Sonantic, a technology development company, claims to have created an artificial intelligence that is capable of crying, announcing her as the first to do so. The objective is to offer a synthetic alternative to the use of real interpreters to contribute their voices and performances to video games.
The system focuses on a process of text to speech that simulates the emotions implicit in an interpretation from the script that is supplied and is capable of generating the simulated cry.
"The objective of the company is to really capture this deep emotion using machine learning"Felix Vaughan, one of the researchers involved in the development of this tool, has said. To demonstrate his work, they have published an illustrative video that you can see below:
In the video demonstration we see two characters, mother and daughter, in an interaction with reactions that they are "completely computer generated". This means that, unlike what is currently happening, no interpreters have been used to double the characters we see on screen.
Realistic emotions without hiring actors
For the base of tears voice samples from selected human actors are used
As they explain, the actors and actresses who have worked with them generate income when developers use their synthetic voices thanks to licensing agreement oriented to the generation of passive income for them.
In short, the objective of Sonantic has been to create a system of dialogues with much more natural synthetic voice and although they have started with crying and
