Julia B. Hirschberg

Julia Hirschberg was the first to create fluent automated text-to-speech systems by combining discourse structures and prosodic features in speech processing. Her contribution was to identify the role of prosodic information in speech and then to use this knowledge to produce modern realistic text-to-speech synthesis systems. Her work allowed the subsequent automatic detection of basic speaker states such as anger, fear, or surprise and derived states such as uncertainty or deception. Her work is widely used in today's text-to-speech systems, voice-response systems, and spoken-dialogue systems.