Experimental programs that apply speechacknowledgment to do real-time language translations have beenkicking about for years now, but Microsoft brought the thought to anew level at its TechFest 2012event last week.
As other translators, the softwaredeveloped by Microsoft Research lets you to talk to it in your nativetongue and send it out the speaker of a device as another language --Spanish, French, Chinese, or such. What arrives out of that speaker,though, is not the ersatz speech of a computer robot, but an ersatzapproximation of what you sound like.
What is more, the software will make a3D image of your head that creates it feel like if you are speakingthe translation. Called Monolingual TTS, the system presently has 26languages in its repertory. Introducing the system on a voice doesbring more time as compare to it does in your typical speechrecognition program -- about an hour of training Monolingual TTS inyour vocal tones. If the application can be thrust into asmartphone, it would be a boon to international travelers. Languagetraining is even a good fit for the technology.
Microsoft's vision in the translationspace is broader as compare to what has been brought forth by itsrivals. It needs to apply lifelike virtual avatars that not justmimic your feels but as well your voice and the movements of yourlips when you speak. The results made from that vision can beimpressive, but they can be creepy, as well. Google, which has abroadly-applied online text-dependent translator, has even beendelving into speech-based translation.
Although it would look that Apple'sSiri voice application would be a natural for speech translationperformances, it does not native support the task. Still, if youjailbreak an iPhone 4S, which is not encouraged by Apple, you can runan app called Lingual which lets you to tell something in your nativelanguage and Siri will display it with the iPhone's display in one of30 languages.