一作ACL 2026Poster· Natural Language Processing
U ni V ocal: Unified Speech-Singing Code-Switching Synthesis
YuFei Shi · Qian Chen · Wen Wang · Xiangang Li · Zhen-Hua Ling · Yang Ai
摘要
We propose UniVocal, a unified framework that implicitly infers vocal modes from text context to pioneer Speech-Singing Code-Switching (SCS) Synthesis—a task where transitions are autonomously driven by textual semantics, akin to seamless human language blending. Unlike single-mode generation or...