Digital Hubbub – IEEE Spectrum

Cutting costs by merging functions

Still, how do engineers cram what used to be thousands of dollars worth of video and computer equipment into an under-$500 box? By designing chips to do multiple duty, points out Anthony Simon, director of marketing for chip maker Conexant. For example, adding cable-modem functions to a video chip cuts between $20 and $40 from the cost of a set-top box.

And at PVR maker TiVo Inc. (Alviso, Calif.), product marketing director Ted Malone is proud of the subtle economies the company engineered into its custom disk-controller chip. The chip can read data streams from the disk surface in whatever order is most efficient for the head and then reassemble the information before handing it off to the video section.

Meanwhile, the price of hard-disk drives has put enormous volumes of storage within reach of even a run-of-the-mill set-top box. Currently, a 40-GB drive, which stores more than 50 hours of video, sells for about $80 retail and much less wholesale. Even a small fraction of that disk space can store dozens of hours of audio and thousands of digital photos.

Some of the money trimmed from hardware budgets is going to software R&D. When a hub has thousands of items available for its users to record, watch, or listen to, the conventional scrolling-list display of current set-top boxes is unusable, says Jakob Nielsen of the Nielsen Norman Group (Fremont, Calif.). Instead, a hub needs a search engine that, with only a few button-pushes, could find all movie musicals starring, for instance, Elvis Presley, or action dramas with Jackie Chan, or new episodes of your favorite home-improvement show.

The evolving interface

A first step up is an interface like TiVo’s, where you peck out the name of the show on a virtual on-screen keyboard: as you type each letter, an adjacent display of potential matching titles gets shorter until only a few choices remain. Once the right show is found, recording its episodes is a matter of pressing just a button or two.

Moxi is simplifying matters further by mapping the letters most likely to be typed next to the numbers 1 through 9 on the remote’s keypad. Typing text on a numeric keypad will be familiar to the millions of people who send text messages by cellular phone.

But this is just the beginning and there is a long way to go. As entertainment hubs start filling up with more than video, tools for browsing the material will become ever more important, notes Jakob Nielsen. Simple groupings according to title, artist, or kind of media won’t be enough. Imagine trying to find movies with Elvis in a non-singing role scattered through a list of his more than 50 films, his hundreds of recorded songs, digitized photos, video clips, and a long list of Elvis impersonators, to say nothing of navigating around people with the same name, like figure-skater Elvis Stojko. Then imagine doing it with a dozen-odd buttons on a remote control; that is the goal.

Although software designers at digital hub companies speak confidently about handling multimedia complexity, Nielsen thinks such talk is bravado. It could be five years, or 20, or even longer before they develop interfaces with the intelligence to make mixed collections of audio, video, and digital photos (and accompanying text annotations) accessible without long manual searches. Eventually, Nielsen says, digital filing systems may recognize a face in an image or a voice on a soundtrack and read the identifying snippets of metadata that may accompany a file.