A Comparative Study of the Performance of HMM, DNN, and RNN based Speech Synthesis Systems Trained on Very Large Speaker-Dependent Corpora

Xin Wang, Shinji Takaki, Junichi Yamagishi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This study investigates the impact of the amount of training data on the performance of parametric speech synthesis systems. A Japanese corpus with 100 hours’ audio recordings of a male voice and another corpus with 50 hours’ recordings of a female voice were utilized to train systems based on hidden Markov model (HMM), feed-forward neural network and recurrent neural network (RNN). The results show that the improvement on the accuracy of the predicted spectral features gradually diminishes as the amount of training data increases. However, different from the “diminishing returns” in the spectral stream, the accuracy of the predicted F0 trajectory by the HMM and RNN systems tends to consistently benefit from the increasing amount of training data.
Original languageEnglish
Title of host publicationProceedings of 9th ISCA Speech Synthesis Workshop
Pages125-128
Number of pages4
DOIs
Publication statusPublished - 15 Sep 2016
Event9th ISCA Speech Synthesis Workshop - Sunnyvale, United States
Duration: 13 Sep 201615 Sep 2016
http://ssw9.talp.cat/

Conference

Conference9th ISCA Speech Synthesis Workshop
Abbreviated titleISCA 2016
Country/TerritoryUnited States
CitySunnyvale
Period13/09/1615/09/16
Internet address

Fingerprint

Dive into the research topics of 'A Comparative Study of the Performance of HMM, DNN, and RNN based Speech Synthesis Systems Trained on Very Large Speaker-Dependent Corpora'. Together they form a unique fingerprint.

Cite this