The music and horse-dance choreography
PSY’s Gangnam Style music video is the reference for the crossed-arm horse dance and lasso gesture. This is a choreography-led format: the full body and space around the limbs matter.
Watch PSY’s video ↗Sources & scene guide
One song. Different video formats.
The night-selfie version is a three-person camera performance: a group shot, individual close-ups and a closing shot of the lead performer in the default Seoul reference. Solo, Duo and Trio describe how many identities you replace, not three different dances. The classic horse dance uses crossed arms and bouncing footwork instead.
The music and the camera template have different histories. PSY’s original video is the source for the recognizable horse-dance choreography. The street-selfie format is a separate creator-led reference; a shared song does not make the two scenes interchangeable.
The street-selfie reference
The linked @emiliyyam post is a reference for the group selfie and rotating close-ups. Here, the camera performance defines the template; it is not a demonstration of the classic horse-dance steps.
Open the creator post ↗Photo-based remixes
AI versions use new identity photos to recreate a recognizable scene. A remix post and a commercial template are different sources: neither by itself proves the first inventor, model quality or soundtrack permission.
Open an AI remix reference ↗These links identify reference posts, not a verified first inventor. We do not assign a first-post date or view count without checking the original platform.
Source notes reviewed · 2026-10-26
What happens in the selfie sequence?
Follow three phases rather than copying unverified second-by-second timings. The selected full reference determines the exact cuts.
Group frame
Identify the center, left and right performers before the camera moves.
Individual close-ups
Each close-up must keep the identity assigned to that reference role. Follow the selected clip’s cut order, not the upload number.
Return to the lead
The default Seoul reference closes on the center performer. Keep the lead identity through the final close-up; do not add a group shot that is absent from the reference.
What changes. What may vary.
How does photo-to-video recreation work?
Your photos set identity
Faces, hair, fur and visible clothing guide the result. One clear subject per photo helps avoid confusing two roles. Photos do not need to contain a dance pose.
The template sets the scene
A fixed reference guides movement, framing and camera rhythm. Center-only replacement keeps two reference roles in the intended composition; it is not a frame-perfect face-swap promise.
Review the full result
Generative video can change faces, hands, paws, clothing or timing, especially during overlaps and close-ups. Output review must cover each available scene and cast combination.
Audio depends on the selected reference and final offer. A clip with sound is not a license to PSY’s recording. Check the soundtrack before payment. The available soundtrack is shown with the selected option.
Download your video as a watermark-free MP4. Check the complete reference and selected output specification before generating.
Prepare better photos →