Η Google evolve the translation service with a different approach that helps the user speak a foreign language better, even if they do not know it.
Πis about Google Translatetron, which uses the user's voice to translate and at the same time compile in real time what it says.
The tool does not follow the classic way of translating speech into text and from there to transcript and again to speech. Instead, Google is launching a different end-to-end technique that automatically compiles the user's voice into a foreign language. The process is done in parts by separating the speech and analyzing its spectrum (analysis of voice frequencies) to create a new spectrum in the language desired by the user.
The result of this process comes much faster and there is no risk of losing anything along the way. The translation - compilation is done in real time and sounds a bit like a robotic voice, but retains several elements of the user's tone.
At the moment we are in a rather early stage of development Google Translatetron, but over time its operation will be constantly improved.
[the_ad_group id = ”966 ″]