Microsoft VALL-E can imitate any voice with a sample time of just 3 seconds

We recently talked about how Microsoft had been considering adding AI-powered tools from ChatGPT to Bing.but it seems that his interest in artificial intelligence comes from before and is strong, because as we can read in this linkMicrosoft has released VALL-E, a tool capable of imitating any voice with a small sample of 3 seconds.

This tool is trained on 60,000 hours of English voice dataTherefore, if someone wants to use this tool, they must use voice samples in English, something that, although it will not be difficult if we want to imitate the voices of famous people from the Anglo-Saxon world, for the rest of the cases it can be more complicated.

One of the main improvements of this model in relation to other forms of speech generation is the fact that emotions can be reflectedso that, for example, if we have a happy voice sample, the result will be happy, whereas, if the sample sounds sleepy, the result will be too.

For now, this tool is not available for public useWell, as we mentioned initially, one of the main sources of samples can be Anglo-Saxon-speaking celebrities.which can be used maliciously.

End of Article. Tell us something in the Comments!