If you feel a certain strangeness when you are answered by an Artificial Intelligence and it has a robotic voice, you will like this news. A Microsoft announced an AI that is capable of imitating any human voice in less than five seconds, VALL-E.
The program listens, synthesizes and imitates the human voice in different contexts. The niche to be reached by the company is software "text-to-speech” or “text-to-speech”, in free translation.
see more
Alert: THIS poisonous plant landed a young man in the hospital
Google develops AI tool to help journalists in…
More natural, less robotic
The idea of this AI is precisely to make an automated voice as natural as possible. More than that, according to Microsoft, it is an attempt to avoid “depersonalized” voices, like those that exist today in applications such as Google translator.
The company relied on more than 60,000 hours of recordings. The audios served as a basis for the AI to identify the various nuances and tones of the voice. In addition, it was also possible to identify the speech humor. Listen below.
VALL-E also synthesizes voice variations for the same input text. pic.twitter.com/Yy9hj05Qa3
— Amogh Vaishampayan (@amogh42) January 7, 2023
AI imitating human voice can be matched
VALL-E can be combined with other AIs, according to Microsoft. An example is the GPT-3, an OpenIA conversation and text generator.
Both technologies are generative. This means they can create content from samples. Therefore, they need a large database to run well.
So far, AI that imitates human voice only works in English. It is possible that, in the coming months, other languages – including our Portuguese – will also be available.
came to stay
This is just another example of how AI is becoming more and more a part of our daily lives. If before this was an exclusive topic for science fiction movies or programming students, today it is an important part of our routine.
Therefore, it is good to get used to the possibility of dealing with this technology more and more.
Graduated in Social Communication at the Federal University of Goiás. Passionate about digital media, pop culture, technology, politics and psychoanalysis.