A new Google speech synthesizer talks like a real person

Google introduced the second-generation Tacotron speech synthesizer. Thanks to the competent use of AI   engineers of the company managed to synthesize a human voice, which is not easy to distinguish from   of the original.

What Tacotron 2 can do

Thanks to AI   The synthesizer of speech has learned to speak like a living person, and not the Daleck of "Doctor Who." The   there, he   pauses after commas and   points, focuses on   beginning sentence, observes stresses and   keeps the pace, like a living person. By   link is its comparison with   a living person. Try not to guess   looking. A   unusual words sometimes put him in   dead end.

How it works

Google has broken the process of synthesizing speech into   2 stages and   divided them   between two neuronets. The first neural network turns the text into   audio spectrogram   - a kind of screenshot of the equalizer with   a clear sequence of sound frequencies. A   The second WaveNet neural network interprets this file and   turns it into a   speech. Thanks to this Google   you need to connect to   internet to talk.

What's next

Google has not yet   discloses plans for   implementation of Tacotron 2. However,   You need to be a genius to assume that if everything works, then   already in   will soon appear in   Google products like voice assistant, translator or Google maps.