Recording your own performance happens in a Guided phone MV project. Auto teaser and Full AI fill projects don’t offer recording, so choose Guided phone MV when you create the project if you plan to film. Recording and uploading takes cost no credits.
Film a take in the app
- On Scenes, use the scene’s Record button, or “Record scene 3” on the next-up strip. The Record tab opens with that scene selected, along with any recording instructions the director wrote.
- Choose “Start Recording” to open the camera. Your browser asks for permission the first time.
- Choose “Start Recording” again. Your song plays from the start of the scene during a short countdown, so you can find your place, then recording starts.
- Perform along with the track. Recording stops by itself at the end of the scene. “Stop Recording” ends it early and keeps what you filmed as a new take. “Stop” asks you to confirm, then throws the take away.
The take uploads (later takes show a progress bar), and you see Scene 3 recorded! The scene is now ready.
If the song has no audio linked, you can still record, with no guide track.
Upload footage instead
Use “Upload take” on the Record tab to pick a video file from your device, such as footage from another camera. Video files can be up to 500 MB.
The newest take becomes the scene’s main one, and earlier takes are kept.
Choose which take the render uses
When a scene has more than one take, they appear as Take 1, Take 2 and so on, with the current one marked “Primary”. Select a take to watch it, then choose “Use for assemble” to make it primary. You see Primary take set.
A scene can have a take and an AI clip at once. “Assemble uses:” under the player shows which one the render will use. Tap the badge to switch; a pin shows the choice is fixed, either by you or by your newest take or clip. You see Assemble media updated. Left alone, the most recent one wins.
Check a take against the song
“Sync Audio” plays the take with the song and scrolling lyrics. Nudge it by a tenth of a second, or use the slider up to two seconds either way, then choose “Confirm Sync & Save Offset”. You see Audio sync saved. You can mute the recording or the song while you judge it.
If a take still looks off in the rendered video, record it again with the guide track, or lip-sync it.
Enhance a take
Under “More” on the Record tab, “Enhance” re-renders your primary take with the video model, guided by the scene’s description, over the scene’s stretch of the song. It’s billed like a generated clip, and the dialog shows the price. The result becomes the scene’s clip for the render. This is only offered on Guided phone MV.
Lip-sync
Lip-sync redraws the mouth on an existing clip to match the song.
On Guided phone MV, go to the Record tab and choose “More”, then “Lip-sync”. It’s there once the scene has a recorded or uploaded take. For a scene with only an AI clip, ask the Visual Director in chat to lip-sync it.
On Auto teaser and Full AI fill, open the scene, then in Scene tools choose Lip-sync to song. It’s there once the clip is ready. “Lip-sync all” in the scene list does the ready performance scenes, up to 20 at a time. It leaves out any already in sync, so they aren’t charged again. To run one again, use the scene’s own lip-sync.
You see Lip-sync this scene? with the price. Choose Continue. The result is saved as the scene’s enhanced clip and used in the render by default. On Guided phone MV with a take, you can switch back with the “Assemble uses:” badge. On Auto teaser and Full AI fill there’s no switch back: Re-generate makes a new clip and bills again.
Cost: 20 credits per second of the scene, 2 seconds minimum. Lip-sync covers the first 10 seconds of a scene at most, so split longer ones. The dialog shows the real figure; prices can change. It costs no credits on a workspace using its own key for video.
Clips made with Performance mode (singing close-up) are already generated from the vocal and don’t need this. See generating clips.
When lip-sync doesn’t work
- “This scene has no performer on camera to lip-sync.” B-roll scenes and shots with no cast can’t be lip-synced. Give the scene a cast on its page, or leave it as it is.
- A message about needing a linked song or song audio. Link a song with audio in the bar at the top of the studio.
- A message that the scene has no video clip. There’s nothing to sync yet, so record or generate one first.
- Lip-sync failed, sometimes with a reason. It needs a clear face in the clip and song audio linked. Try a take where the face is larger and well lit, then run it again.
When recording doesn’t work
- “Unable to access camera” means your browser has blocked it. Allow the camera for this site in the browser’s settings, then “Try Again”.
- “Camera API not supported in this browser.” Switch browsers, or film with your phone’s camera app and use “Upload take”.
- “Could not play your song.” Check your volume and that the track is linked, then start again.
- “Failed to upload recording.” Check your connection and the file size, then upload again.
What to do next
When every scene shows as ready, assemble and edit your cut.
