Just upload a picture and audio, and she can speak or sing. The Wan2.2 S2V model can generate continuous videos of over 15 seconds at a time. I optimized the workflow and the result was excellent. The workflow I built is in the attachment. Everyone can download and experience it.