Create a 10-second ultra-realistic vertical (9:16) comedy video using my reference images. Keep both characters' faces, hairstyle, clothing, body shape, and identity exactly the same throughout the video. Never change or regenerate their faces.
Timing (Strict)
0.0–1.5 sec
Medium shot.
Wife: "রান্নাডা ক্যানং হইল?"
(Wait 0.2 sec after she finishes.)
1.7–3.8 sec
Close-up of husband tasting the curry.
Husband: "বাড়িয়া! খালি একডা সমস্যা।"
(Wait 0.2 sec.)
4.0–4.8 sec
Close-up of wife.
Wife: "কী?"
(Wait 0.2 sec.)
5.0–8.8 sec
Close-up of husband.
Husband: "মুই তরকারি খাছি না... লবণটাই মোরে খাইতাছে!"
Add:
- Funny record scratch before the punchline.
- "Ba dum tss" immediately after the last word.
- Light audience laughter.
8.8–10.0 sec
Split-screen showing both characters simultaneously.
- Husband: funny salty face while drinking water.
- Wife: embarrassed smile with raised eyebrows.
Freeze for the last 0.5 second with a slight zoom-in.
Strict AI Instructions
- Total duration must not exceed 10.0 seconds.
- Each character must wait until the other finishes speaking. Never overlap or interrupt dialogue.
- Insert a 0.2-second silent gap between every dialogue.
- Do not speed up speech unnaturally.
- Prioritize complete dialogue over visual movement.
- Use fast comedy editing with smooth cuts only after each line ends.
- Maintain perfect facial consistency from the reference images throughout the video.