2018.09 updates

https://azure.microsoft.com/en-us/blog/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. With the human-like natural prosody and clear articulation of words, Neural TTS has significantly reduced listening fatigue when you interact with AI systems.