Speech recognition, personalization, mass storage among
the advances that will drive the box of 2010

Forty years ago, George Jetson and his family buzzed their way into the American consciousness, stoking a futuristic lifestyle in the imaginations of millions of TV viewers. Today, the digital set-top is being positioned as the "gateway" device that will control gobs of information flowing into, out of, and around, the home.

Bob McIntyre
Bob McIntyre,
CTO, Scientific-Atlanta
Maybe it's time to give the animators and storywriters some credit. In the year 2010, a mere eight years from now, the digital set-top box will inch closer to the everyday expectations of George Jetson.

Clearly, the sheer horsepower to drive the set-top and applications of the future will be available. Using Moore's Law, the ultimate metric for future technological advancements, processing power, based on Intel Corp.'s projected 3 GHz release scheduled for the end of the year, will undergo at least three 18-month cycles before 2010, putting central processing unit (CPU) strength at a rippling 24 GHz. Set-tops of today use a fraction of available CPU power, but there are other metrics that will determine the capabilities of the set-top that will debut less than a decade from now.

Memory and network bandwidth cycles–the latter driven by fiber optics speed/capacity enhancements–are on track to double even more quickly than the 18 months that Moore's Law dictates, according to Bob McIntyre, chief technology officer for Scientific-Atlanta Inc. "Bandwidth is working at a rate that's double Moore's Law," he says. Consequently, set-top boxes of the future will not be CPU, bandwidth or memory restricted.

Visions OF 2010

Clues about future set-top functionality are seeping into the marketplace, with companies like Digeo Inc. leading the way with set-top/media center reference designs and software that converges multiple applications and formats, including video tuning, personal video recording (PVR), DVD and MP3 playback.

The media center, says Eric Roza, vice president of marketing for Digeo, "will be serving every room in the house," with different types of content, regardless of its source and regardless of whether it's streamed, stored or broadcast.

PVR giant Tivo Inc. is advancing its own platform, and is collaborating with RealNetworks Inc. to add streaming playback functionality to Tivo boxes, as well as digital photo display and organization, notes Tivo Executive Producer Jim Monroe.

The proliferation of content choices will require better navigation by 2010, says Roza. Consequently, evolving the remote control to a point that it may include a scroll wheel, data entry pad and voice recognition capability will be commonplace. Monroe adds that the variety and quantity of content will require the Tivo platform to become, in effect, a media management system.

Combining navigation and management suggests that a wireless personal digital assistant (PDA) display device will help viewers determine what programs they want to record, without disrupting the programming they're watching, and will provide set-top details such as how much storage is available, says Anthony Simon, director of digital cable set-top box products for Conexant Systems Inc.

"Over time, we see technology becoming more discreet and disappearing," says Simon Poulter, director of corporate communications for Philips Consumer Electronics/Digital Networks.

Instead of piles of boxes and cables running around the home, these devices will begin to disappear from view, perhaps tucked away in the attic, or consolidated into another central location in the home.

In the vein of Ray Bradbury's futuristic novel Fahrenheit 451, the house of tomorrow, adds Poulter, could be outfitted with flat-panel displays that have grown in size, effectively becoming "video walls."

Once a person walks into a room, the set-top/server will "recognize" the person through, for example, a badge with a Bluetooth-enabled transmitter. Once recognized, the system will refer to personal preferences stored in metadata, such as what music the person likes and/or what news or financial news channel is preferred at what time of day. A discreetly located set-top will respond accordingly.

McIntyre says that the set-top, or "network edge device," as he prefers to refer to it, "will be a triple play (voice, video and data) converged router" that constantly searches for content, manages unlimited bandwidth in both upstream and downstream directions, is unencumbered by wires, and displays high-definition video. Content, he adds, will be controlled by voice commands.

Philips has packaged its vision of home entertainment into a concept called "Ambient Intelligence," a rather broad and abstract concept that relies on "smart" technology that can be beckoned from anywhere.

"The building blocks are there for this ambient technology," says Poulter, referring to elements such as Internet audio, PVR and speech recognition–all of which are manifested in Philips consumer electronics products available today.

Philips, for example, has utilized speech recognition technology in its mobile GSM (Global System for Mobile Communications) phones for three or four years.

"Voice recognition will clearly be a no-brainer," agrees McIntyre. Two possible implementations of voice recognition are possible, he says. One implementation involves a thumbprint biometric circuit on a remote/handset, which recognizes the person using the unit and refers to that person's personalized preferences, including programming and musical tastes. The other paradigm discards the remote and allows the user to talk directly to the video display.

In fact, McIntyre says that the S-A Explorer platform is today capable of voice recognition and calling up specific customer profiles that tune to a favorite "news" channel when "news" is spoken.

Philips is currently researching the feasibility of incorporating speech recognition in TV sets, allowing a consumer to verbalize instructions such as "volume up," "volume down" or "change channel," resulting in a corresponding action in the TV.

Speech recognition aside, the visual aspect of the future set-top might be best described as out of sight, but not necessarily out of mind. According to Rahul Mehra, director of advanced services for Pace Micro Technology plc, "the form factor [of the set-top] might not necessarily matter that much as a home server may be hidden away in a cupboard, doing its job, out of sight."

