AI RESEARCH

Mechanisms of Multimodal Synchronization: Insights from Decoder-Based Video-Text-to-Speech Synthesis

arXiv CS.CV

ArXi:2411.17690v3 Announce Type: replace-cross Unified decoder-only transformers have shown promise for multimodal generation, yet the mechanisms by which they synchronize modalities with heterogeneous sampling rates remain underexplored. We investigate these mechanisms through video-text-to-speech (VTTS) synthesis-a controlled task requiring fine-grained temporal alignment between sparse text, video, and continuous speech.