How to get the exact Bad Times moves in an AI video

• 0–1.5 s: bent over, head in both hands, straightening up.
• 1.5–2 s: both palms on the cheeks, staring at the phone.
• 2–6 s: hand on the hip, the other flapping and slapping the forehead.
• 6–10.5 s: both palms up, "hold on, listen to me", swaying.
• 10.5–12 s: a turn away, then the charge at the camera.
• 12–15 s: the face at the lens, a finger jabbing at it, a palm slapped at the screen.

An AI video model given a photo and a text prompt invents its own timing. It tends to skip the turn, end before the charge, or keep the camera still when the original shakes and closes in. Motion tools that copy a reference video do better, but you have to supply the clip, and some keep your photo's background instead of the kitchen.

The motion, camera and timing come from the kitchen take itself, and Seedance 2.0 makes every frame new with you in it. It is not a face swap and not a template: your face, hair and clothes come from your photo, the moves come from the original. The free preview shows the first 4 seconds so you can see the timing before you pay.