Microsoft hopes WP 7 speech features surpass Android, iPhone

IDG News Service - Hoping to leapfrog over Google and Apple's successes in the smartphone market, Microsoft plans to use cloud-based speech recognition and natural language processing technology to offer user interface capabilities not found on the iPhone or Android devices.

"We believe speech is not a separate application. Rather it is an integral part of the user experience," said Zig Serafin, Microsoft unified communications general manager, before an audience at the SpeechTEK 2010 conference, being held this week in New York.

In order to boost Windows Phone 7's capabilities for understanding a voice command and delivering the requested result, the company plans to tie in Windows Phone 7 handsets with the company's Tellme cloud-based voice recognition and natural language processing service, said Serafin, in a subsequent interview with IDG News Service.

Microsoft purchased the company that created this service, Tellme Networks, in 2007.

Before the SpeechTEK audience, Serafin chastised the Android and iPhone operating systems for using icons as the chief form of interaction. "Most smartphones are a grid of icons, much like Windows 3.1," he said.

Talking to the phone is a more natural way of telling it what to do, he said. "When you move to a device that doesn't have a large keyboard, voice is such a compelling complement to that experience," Serafin said.

He then had Microsoft marketing director Ilya Bukshteyn run through a demonstration of how a Windows Phone 7 could use speech recognition and natural language processing, or the means by which a computer interprets what a person says. Bukshteyn asked the phone to call "Paul," and a voice emitting from the handset responded with a number of different contacts with the first name of Paul. Bukshteyn responded with the specific full name and the phone proceeded to call that person.

Bukshteyn also told the phone to open an album of pictures, and a picture view app came up on the screen, showing not only pictures taken by the user, but also taken by the user's friends that were posted on social networking sites.

In a third example, Bukshteyn asked for a list of nearby Chinese restaurants. The request was conveyed to the Bing search service, which returned a list of restaurants and their locations on a map.

While the iPhone and the Android variants do use some voice recognition capabilities, Microsoft's phone service will be different in a number of respects, Serafin said. For one, it will be not be restricted to being used in just a few apps, but rather could be used to control the entire phone. The second way in which this service will be unique is that it will be interactive. If given an ambiguous command, the handset or appropriate service can ask the user to clarify the request.