DISH Network via a single box.
But the download-and-play model isn't a real solution to the Internet/HDTV intersection problem. The more time-sensitive a program is—sports, the news—the less sense downloading makes. The same goes for massively popular shows: Who wants to wait a day to download American Idol? Plus, why would broadcasters want to support 30 million simultaneous, bandwidth-hogging downloads when they could send out a single broadcast signal instead?
So, how will the Web and HDTV get reconciled? It's instructive to look at what is happening with cell phones. Wireless carriers now provide cell-phone videos as individual downloads, which increasingly threaten to swamp their data networks. In response, companies like Verizon Wireless are building separate digital broadcast networks to offer television channels specifically formatted for cell phones. These channels won't have high-enough resolution for living room TVs, but they'll offer a better picture than what cell phones currently receive. Wireless carriers will reserve data downloads for niche offerings that customers order individually.
That's a likely model for the big-screen TV system, too. Set-top boxes like TiVo already supplement traditional television with download capabilities, and such options will undoubtedly grow. In most cases, the quality will be lower than what comes in over broadcast. But as the popularity of today's low-fi Web video already shows, people will trade quality for convenience in some cases—to catch a missed TV episode that won't hit DVD for a year, or to watch Internet-only gems like the hilarious news blog Rocketboom.
But given the choice between the Web and traditional broadcasts, most people will pick the higher performance (and instant gratification) of traditional TV. And don't just assume that the Web will catch up in a year or two. While Internet bandwidth and video quality will improve, so will our standards. There's already talk of increasing the color gamut and resolution of HDTV. Television quality is a moving target that the Internet already fails to hit.