In order to better understand different speech synthesis techniques on the same data, we have devised a challenge that will help us better compare research techniques in building corpus-based speech synthesizers.

The basic challenge is to take the publicly available CMU ARCTIC speech databases and build a synthetic voice. Unknown sentences from an independent source will be generated and each participant will synthesize them with their system. The speech will then be put on the web for evaluation. The results were presented at a special session at Interspeech 2005 -- Eurospeech in Lisboa.

• The Blizzard Challenge -- 2005: Evaluating Corpus-Based Speech Synthesis on Common Datasets (Alan W. Black, Keiichi Tokuda) PDF
• A Probabilistic Approach to Unit Selection for Corpus-Based Speech Synthesis Shinsuke Sakai, Han Shu PDF
• The Blizzard Challenge 2005 CMU Entry -- A Method for Improving Speech Synthesis Systems (John Kominek, Christina L. Bennett, Brian Langner, Arthur R. Toth) PDF
• Automatic Personal Synthetic Voice Construction (H. Timothy Bunnell, Chris Pennington, Debra Yarrington, John Gray) PDF
• An Overview of Nitech HMM-Based Speech Synthesis System for Blizzard Challenge 2005 (Heiga Zen, Tomoki Toda) PDF
• On Building a Concatenative Speech Synthesis System from the Blizzard Challenge Speech Databases (Wael Hamza, Raimo Bakis, Zhi Wei Shuang, Heiga Zen) PDF
• Multisyn Voices from ARCTIC Data for the Blizzard Challenge (Robert A.J. Clark, Korin Richmond, Simon King) PDF
• Large Scale Evaluation of Corpus-Based Synthesizers: Results and Lessons from the Blizzard Challenge 2005 (Christina L. Bennett) PDF