Apple gets edge with use of voice-recognition technology
— -- Say what?
The voice-driven Siri virtual personal assistant that Apple touted as a major feature in the brand new iPhone 4S says a lot about how far speech recognition in phones has come.
Siri not only recognizes what it is that you are asking — "wake me up at 6:30", "any decent sushi restaurants around here — but it understands the intent of your words and can respond accordingly.
Siri relies on artificial intelligence and is tightly integrated with the various apps on the phone. Results are tailored based on your location and through partnerships with the likes of Yelp.
Mike Thompson, who heads up the mobile business at Nuance Communications, called the Siri news "an exclamation point on the mainstreaming of the technology." Adds Dan Miller, senior analyst at Opus Research, "I'm calling it an inflection point of sorts."
As a leader in the field, Nuance has partnerships with Apple (and Siri), and several other companies, and is the outfit behind Dragon Naturally Speaking speech-recognition products for computers and mobile.
Analyst Shyam Patil of Raymond James estimates that Apple is paying Nuance a recurring $5 million to $10 million in revenue every quarter over the life of an estimated two to three year contract. "If anyone can get speech right, its Apple," he says.
Of course Nuance is by no means alone. A veritable who's-who of tech companies including Google, Microsoft and IBM spend a lot of time on speech-related research, as do startups such as Vlingo.
"Our vision is complete ubiquity of this type of technology," says Mike Cohen who leads the speech technology efforts at Google. That means on every application, webpage, social network, in the car and so on. Google executive chairman Eric Schmidt has said that 25% of Google searches on Android phones are triggered by voice. Along with voice search, Android users can employ voice actions ("note to self -- grocery list: banana, milk, eggs, pizza") to get things done.
Until its inclusion as part of the iPhone 4S, a version of Siri was available on iPhones as a third party app. Apple bought the Siri startup in April 2010.
Though the app disappears, the promise of the Siri approach is in the way it is tied into the very fabric of the 4S and the way in which you can pose a question. For example, "do I need an umbrella today?" yields the same result as "what is the weather like today?"
"Its not that you get to use your voice, but you get to talk in a way that you feel comfortable with," Miller says. "The systems themselves are less robotic, it doesn't seem as gimmicky. There is a true Star Trek effect."
Good as it has gotten, voice recognition is far from perfect. Apple says Siri on the iPhone 4S is in beta. Siri isn't completely hands-free either. Though it can help cut down on certain steps as you navigate tasks, a typical Siri interaction may involve a combination of using your voice, looking at results with your eyes and touching the phone.
Thompson says: "We're in the 4th inning—the rate of change and innovation is faster than ever before in speech. The accuracy and performance is getting better. (In the) next five innings we'll see greater and greater natural language."
Popular Reads
ABC News Live