Speech Synthesis - From Models to Signals and Back - Day 6 Afternoon

Explore the comprehensive evolution of speech synthesis technology through this 3-hour 20-minute lecture by Jindřich Matoušek from the University of West Bohemia in Pilsen, delivered as part of the JSALT 2025 summer school at Johns Hopkins University's Center for Language & Speech Processing. Delve into the historical development of speech synthesis systems, examining how the field has progressed from early computational models to sophisticated modern approaches. Analyze current state-of-the-art techniques in converting textual input to natural-sounding speech, understanding the complex processes involved in transforming linguistic models into acoustic signals. Investigate the bidirectional relationship between speech models and signal processing, gaining insights into how advances in one area drive innovations in the other. Discover emerging trends and future research directions that will shape the next generation of speech synthesis technologies, with particular attention to recent breakthroughs in neural approaches and their implications for creating more natural and expressive synthetic speech.