Speech synthesis
with deep neural network
can surprise you
because the algorithm
can understand
in a generative way
English
or Mandarin
which is better
than the parametric
or concatenative
version
and you can
hear breathing
and mouth
movement sounds
as a new reality.