Development of technologies for next-generation voice changer towards barrier-free speech communication

语音产生机制中的物理约束有时会导致语音交流中的各种障碍，如发声障碍、歌唱表达的局限性等，统计语音转换技术能够灵活地转换语音，使语音产生超出物理约束，是突破这些障碍的一种潜在技术。 为了将该技术应用于语音通信中，必须开发一种高质量、实时的转换算法。 在这个研究项目中，我们开发了基于实时统计语音转换的下一代语音转换器的基本技术。 此外，我们还开发了使用新一代变声器的各种应用程序，以打破现有的语音通信障碍。

Physical constraints in our speech production mechanisms sometimes cause various barriers in speech communication, such as vocal disorder, limitation of singing expressions, and so on. Statistical voice conversion capable of flexibly converting speech sounds is a potential technology to break down these barriers by augmenting speech production beyond the physical constraints. To use this technology in speech communication, it is essential to develop a high-quality and real-time conversion algorithm. In this research project, we have developed basic technologies for a next-generation voice changer based on real-time statistical voice conversion. Moreover, we have also developed various applications using the next-generation voice changer to break down the existing barriers in speech communication.