DiVISe: Direct Visual-Input Speech Synthesis Preserving Speaker Characteristics And Intelligibility

Re-assign community