Parrots are the one of the best
example that imitates the human voice if we put much effort on training it. However,
assuming they can copy the speech and your voice sounds the same as you.
Terrible, right? Then replace the Parrot with artificial intelligence
algorithms, do you have the latest progress in machine learning.
So no its became real as Montreal-based
startups Lyrebird is making tremendous progress in speech synthesis technology.
In other words, their algorithms are good for imitating people. Ordinary people
can easily copy Liyaneisen, Arnold Schwarzenegger and Clint Eastwood's voice.
They have a simple slogan, you can imitate.
However, Lyrebird does not imitate,
it is to learn. Just use only one minute of data recording from a particular
person, their AI can pick up all the functions, unique sound. Of course, there
is not enough DNA data in a minute to create a representative style speaker. Consider
that do not put words in your mouth with that in mind, you should consider all
the possibilities.
Instead of boring sound books from speaker Morgan Freeman's
voice. The sound weakened the people to recover the sound. Completing the
complicated Ponzi phone system. The possibilities are endless. Mentioning that
perhaps the most appalling Lyrebird app will be paired with VOCO, Adobe's
secret design allows users to easily edit the voice of the text. Then add
face2face, which makes convincing editing facial expressions.
Now you can
effectively put words into someone's mouth, and there is video evidence. If you
believe that Fake News 2016 is wrong, wait until the material is good to use
voice editing software. Fortunately, as everyone wears adolescent adolescence
Lyrebird still have some voice cracks here and there, VOCO are not yet
available to the public, but only an experiment at Face2Face University. But
not too convenient, because these tools will quickly mastered the boundaries
between true and false public almost disappear.
