By Higgsfield Genjutsu editorial teamPublished Nov 2, 2026Updated Nov 2, 2026
Does Genjutsu keep the original audio?
Do not assume that the final Genjutsu file preserves the source soundtrack. The current workflow is designed around transforming the visual performance, and the site does not present itself as a music, voice or audio-finishing tool.
Check the downloaded result before planning a publish. If the sound is missing, changed or out of sync, add the approved audio in an editor and review the final sequence there.
Keep visual generation and audio finishing as separate jobs.| Task | Genjutsu | Post-production |
|---|
| Rebuild the visual performer or scene | Core workflow | Review the generated clip |
|---|
| Generate the song or voice performance | Not provided | Use a music, voice or recording workflow |
|---|
| Align the final track with the full edit | Not a timeline tool | Sequence, sync and mix in an editor |
|---|
| Add exact lyrics, captions or disclosures | Do not rely on generated text | Add deterministic text in post |
|---|
Why does the source video still matter for timing?
Even when audio is finished separately, the source video gives the visual performance its rhythm. Head movement, gestures, footwork and camera timing come from the clip; that shape can be matched to a new track in post.
For a lip-sync-focused shot, use footage with clear mouth and head movement and budget time for manual review. Genjutsu follows the source timing, but that is not the same as guaranteed frame-accurate replacement of a new vocal.
- Use a clean performance with readable movement.
- Keep the phrase short enough to review frame by frame.
- Record or license the final voice or song separately.
- Compare mouth, head and gesture timing after generation.
- Make final speed and cut changes in the editor.
Can Genjutsu make an audio-driven video from a song?
No. On this site, the tool needs a source video and one character image. Uploading a song does not give the model choreography, camera motion or a complete visual performance.
You can use audio as creative direction for the final edit, but the visual source still has to exist. A music-video project is therefore a sequence of generated performance shots plus an editor for sound and assembly.
How should you finish Genjutsu audio in an editor?
Import the approved visual clips, the licensed track or voice file, captions and any sound design into an editing project. Set the final cut to the music or dialogue, then review every transition at phone size.
Use a limiter or loudness target appropriate to the destination, but do not assume the platform will fix clipped or badly mixed audio. Keep a clean master outside the social export.
- Place the visual clips on the main timeline.
- Add the approved music or voice track.
- Review sync at every cut and close-up.
- Add exact captions and synthetic-media disclosure.
- Normalise loudness and export a platform-ready master.
What audio rights do you need?
Rights to the visual source do not include rights to a song, master recording, voice sample or another person’s likeness. Each asset may have a different owner and licence.
Use original recordings, licensed music or platform sounds within their actual terms. Do not treat “available on social media” as permission to redistribute or monetise.
Does Genjutsu create lip sync?
The generated visuals follow the timing of the uploaded source performance, but the site does not provide a dedicated lip-sync replacement workflow for arbitrary audio. Audio replacement and exact mouth-shape review remain editing tasks.
If the final deliverable depends on matching new spoken words, use a dedicated dubbing or lip-sync workflow or plan additional post-production. Do not promise a client perfect lip sync from a standard recast alone.
What should you check in the generated file?
Play the complete download with audio enabled before muting it. Confirm whether the file contains sound, then check the duration and the first and last frames. A browser preview with muted autoplay can hide an audio problem.
If the file is silent, that may be expected for this visual workflow. Add the final audio in post. If the file contains unexpected sound, confirm its rights before using it.
- Audio track present or absent?
- Does its duration match the picture?
- Any clicks, dropouts or distortion?
- Is the mouth timing acceptable?
- Do you hold rights to every sound used?
Key takeaways
- Do not assume the source audio survives the visual recast.
- Genjutsu does not generate the song, voice or final soundtrack.
- Lip sync, captions, mixing and sequencing belong in an editor.
- The source video still defines visual motion and timing.
- Verify every sound, likeness and music right before publishing.
Higgsfield Genjutsu