The American company will update its text-to-speech engine to make the voice of its transcription system more natural, regardless of the language selected.
Google is at the forefront of text recognition and understanding, but its voice transcription system has always sounded somewhat robotic.
Voice transcription widely used in Google services
The voice synthesis engine is used across many software. You hear it for example when you use Google Maps in the car, at each direction indication, or in Google Translate to listen to a translation or make it heard to your interlocutor.
Google’s technology is therefore well established in many of its services, but the American company does not want to rest on its laurels and continues to refine it.
The Mountain View company has just announced a major update to its voice synthesis engine. As she explains on her blog for developers, this improvement will be available for the 421 voices and 67 languages supported by Google through its applications.
Slightly improved intonation and comprehension
The update will be made via the Google Play Store and will be transparent for the end user, who will not have to download new versions of their applications using voice transcription.
Google provides on its blog some examples of the new voices offered in comparison with the current voices used through its online services. We can thus realize the improvement of the intonations during the transcription but also note a better clarity.
The changes are currently minimal and we have not been overwhelmed by the results of this new synthetic voice in the English language. Indeed, the audio file is more pleasant to listen to, but the differences are slight.
Unfortunately Google does not offer an example in French to give us a better idea. We will now have to wait a few more weeks, the time for the company to deploy its update on all compatible devices.
Source : The Verge
1