Multi-Dialectal HMM-based Speech Synthesis

This is a demonstration for HMM-based speech synthesis system
using the Unisyn Lexicon for multi-dialectal speech synthesis.
The Unisyn lexicon allows the encoding of multiple accents of English.
This can be used for HMM-based speech synthesis and acheive several
distinctive dialectal voices and multi-accented average voice models easily.

Demonstrations (English)

Dialect
U.S. English
U.K. English
Scottish English
Indian English
Sample 1
Sample 2
Sample 3
Sample 4


Application (Interpolation of Dialects)

From Indian English to American English
From American English to Indian English

Reference:
Junichi Yamagishi, Heiga Zen, Yi-Jian Wu, Tomoki Toda, Keiichi Tokuda
``The HTS-2008 System: Yet Another Evaluation of the Speaker-Adaptive
HMM-based Speech Synthesis System in The 2008 Blizzard Challenge,''
Proc. Blizzard Challenge 2008, 2008.