See me, hear me

Applications are coming to market that will have TV viewers not only talking to their TVs, but listening to the TV talk back to them. U.K.-based Media Logic Systems Ltd. has been involved in trials of its iSeeTV service, which lets shoppers tune to a shopping channel, dial a number on their telephone, and begin talking to an on-screen "shopping assistant." The service has attracted a financial services/insurance firm and the U.K.'s National Health Service, which allows patients to see and talk with an NHS-qualified nurse for medical advice.

This iSeeTV application allows viewers to
access a real-time video call center.

"What we're basically doing is video-enabling a call center," says Glyn Radcliffe-Brine, commercial director of Media Logic. The service is particularly adept at processing complex sales queries, such as auto financing and insurance, and real estate, he adds.

Radcliffe-Brine says a hotel concierge service using iSeeTV will debut on the West Coast of the U.S. this year (at an unspecified location with an undisclosed cable operator), letting patrons tune to a channel and ask questions about nearby attractions and activities. He expects the service to eventually be rolled out to the operator's cable subscribers.

Digeo, Roza points out, has demonstrated a videophone application by plugging a Web-camera into the universal service bus port of its set-top–a functionality he expects will proliferate. "It's hard to believe people won't want the option to videoconference with their family," he says.

A.I.: more than a movie title

Not only will the set-top of the future allow for person-to-person interaction, but reliance on machine intelligence will increase as personalization services proliferate. Interactive application developers have long trumpeted the benefits of extending personalization beyond TV entertainment.

But developers at BIAP Systems Inc. are hoping to push personalization of content to the set-top box and TV. BIAP has demonstrated so-called embedded "artificial intelligence" software for set-tops, based on "intelligent agents," which, once configured, venture out and retrieve selected content from data sources, such as Web sites or operator-controlled databases, and convert the content to an appropriate format for display as an overlay. BIAP's agents can extract data from formats such as XML or HTML. Because the agents reside in software in the set-top, they require no server-side support.

AI in this context, says Louis Slothouber, chief scientist for BIAP, is a set of technologies for processing information, and most importantly, organizing, retrieving and filtering data.

BIAP has introduced a service, dubbed PiTV, which allows subscribers to personalize on-screen crawls, scrolls and alerts, such as local news and weather, sports scores, stock quotes and e-mail alerts, with updates delivered automatically, regardless of what video channel is displayed.

"Our artificial intelligence expertise will grow along with set-top box capability," says John Gregoire, executive vice president of BIAP's broadband TV business unit. In its current form, PiTV is designed to run on today's class of set-tops via a software program that soaks up about 1 megabyte.


Storage devices in set-top boxes are becoming a standard feature as personal video recording moves from niche to must-have status. The standard 3.5-inch form factor of hard drives will shrink to either 2.5-inch or 1.8-inch sizes, depending on what the market is demanding, says Rob Pait, senior sales and marketing manager for consumer electronics for storage developer Seagate Technology LLC.

Because current storage technologies are beginning to reach their limits, Pait notes, "the whole storage industry is looking at different ways to record information on platters."

Three contenders for new storage technologies are: perpendicular recording, in which bits are "stood up" on end in a vertical orientation; heat-assisted magnetic recording (HAMR), which uses a laser-generated beam to heat the recording medium for greater bit density; and holographic recording. Despite the choices, it's too early to predict a winner in the consumer hard drive market.

In terms of capacity, "we don't see the curve ramping quite as sharply as it has in the last three years," says Pait. During that time, consumer hard drive capacity moved from 4 gigabytes to 80 gigabytes. As far as hitting the next magic threshold of 1 terabyte, Pait notes, "it's hard to say" that hard drives will reach that capacity via a 2.5-inch or 1.8-inch drive. However, he adds, "it's easy to see hundreds of gigabytes of storage."

But another storage metric that's also important is drive read/write times. Current ATA (Advanced Technology Attachment, a.k.a. IDE) data rates are 100 megabits per second, while first-generation Serial ATA drives push 150 Mbps. Second-generation Serial ATA drives, expected to debut in 2004 or 2005, will be capable of 300 Mbps, and Pait believes 600 Mbps will be the next leap. These increased throughputs will help drives handle more audio and video streams as applications such as PVR proliferate.

Advanced display

A few years ago, the notion of holographic TV was floated around the cable industry, and captured the imaginations of the futurists, but the technology most likely will not be ready for prime time by 2010, despite the massive CPU and storage capabilities expected to be available.

"Sheer processing power is only part of the solution," says Stephen Benton, founder of the spatial imaging group of the Massachusetts Institute of Technology's Media Lab. One roadblock is limited device bus bandwidth.

More likely is advanced 3-D technology using displays with cameras that can determine where viewers are located in a room. In this way, different programming can be beamed to different people in the room, and/or different images beamed to the left and right eye, thus creating 3D images without glasses. Prototype boxes have been developed that send two images.

The fly in the ointment of any discussion of future set-tops is a series of issues, including the resolution of digital content copying and distribution, costs and operators' business plans based on consumers' perceived acceptance of new technologies. But clearly the horsepower and advanced applications will be on the shelf if and when operators decide to take the next step toward the futuristic home that George Jetson has dangled in front of our eyes for so many years.