ResearchBenchmarksVideo-to-Speech Synthesis on LRS3 (test) using GE2E and VoxSimFollow0.65GE2EOurs0.607360.618430.62950.64057Mar 21, 2025Evaluation ResultsMethodMethodLinksGE2EVoxSimOursSteps=10Steps=102025.030.650.495OursSteps=1000Steps=10002025.030.650.494DiffV2SSteps=1000Steps=10002025.030.6210.433LTBS2025.030.6090.399