Establishment of Technical and Social Foundations for Constructing a Universal Speech Model Based on Massive Data

Author: Keiichi Tokuda (Principal Investigator), Akinobu Lee (Co-investigator), Yoshihiko Nankado (Co-investigator), Tomoki Toda (Co-investigator), Junichi Yamagishi (Co-investigator), Keiichiro Oura (Research Collaborator), Kei Hashimoto (Research Collaborator)

Period: April 2013 – March 2016
Grant type: Japan Society for the Promotion of Science, Grants-in-Aid for Scientific Research (B)
Issue number: 25280061
URL: https://kaken.nii.ac.jp/ja/grant/KAKENHI-PROJECT-25280061/

In this research, we aimed to establish a social infrastructure for the permanent collection, accumulation, sharing, and maintenance of voice data, and proposed a framework for collecting more voice data by developing voice recording tools and designing appropriate licenses for voice data. We also proposed a framework for developing a voice model that can express various voice qualities and building a voice synthesis system from unstructured voice data, and worked on building a voice synthesis system that can realize various speech expressions.

Back to list