Google is evidently working on real-time mobile translation tech

Android speech recognition
Could Google tech help us communicate everywhere?

Google has its sights set on the future with projects like Google Fiber and Google Glass, and now it's adding real time voice-to-voice translation to that list as well.

Google's Vice President of Android Hugo Barra said this week that Google is now in the early stages of creating real-time translation software that it hopes to perfect within the next "several years," according to The UK Times.

The company already has prototype phones that can translate speech in real time, so that a user speaks into the device in one language and the person on the other end hears it in a different one, like the fictional Babel fish in "The Hitchhiker's Guide to the Galaxy" or the TARDIS in "Doctor Who."

"That is where we're headed," Barra told the publication. "We've got tons of prototypes of that sort of interaction, and I've played with it every other week to see how much progress we've made."

Same old hurdles

Google's speech-to-speech translation project is reportedly being developed as part of Google Now, Google services suite that's being designed to predict your needs before you know them.

The real-time translation is reportedly better for certain language pairs, such as Portuguese and English, but accuracy remains an issue.

Anyone who's tried to use Apple's Siri or Android's voice-to-text services knows that a little background noise can cause a lot of inaccuracies, and that's something Google is wrestling with still.

Google Translate

Translations per day: a billion and one

The groundwork for real-time, voice-to-voice translation certainly exists, though, between that speech recognition software and Google's online Google Translate service.

Google said that on that service alone it translates a billion entries per day in 71 languages, and it just added new languages from places like the Philippines, South East Asia and Indonesia.

Don't stop me now

Google discussed voice translation software back in 2010, when Google Distinguished Research Scientist and head of machine translation Franz Och offered this:

"We think speech-to-speech translation should be possible and work reasonably well in a few years' time. Clearly, for it to work smoothly, you need a combination of high-accuracy machine translation and high-accuracy voice recognition, and that's what we're working on.

"If you look at the progress in machine translation and corresponding advances in voice recognition, there has been huge progress recently."

It would have been nice if he was right - we'd probably have real-time voice translation on our Galaxy S4 right now. But at least we know they're still working on it.

  • TechRadar spent a week with Google's newest search tool and wrote about what it's like living with Google Now.
Michael Rougeau

Michael Rougeau is a former freelance news writer for TechRadar. Studying at Goldsmiths, University of London, and Northeastern University, Michael has bylines at Kotaku, 1UP, G4, Complex Magazine, Digital Trends, GamesRadar, GameSpot, IFC, Animal New York, @Gamer, Inside the Magic, Comic Book Resources, Zap2It, TabTimes, GameZone, Cheat Code Central, Gameshark, Gameranx, The Industry, Debonair Mag, Kombo, and others.


Micheal also spent time as the Games Editor for Playboy.com, and was the managing editor at GameSpot before becoming an Animal Care Manager for Wags and Walks.