Software to make your media center 'listen'
One Voice program uses speech-recognition technology that lets users control gadgets via verbal commands.
Its One Voice software add-on for Windows XP Media Center is designed to allow users to
"If you have 900 music albums, you have to see the screen, scroll through menus and selections, or type in the artist's name to find what you want. With the voice control, you can simply say 'Play John Mayer,' and it scans through your collection, makes a match and plays it," said Adam Fitch, sales and marketing manager for One Voice.
To avoid your stereo switching between Rolling Stones and Beatles while you argue about which band is the greatest, there is a keyword that triggers the system into action. The default keyword is "One Voice," but consumers may choose any name or word they like.
The One Voice program, priced at $149.95, also can be applied to slide-show presentations on a PC, TV-show recordings ("Record channel 7 at 6 p.m."), voice over Internet Protocol phone calls ("Call Mom"), and the composing, sending and reading of e-mails and text messages.
According to Fitch, the basic operations of the program are based on a set of simple commands such as "play" and "record." When the application is launched it scrolls through the computer, gathers the names of picture folders, artist names, albums and genres, and automatically creates phonetic matches to each name or word in the inventory. "It has proven to be extremely accurate as long as you pronounce it as an American speaker," Fitch said.
One challenge the company still faces in marketing the software is the perception that voice-recognition technology is unreliable, he said.
"Our biggest obstacle right now," he said, "is all the voice recognition on the market that really came out too soon. The difficult part is getting people over the fact that they used voice recognition before and it just didn?t work."
One Voice, Fitch added, will understand both your 3-year-old child, your cousin from northern England and even you when you come home late and tired. The speech engine is pre-trained on a multitude of different accents and pronunciations and designed to cope with high levels of background noise.
Consumers also can create their own set of commands and program them to either prompt an audio response from the computer or simply execute tasks. "I have it in my home in San Diego," Fitch said. "When I say 'Surf's up!' it will go right on the television to a Web site that shows me the Webcam for Pacific Beach with the surf report," Fitch said.
Launched two months ago, the software is priced at $149.95.
One Voice hopes that home-automation companies will start adding speech recognition to their products. The software has been opened up for research to companies like Honeywell and Lifeware, according to Fitch. "Their customers are going to be able to walk in and say, 'Set the thermometer to 72 degrees, get the Beatles playing, dim the lights and start dinner.'"