Is the Internet Ready for Video?

As someone who makes a living providing digitized video for on-line applications, I have to let you in on a little secret: despite what you may have heard, video on the Internet is not quite there yet. Download times are too long for traditional multimedia video content. Image and audio quality of streaming video aren't high enough to satisfy anyone other than the technically enamoured. And multicast techniques are still experimental and available to only a few.

Yet I'm a firm believer that video will be moving onto the Internet. The combination of a widely available interactive distribution medium with the proven power of video is simply too powerful a capability to give up on merely because of temporary technical limitations. And just because it is not yet fully realized doesn't mean your can't be leveraging your video skills into the new world of the net. But before we get into futures, let's look at where things stand today.

As with most issues on the Internet, the root problem is bandwidth; video uses a lot, and the Internet supplies comparatively little. While new technologies like cable modems and xDSL telephone connections--with speeds of 3-30 Mbps--are hyped by their respective suppliers, most Internet users are stuck with a 14.4K or 28.8K modem connection. Those are very slow speeds when it comes to dealing with the bandwidth requirements of video.

Even the highly compressed content that has come to be accepted as multimedia video consumes 2.4M bits/second, more than 80 times the bandwidth of that 28.8K modem connection. While some would have you believe that higher bandwidth connections to the Internet are just around the corner, high-speed access outside of major metropolitan areas is likely to be quite a few years away, as building on existing telephone and CATV infrastructure really only works in relatively densely populated areas. Look at the poor track record of ISDN availability to see how quickly network upgrades (don't) occur.

But in testimony to both the genuine and hype value of video, bandwidth limitations have not stopped people from putting video on today's Internet. The first way this was done, and still the most commonly used approach, is to simply place the same type of video content used on CD-ROM's--QuickTime, AVI, and MPEG-1 files--onto the net for downloading. This has the advantage of leveraging existing content, but the major disadvantage of requiring as much as an hour to download a 60-second video clip. Unless the content is particularly compelling, that's asking for a great deal of patience from consumers. The only areas that have achieved much success are popular media like clips from Hollywood blockbusters and so-called "adult" content.

To reduce download times, video destined for the Internet typically has its frame rate reduced to 10 fps or less and its size reduced to a 160x120 pixel window or even smaller. New compression tools specifically aimed at the net, such as WebMotion (Terran Interactive, 408-278-9065, <http://www.terran-int.com/>), help make the job easier, but this is still going back to the quality level of the early days of CD-ROM video and "dancing postage stamps," an image that the CD-ROM world has finally left behind.

To avoid the lengthy download times of traditional multimedia video, numerous companies have been working on streaming video solutions. Rather than downloading all of the video data to your local hard disk and then playing the video from there, streaming sends the video data and plays it at the same rate, using buffering to accommodate the variable delivery rate of the Internet. This is certainly a better approach, but requires intense levels of compression--as much as 500:1--to make it close to practical at common network connection speeds.

Drawing on existing video-conferencing technologies--such as H.263 video compression--and reducing audio sample rates to 8 KHz and below, it is possible to stream video over Internet connections as slow as 28.8K. VDOnet Corp. (415-846-7700, <http://www.vdo.net/>) was one of the first to provide true streaming video at modem speeds using their VDOlive software. However, video and audio quality is poor at these low speeds, limiting just how useful the product is. One common use for streaming video is as a preview of a longer, higher-quality video clip that can be downloaded.

Apple has taken a hybrid approach with their QuickTime technology. Rather than creating a new streaming format--forcing existing QuickTime content to be recreated--they've added pseudo-streaming to QuickTime in the form of a Netscape Navigator plug-in. As the video clip is downloaded from the server, the plug-in monitors the download rate and the data requirements of the specific clip to know when it can begin to play the video. This doesn't provide immediate playback as the video is downloaded--as in true streaming video--but it does allow you to see the video before the download is complete. Used with appropriate levels of compression, the effect can be satisfactory.

But all that's the bad news. The good news is that the benefits of Internet video provide ample motivation for solving the technical issues. Many of the big players on the Web have large existing catalogs of video content that is ready to market once the problems are solved. These potential of these large players can help attract the capital needed to upgrade our network connections. And the power of video as an entertainment, communications, training, and motivational medium is certainly well established.

The Internet is also more than just the World Wide Web, and Internet video is more than watching video on a web site. Video conferencing via the Internet is successful simply because the demands for video and audio quality are lower. Emerging multicast technologies like the MBONE and QuickTime Live provide real-time streaming video feeds from major events. They are not interactive in the way we think of the Web as being interactive, but they build on Internet technology to provide a new way of delivering video.

And perhaps even more important than the Internet are the many "intranets" being built inside corporations. Because an intranet is contained within a company, bandwidth limitations are greatly reduced. While it may drive the network administrator crazy, upgrading a corporate network to higher speeds is certainly less expensive than upgrading the world's telecommunications infrastructure.

Intranets delivering 10Mbps Ethernet speeds or better to the desktop, with suitably higher speed network backbones, are fully capable of supporting high-quality streaming video. Both video conferencing and training are natural uses for such a capability. There are a number of companies, such as VXtreme (415-614-0700, <http://www.vxtreme.com/>), which specifically targeted the intranet market with their products but are now moving it out onto the Internet.

Much of the power of these networks, Inter- or intra-, comes from the idea that while the network provides a transmission means, both ends are under the control of a software developer. As such a developer, I can create software that will both generate the content and display it. If I can convince you to download my browser plug-in, I can provide my own unique type of content to you. Compare this to the transition from NTSC to HDTV where broadcasters, equipment makers, cable operators, the FCC, and hundreds of others must all agree on one method before anything can happen. Or not happen.

This freedom to experiment inexpensively provides an opportunity for people other than a few large entrenched companies to try their ideas. Such experimentation has led not only to the many methods of delivering traditional video content we've already discussed, but also to interactive media that draw on video technology but then move beyond it, taking advantage of the inherent interactivity of the Internet. The desire for Interactive TV was valid, it just tried to use the wrong technology.

The technical problems of limited network bandwidth are real, but so are the benefits of video on the net. If you want to get involved with it now, be prepared to experiment, hit dead-ends, backtrack, and endure some real frustration. But also know that you will get the satisfaction of being a pioneer in an area that will take all the power of video as we know it today and move it to the next level.

Bob Currier is President of Synthetic Aperture, a multimedia production company specializing in digital video and QuickTime VR. He also serves as Sysop of the Macintosh Multimedia Forum on CompuServe.

He can be reached at rcurrier@synthetic-ap.com. Be sure to visit the Synthetic Aperture web site at <http://www.synthetic-ap.com/> for more tutorial information, sample content, and information on new media services.

This article orignally appeared in a slightly different form in the November 29, 1996 issue of TV Technology magazine.