Developing an AI-Powered Karaoke Experience

Explore the development of an AI-powered karaoke experience in this conference talk from the Audio Developer Conference (ADC23). Dive into the challenges and solutions for creating a fully automatic and integrated karaoke system adapted for mobile and web platforms. Learn about cutting-edge deep learning technologies applied to audio source separation, voice transcription, real-time stems remixing, pitch and tempo control, and singing quality assessment. Gain insights from Thomas Hézard, leader of the Audio Research & Development team at MWM, and Clément Tabary, a deep-learning research engineer, as they discuss their innovative approach to enhancing the traditional karaoke experience using advanced signal processing algorithms and optimized implementations.