Researchers have introduced LuxEmo, a new 21-hour corpus of expressive speech for the Luxembourgish language, addressing the underrepresentation of low-resource languages in speech technology. The corpus, derived from Radio Télévision Luxembourg (RTL) youth broadcasts, features four emotion categories and was curated using a semi-automatic workflow involving automated detection and human validation. The work also includes benchmarking of five text-to-speech (TTS) systems to evaluate their performance on Luxembourgish, assessing cross-lingual transfer, adaptation, and prosody transfer capabilities. AI
IMPACT This release aims to improve speech technology for underrepresented languages, potentially enabling new applications and research in expressive TTS for Luxembourgish.
RANK_REASON The cluster describes the release of a new academic paper detailing a novel dataset for a low-resource language.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →