Vivos

Creation of Expressive Voices: Multimedia Applications

This project aims to fill the current void in post-production sound tools for image and voice, to be used for film, video and multimedia productions. The project hopes to put in place a group of synthesis and processing techniques, concentrating specifically on the expressive quality of the voices created, through a combination of several methods such as:

• transforming the voice identity: change the characteristics of the vocal identity from an existing recording, modification of the voice's timbre,
• converting the voice type and quality: passing from a man's voice to the voice of a woman or child, etc.,
• synthesis of an expressive voice: modification of the expression inherent to a given voice,
• text-to-speech: production of a synthesized voice from a text document,
• synthesis from a group of actors and people: synthesis consisting of using an existing voice, characterized from a group of recordings, saying an arbitrary text.

There are several foreseeable applications, including; film dubbed with a synthesized voice, the creation of voices for cartoons, the synthesis of the voice of an actor or a real person from an unpublished text, the modification of a recorded voice's expression or accent, the creation of several vocal identities (characters) using just one actor, vocal synthesis in games that use a system for generative text production and the expressive qualities in connection with the nar-rative context, the creation of live multilingual sport commentaries, etc.

Project Details

Participants