Microsoft’s new neural text-to-speech service helps machines speak like people
Microsoft has reached a milestone in text-to-speech synthesis with a production system that uses deep neural networks to make the voices of computers nearly indistinguishable from recordings of people.