Microsoft Develops Translation Tool That Mimics User’s Own Voice
With the advent of Siri for the iPhone, text to speech is the new frontier for translation apps. Where Google Translate, and other text translator apps were once life support for travelers and foreign language speakers, now audio and voice are taking over the trend.
While the semi-robotic voice of Siri and other audio tools are sufficient substitutes for the lack of a speaker’s own Bi-lingual abilities, Microsoft wants to take the concept to the next level by creating a prototype that acts as a translation tool that retains the original voice, including tone and accent.
Unveiled at Microsoft’s recent TechFest 2012 event and developed by Microsoft Research Asia, the speaker uses a system to model the inflections in their voice (this takes an hour of learning by the system), but rather than produce a robotic translation, as per other services, Text To Speech repeats the phrase in the selected language but keeps speaker’s original voice.
Right now, the prototype works in 26 languages but is still not 100% perfect. But once Microsoft perfects this application, it is expected to be a great hit on smart phones.
Although, personally, I can’t wait for the hi-jinks that will undoubtedly ensue from a device that perfectly mimics the sound of your roommate’s voice.