Last week, one of the largest technology companies in the world revealed a hack that uses the best of artificial intelligence (AI). Microsoft has developed an artificial intelligence capable of reproducing and imitating any human voice within seconds of analysis. The new technology was christened as VALL-E.
Microsoft is developing amazing artificial intelligence
According to technology analysts, the new Artificial intelligence Capable of enhancing the use of voice to feed text-to-type algorithms – which convert texts to speech.
As the company points out, the new AI needs just three seconds to hear and imitate a person’s voice in different situations and several different contexts. Microsoft itself stated that the novelty is a new language model for assembling text into speech in a digital and authentic way.
For some, this is an obvious attempt to expose the automated voice that is in Google Assistant, as it is in Translate, or in other competing apps.
More than 60,000 hours of recordings have been used
the Artificial intelligence It contains a database with more than 60,000 hours of recordings. They were used as a training base for acquiring intonations, accents, intonations, and many other characteristics of each type of articulation.
The artificial intelligence is able to identify the sound of the external environment and even the mood differences of the person speaking. In other words, this AI is able to faithfully imitate the speaker’s voice and very closely to reality.
The development result, according to the company, makes it possible to keep “the emotion of the speaker and the sonic environment of the soundboard in tuning.”
Even Microsoft’s novelty can be used in conjunction with another type of AI. Imagine using this AI with a text generator, for example. It would be as if a machine could create ceilings and produce placements at the same time. Scary right?
“Web geek. Wannabe thinker. Reader. Freelance travel evangelist. Pop culture aficionado. Certified music scholar.”