SonarÀgora

Synthetic singing: Towards natural, personalized and user controllable voices Masterclass by Voctro Labs

ES

  • Friday 17 13:30 - 14:30

    Sónar by Day - SonarÀgora

Synthetic singing: Towards natural, personalized and user controllable voices Masterclass by Voctro Labs

Designing machines that can simulate the human voice has been a dream of scientists and engineers for decades. Where once computer music pioneers manipulating early text-to-speech synthesis systems to produce robotic melodies, today’s singing synthesisers are complex generative systems based on deep neural networks. 

Barcelona’s Voctro Labs has spent 10 years developing singing synthesisers, and their presentation at Sónar+D will explore the current state of play of these technologies. How do different approaches model the human voice? How can artists and practitioners interact with such synthesisers? What are the future challenges and next steps for the technology? The interactive talk will draw on their work in everything from academic and industrial research to creating experimental prototypes and applications.