“I suppose if they keep track of [the] IP address, they could adapt” to a given user’s voice, says Jim Glass, a speech recognition expert at MIT. Glass notes that the mobile phone provides an acoustic environment very different from that of a laptop or desktop computer; for one thing, a phone’s microphone is reliably placed right at the user’s mouth, unlike computer microphone setups in homes or offices. “This is the beta version of Chrome,” says Glass. “They’ll be collecting data, and we can be sure they will be refining their models–that’s the nature of the speech-recognition game.”
Even if it’s rough around the edges, sometimes the technology impresses. I tried once again and got back “the final warning sounds of the dress rehearsal at laurel players with nothing to do with stand there.” Not so bad. And the Chrome app nailed it to a letter when all I said was “the quick brown fox jumps over the lazy dog.”
Third-party programmers have also begun creating Web pages capable of using the new feature of Chrome. Already available for trial is a browser plugin called Speechify that lets you search Google, Hulu, YouTube, Amazon, and other sites using voice with Chrome.
Other inventive uses could soon follow. “Games could be taking keyboard, mouse, touch, accelerometer, and speech input together,” says Karl Westin, an expert on HTML5 who works for Nerd Communications, based in Berlin, Germany. “Having an aeroplane game where you could actually scream ‘up, UP, UUUPPP!’ could be fantastic.”
But the technology is more than just a toy—it also points the way to a much more capable Web. HTML4, the last major version of the HTML language, emerged in 1997. Since then, plugins like Silverlight and Flash have added media-processing capabilities to the Web. But HTML5 enables media playback and offline storage via the browser.
“The insight we had was that more and more people were spending all their time in the browser,” says Google’s Brian Rakowski, group product manager for Chrome. E-mail and instant messaging increasingly take place in browsers rather than in separate e-mail or AIM applications. “We’d like it to be case that you never have to install a native application again,” says Rakowski. “The Web should be able to do all of it.”
Hear more from Google at EmTech 2014.