Special Section on Processing Natural Speech Variability for Improved Verbal Human-Computer Interaction

摘要

Speech processing techniques such as speech synthesis and speech recognition have been improved by statistical methods and are expected to be fundamental techniques for natural communication between humans and machines. However, to improve verbal human-computer interaction, we need to carefully consider natural speech variability such as speaking styles, emotions, speakers' voices, environmental noise, and different kinds of expressions.