Infoscience

Report

The SIWIS French Speech Synthesis Database ? Design and recording of a high quality French database for speech synthesis

We describe the design and recording of a high quality French speech corpus, aimed at building TTS systems, investigate multiple styles, and emphasis. The data was recorded by a French voice talent, and contains about ten hours of speech, including emphasised words in many different contexts. The database contains more than ten hours of speech and is freely available.

    Reference

    • EPFL-REPORT-225946

    Record created on 2017-02-19, modified on 2017-05-10

Related material