Research Interests †

Acoustic modeling for statistical speech recognition & synthesis †

• Phonetic, dimensional & state positional decision tree (PDS-DT) based context clustering (From 2001)
• Trajectory model derived from HMM (trajectory HMM) (From 2002)
• Hidden semi-Markov model (HSMM) based speech synthesis (From 2003)

HMM-based speech synthesis system (HTS) †

• Former maintainer

Applying HTS to various languages †

• Japanese
• English
• German
• Brazilian Portuguese
• Mandarin

Developing Festival speech synthesis system †

• Implementing HTS-related modules for Festival

Automatic voice building †

• Accent recognition

Text processing for TTS †

• Designing prompt texts for recording
• Predicting prosodic events using statistical models