A Holistic Cascade System, benchmark, and Human Evaluation Protocol for Expressive Speech-to-Speech Translation
This work addresses the lack of standard benchmarks and evaluation protocols in expressive S2ST, which is incremental as it builds on existing techniques but integrates them into a more comprehensive framework.
The paper tackles the problem of expressive speech-to-speech translation by proposing a holistic cascade system that combines multiple prosody transfer techniques, outperforming single-aspect systems and enabling bilingual annotators to assess expressive preservation.
Expressive speech-to-speech translation (S2ST) aims to transfer prosodic attributes of source speech to target speech while maintaining translation accuracy. Existing research in expressive S2ST is limited, typically focusing on a single expressivity aspect at a time. Likewise, this research area lacks standard evaluation protocols and well-curated benchmark datasets. In this work, we propose a holistic cascade system for expressive S2ST, combining multiple prosody transfer techniques previously considered only in isolation. We curate a benchmark expressivity test set in the TV series domain and explored a second dataset in the audiobook domain. Finally, we present a human evaluation protocol to assess multiple expressive dimensions across speech pairs. Experimental results indicate that bi-lingual annotators can assess the quality of expressive preservation in S2ST systems, and the holistic modeling approach outperforms single-aspect systems. Audio samples can be accessed through our demo webpage: https://facebookresearch.github.io/speech_translation/cascade_expressive_s2st.