CNET también está disponible en español.

Ir a español

Don't show this again

Christmas Gift Guide
Culture

Talking to your TV

After finding success in cars and call centers, IBM's speech recognition team works to add voice commands to consumer gadgets.

Go ahead and yell at the TV. If IBM has its way, your home entertainment gadgets will soon listen to you.

Big Blue announced a partnership Wednesday with browser maker Opera Software to produce a voice-activated programming guide. IBM engineers expect it to become the foundation for numerous consumer gadgets that will use speech recognition to replace bewildering menus and clumsy control systems.

"With a lot of these new entertainment devices, you need the ability to navigate through a lot of information, which means cramming more and more buttons onto the remote control," said Gene Cox, director of IBM's contact center product and solutions management group.

Speech recognition promises a much more elegant approach, Cox said. Instead of using a remote to hack through a complex menu structure, just say "record Arnold Schwarzenegger" to your digital video recorder, for example, and it'll grab the next showing of "The Terminator."

IBM's ViaVoice technology is already widely used in corporate call centers and, more recently, for navigation systems in cars, where clicking buttons to scroll through complex menus would pose a safety hazard. Cox is confident that the technology will filter down to other devices, as advances in processing power make speech recognition feasible in small devices and proliferating information makes it necessary.

"If you look at something like a TiVo programming guide, it's really a very complex piece of information that poses a lot of choices," Cox said. "Speech really helps simplify the process if you have anything more complex than play, pause and record."

The first applications of the voice-equipped Opera browser will likely be on media-centric PCs, but the browser is slim and flexible enough for the technology to filter down to all manner of devices, Cox said. It's not a stretch to imagine a voice-activated iPod, where a brisk "Play Johnny Cash" would replace scrolling through menu trees.

"Matching the name of an artist or album to something in a database--that's not a terribly difficult problem to solve," Cox said.