The USTC System for Blizzard Challenge 2008

This paper introduces the speech synthesis system developed by USTC for Blizzard Challenge 2008. Two synthetic voices from the released UK English database are built using the HMMbased unit selection synthesis method, which is a hybrid of statistical parametric synthesis and unit-selection techniques. In this method, the optimal sequence of phone-sized candidate units is selected from the database following the statistical criterions derived from a set of trained HMMs for different acoustic features. Then the waveforms of selected units are concatenated to generate the synthesized speech. The evaluation results of Blizzard Challenge 2008 show that our system has good performance on similarity, naturalness and intelligibility for both English voices. Index Terms: speech synthesis, Blizzard Challenge, unit selection, hidden Markov model