Japan’s biggest mobile carrier NTT Docomo has developed a real-time Japanese <-> English translation service for mobile phones, the first of its kind. The way it works is that you speak something into the device and wait to hear a voice interpretation of what you just said in another language.
As you can see in the videos embedded below, the service, which uses the cloud for the heavy lifting, isn’t quite “real-time” but pretty close. NTT Docomo says the service can be used for communicating over the phone but also face-to-face.
The company also claims the success rate for speech recognition stands at about 90% in the case of Japanese and about 80% for English (other languages will be added soon). That’s significantly more than the 15-20% back in May, when we introduced the service first.
A test with 400 end consumers in Japan started today and runs through March next year. If everything goes according to plan, NTT Docomo plans to offer the service to all of its 56 million subscribers in the second half of 2012.
This video [JP] shows the service in action:
Here’s another one: