Details of the Initiative
Speech is a convenient medium for conveying information, for example, announcements in town and product usage guides. Latest technology can synthesize such high-quality speech that many papers say has reached the sound quality of the real human voice. On the other hand, using real human voice is the mainstream for such announcements and guides, and there are still many issues to be addressed in order to spread the use of synthetic speech.
In addition to the development of speech synthesis technology, Morise Laboratory has been conducting research on the construction of fundamental speech databases and public relations strategies to promote its social implementation. Current speech synthesis research uses databases of voices reciting sentences called a corpus. The existing corpus sentences are copyrighted and cannot be obtained easily. Therefore, we have constructed public-domain corpus sentences and distributed databases of speech recited by professional voice actors for social implementation of our research. As a result, the speech of corpus sentences recited by many speakers has become available. Speech synthesis software using the voice database we distribute has been released by third parties, and the number of its applications to audio guides for products and explanations in educational videos etc., has been increasing.
