Reproduce “VoxDub: A Bilingual Lip-Synchronized Audio-Visual Dataset for Automatic Video Dubbing” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Kaidi Wang, Wenhao Guan, Peijie Chen, Weijie Wu, Xuetao Lin, Lin Li, Qingyang Hong
VoxDub creates a Chinese–English corpus for automatic video dubbing by segmenting speech-active regions and screening their transcripts, audio quality, faces, and synchronization. The corpus contains 766,708 clips and about 1.58K hours. In ZipVoice TTS on Seed-TTS test-zh, VoxDub-ZH achieved 2.63% WER and 0.613 SIM-o, versus 12.52% and 0.482 for CN-CVS. These results support TTS gains in this comparison, not a general synchronization advantage.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.