How do Voices from Past Speech Synthesis Challenges Compare Today?
Citations Over TimeTop 10% of 2021 papers
Abstract
Shared challenges provide a venue for comparing systems trained on common data using a standardized evaluation, and they also provide an invaluable resource for researchers when the data and evaluation results are publicly released. The Blizzard Challenge and Voice Conversion Challenge are two such challenges for text-to-speech synthesis and for speaker conversion, respectively, and their publicly-available system samples and listening test results comprise a historical record of state-of-the-art synthesis methods over the years. In this paper, we revisit these past challenges and conduct a large-scale listening test with samples from many challenges combined. Our aims are to analyze and compare opinions of a large number of systems together, to determine whether and how opinions change over time, and to collect a large-scale dataset of a diverse variety of synthetic samples and their ratings for further research. We found strong correlations challenge by challenge at the system level between the original results and our new listening test. We also observed the importance of the choice of speaker on synthesis quality.
Related Papers
- → The Influence of Listening English Song to Improve Listening Skill in Listening Class(2017)54 cited
- → Are You "Wanted" for Poor Listening Habits?(2015)
- → Listening to the Other – A New Approach to Counselling and Listening SkillsListening to the Other – A New Approach to Counselling and Listening Skills(2010)
- → Can You Hear Me Now?(2015)
- → White Liberals Need to Stop Talking and Start Listening(2019)