Smalltown Boy AI Prompt Guide

Use Video 1 as the exact motion, choreography, and edit reference. Video 1 controls scene layout, background, lighting, camera angles, framing, cut order, step timing, head movements, body sways, gestures, and performance beats. The two reference images control only identity, species-appropriate anatomy, and wardrobe.

Performer map for every frame and every cut (do not reassign by screen position later):
- Image 1 → the dancer who starts on the LEFT at the opening of Video 1
- Image 2 → the dancer who starts on the RIGHT at the opening of Video 1

Replace both original performers in Video 1 with these two subjects. Track Image 1 and Image 2 to their mapped source performer through the whole clip — including when they alternate foreground/background, cross the frame, overlap, or appear in close-ups.

SAME IDENTITY: preserve each subject’s recognizable face, eyes, hair or fur, skin or coat color, markings, body proportions, and outfit from their image for the entire duration. No identity swaps, no fused bodies, no duplicate faces, no return of original Video 1 performers after a cut, no extra people or animals, no changing markings or outfits mid-clip.

Follow Video 1’s action sequence exactly: alternating front-and-back rhythm, retro 80s dance steps, expressive reactions at the vocal peak (including dramatic head-hold / high-note timing), footwork, pauses, and interactions. Adapt gestures to each subject’s anatomy (human hands vs. pet paws) without changing the timing or choreography. Keep stable ground contact; no foot sliding.

Preserve Video 1’s setting, props, light direction, and shot sequence. Do not invent new locations, camera moves, or choreography. Vertical 9:16 — keep both dancers visible when the reference does. Match the selected segment length.

Audio: keep the original soundtrack from Video 1 — Bronski Beat «Smalltown Boy» — locked to the same timing, cuts, and duration as the reference. Do not remove, replace, remix, or mute the music. Mouth and expression animation should follow the vocal phrasing and beat in that audio track.

Use Video 1 as the exact motion, choreography, and edit reference. Video 1 controls scene layout, background, lighting, camera angles, framing, cut order, step timing, head movements, body sways, gestures, and performance beats. The two reference images control only identity, species-appropriate anatomy, and wardrobe.

Performer map for every frame and every cut (do not reassign by screen position later):
- Image 1 → the dancer who starts on the LEFT at the opening of Video 1
- Image 2 → the dancer who starts on the RIGHT at the opening of Video 1

Replace both original performers in Video 1 with these two subjects. Track Image 1 and Image 2 to their mapped source performer through the whole clip — including when they alternate foreground/background, cross the frame, overlap, or appear in close-ups.

SAME IDENTITY: preserve each subject’s recognizable face, eyes, hair or fur, skin or coat color, markings, body proportions, and outfit from their image for the entire duration. No identity swaps, no fused bodies, no duplicate faces, no return of original Video 1 performers after a cut, no extra people or animals, no changing markings or outfits mid-clip.

Follow Video 1’s action sequence exactly: alternating front-and-back rhythm, retro 80s dance steps, expressive reactions at the vocal peak (including dramatic head-hold / high-note timing), footwork, pauses, and interactions. Adapt gestures to each subject’s anatomy (human hands vs. pet paws) without changing the timing or choreography. Keep stable ground contact; no foot sliding.

Preserve Video 1’s setting, props, light direction, and shot sequence. Do not invent new locations, camera moves, or choreography. Vertical 9:16 — keep both dancers visible when the reference does. Match the selected segment length.

Audio: keep the original soundtrack from Video 1 — Bronski Beat «Smalltown Boy» — locked to the same timing, cuts, and duration as the reference. Do not remove, replace, remix, or mute the music. Mouth and expression animation should follow the vocal phrasing and beat in that audio track.

Use Video 1 as the exact motion, choreography, and edit reference. Video 1 controls scene layout, background, lighting, camera angles, framing, cut order, step timing, head movements, body sways, gestures, and performance beats. The two reference images control only identity, species-appropriate anatomy, and wardrobe.

Performer map for every frame and every cut (do not reassign by screen position later):
- Image 1 → the dancer who starts on the LEFT at the opening of Video 1
- Image 2 → the dancer who starts on the RIGHT at the opening of Video 1

Replace both original performers in Video 1 with these two subjects. Track Image 1 and Image 2 to their mapped source performer through the whole clip — including when they alternate foreground/background, cross the frame, overlap, or appear in close-ups.

SAME IDENTITY: preserve each subject’s recognizable face, eyes, hair or fur, skin or coat color, markings, body proportions, and outfit from their image for the entire duration. No identity swaps, no fused bodies, no duplicate faces, no return of original Video 1 performers after a cut, no extra people or animals, no changing markings or outfits mid-clip.

Follow Video 1’s action sequence exactly: alternating front-and-back rhythm, retro 80s dance steps, expressive reactions at the vocal peak (including dramatic head-hold / high-note timing), footwork, pauses, and interactions. Adapt gestures to each subject’s anatomy (human hands vs. pet paws) without changing the timing or choreography. Keep stable ground contact; no foot sliding.

Preserve Video 1’s setting, props, light direction, and shot sequence. Do not invent new locations, camera moves, or choreography. Vertical 9:16 — keep both dancers visible when the reference does. Match the selected segment length.

Audio: keep the original soundtrack from Video 1 — Bronski Beat «Smalltown Boy» — locked to the same timing, cuts, and duration as the reference. Do not remove, replace, remix, or mute the music. Mouth and expression animation should follow the vocal phrasing and beat in that audio track.