Sign In

Communications of the ACM

ACM TechNews

Can Google Get Web Users Talking?

Google researchers are implementing voice-driven Web search capability on computers, with the goal of making such searches as easy to perform as they are on the Android handset.

Google's Vincent Vanhoucke says that voice recognition is more technically complex on a laptop or desktop computer because it calls for noise-filtering algorithms that mobile speech recognition does not require. Google is pressuring the World Wide Web Consortium to roll out a standard set of HTML markup that permits any Web site or application to call on voice recognition through the Web browser, and it has already enabled a version of this markup in the Chrome browser.

Google's desktop speech recognition is precise for search queries, but has little utility for tasks such as composing email. This challenge could be met by enabling the system to learn the personal mannerisms of each person's pronunciation.

"The advantage of Google's networked approach is that a [speech] model in the cloud can adapt to your voice in all these different places and follow you around," says Princeton University's Keith Vertanen. However, he notes that users may be resistant to the notion of using voice recognition on the desktop, given their familiarity with existing input tools. Vertanen also says that people are less acceptable of glitches when using speech recognition on a desktop because of the keyboard's close proximity.

From Technology Review
View Full Article

Abstracts Copyright © 2011 Information Inc. External Link, Bethesda, Maryland, USA 


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account