Outlier Detection and Removal for HMM-Based Speech Synthesis with an Insufficient Speech Database

Doo Hwa HONG; June Sig SUNG; Kyung Hwan OH; Nam Soo KIM

doi:10.1587/transinf.E95.D.2351

Regular Section

Outlier Detection and Removal for HMM-Based Speech Synthesis with an Insufficient Speech Database

Doo Hwa HONG, June Sig SUNG, Kyung Hwan OH, Nam Soo KIM

Author information

Keywords: HMM-based speech synthesis, decision tree-based clustering, outlier detection, insufficient speech database

JOURNAL FREE ACCESS

2012 Volume E95.D Issue 9 Pages 2351-2354

DOI https://doi.org/10.1587/transinf.E95.D.2351

Details

Abstract

Decision tree-based clustering and parameter estimation are essential steps in the training part of an HMM-based speech synthesis system. These two steps are usually performed based on the maximum likelihood (ML) criterion. However, one of the drawbacks of the ML criterion is that it is sensitive to outliers which usually result in quality degradation of the synthesized speech. In this letter, we propose an approach to detect and remove outliers for HMM-based speech synthesis. Experimental results show that the proposed approach can improve the synthetic speech, particularly when the available training speech database is insufficient.

Corresponding author

Register with J-STAGE for free!