Deep-Learning Algorithm Duplicates Anyone’s Voice After One Minute of Training

A Montreal-based startup called Lyrebird has developed an algorithm that is able to duplicate anyone’s voice after just one minute of training (listening to their voice). This technology is the first of this kind and will help the rapid development of human-computer interfaces.

Thanks to its deep-learning algorithm, the new tool allows researchers to copy the voice of anybody with very little data. In fact, after just one minute of listening to a person’s voice, the system is able to duplicate it completely. This is possible because the algorithm works by comparing the similarities between the new voice and all the other voices it already knows and contains in its database.

The applications? Numerous. For example, some companies would like to let their clients have audio books read in voices of famous people, family members, etc. Others would like to allow people with voice disabilities to train heir synthetic voices to sound like themselves (which, of course, would be possible if there exist recorded samples of their speaking voices). And video game companies want in too: with the new tool, they could allow in-game characters to speak exactly like the human players playing them.


Lyrebird (

Digital Trends (


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s