By Higgsfield Genjutsu editorial teamPublished Oct 31, 2026Updated Oct 31, 2026
Can you make a music video with Genjutsu?
Yes, when the music video starts from real performance footage. Genjutsu recasts a 3 to 30 second source clip, keeping the performer’s movement, camera and timing while references rebuild the cast, wardrobe or world.
It does not generate the song, invent a complete performance from text or replace a full editing suite. The practical workflow is Genjutsu for controlled visual shots, followed by an editor for sequencing, music, captions and final delivery.
Match the tool to the part of the music video you need.| Need | Best workflow | Genjutsu boundary |
|---|
| Recast a real dance, walk or performance | Genjutsu Motion Transfer | Keeps the source action and camera while references rebuild the scene |
|---|
| Generate a performance that was never filmed | Text-to-video or virtual-avatar tool | Genjutsu requires a source video rather than inventing choreography |
|---|
| Cut shots to the song and add sound | Traditional video editor or NLE | Genjutsu does not generate the track or assemble the final timeline |
|---|
What performance footage works best for a music video?
Use clean, single-angle footage with a readable silhouette and room for the camera to move. A 3 to 10 second phrase is easier to control than a full routine. Longer sections can be split into shots and assembled on the beat.
Choose movement that communicates the song: a direct-to-camera line, a walk, a dance phrase, an outfit reveal or a final pose. Keep feet, hands and face visible often enough for the references to support the performance.
- One continuous take without internal cuts or speed ramps.
- Even exposure that preserves face, hands and costume detail.
- A camera path you want to keep in the finished shot.
- A clear opening action and a loopable or decisive ending.
- Source resolution and framing appropriate for the final vertical or widescreen edit.
How do you plan a multi-shot Genjutsu music video?
Treat each source clip as one visual unit. Build a shot list around the performance you actually have: verse close-up, full-body chorus, movement transition and hero pose. Recast one unit at a time instead of forcing a long montage into a single render.
Keep the character references consistent across every unit. Change the source performance and scene deliberately, then assemble the approved clips against the track in the editor.
- Shot 1: 3 to 6 second direct-to-camera hook.
- Shot 2: full-body walk or dance phrase.
- Shot 3: close performance beat or outfit detail.
- Shot 4: final pose, room reveal or transition plate.
- Post: sequence to the song, grade, caption and deliver.
What references keep the music-video performer consistent?
Use one canonical identity set and separate wardrobe sets. The identity set should include front, three-quarter, profile and full-body views. Permanent details such as hairline, tattoos or glasses stay in the identity set; campaign outfits change separately.
If the video includes more than one performer, create a separate reference set and explicit left-right ownership. Avoid mixing multiple people into one identity folder.
How do you prompt a Genjutsu music-video shot?
Protect the source movement, camera and timing before describing the new visual. Then name the identity and the one scene variable you are changing. The references define appearance; the prompt resolves what must stay.
Do not describe a new edit rhythm. Generate the shot from the source performance and match it to the song later.
Performance close-upRecast the performer as the fictional identity in reference set A. Preserve the exact speech timing, head movement, camera height and light direction. Keep the same face and hair throughout.
Full-body chorusUse identity set A and rebuild the dancer’s outfit from wardrobe set B. Preserve every step, gesture, camera move and cut timing from the source performance.
New stageKeep the performer, motion and camera unchanged. Rebuild only the stage from the location reference, matching the floor plane, light direction and lens perspective.
How do music, lip sync and captions fit the workflow?
Genjutsu handles the visible recast. Add the finished track, voice, captions and final lip-sync review in an editor. A source performance with clear speech timing gives the recast a stronger base, but the audio stays a separate rights and editing decision.
Keep legal or lyric text out of the generated frame. Add exact captions and disclosures deterministically in post so every character can be reviewed.
How does Genjutsu compare with AI music-video generators?
Viggle and HeyGen both present music-video workflows, but they emphasize different starting points. HeyGen focuses on AI-generated visuals, avatars and prompt-led production; Viggle is built around character animation and motion sources.
Genjutsu’s difference is the production clip. When you already have a real performance, the recast can preserve the actual weight, camera language and timing instead of generating a new performance.
What should you review before publishing the music video?
Review each shot for identity, body ownership, hand contact, feet, costume and camera. Then watch the assembled sequence at phone size with the final audio. A clean still frame does not prove that the performance survives movement.
Confirm rights for the footage, references, music and likenesses. Disclose synthetic media where required and do not imply that a real artist, brand or collaborator participated without permission.
Why does the performer change between shots?
A new reference set, different reference order, a new source angle or a different light direction can all read as a new person. Outfit changes can also alter silhouette and body proportions.
Return to the canonical identity set before changing the prompt. If only the face drifts, add a matching profile and front view. If the silhouette changes, check the full-body and wardrobe references.
Key takeaways
- Genjutsu music videos begin with real source performance footage, not text alone.
- Plan and generate one short shot at a time, then assemble to the track in an editor.
- Use a canonical identity set and separate outfit sets for a consistent performer.
- Music, voice, captions and final sequencing remain post-production tasks.
- Publish only with rights to the footage, music, references and likenesses.
Higgsfield Genjutsu