AI Dubbing Editor: Fix Words, Timing and Mix

What you can change

Words

Retype any translated line; the original words sit beside it so you can check the meaning

Take and delivery

Ask for another take of a line, or a calmer, more natural or livelier reading of it

Timing

Drag a line on the timeline or nudge it by 0.1 s, up to 3 seconds from where it was said, without crossing its neighbours

Line level

Raise or lower one line by up to 12 dB

Keep original voice

Leave a line as it was said, for a name, a song or a quote that shouldn't be translated

Voice and background

Turn the whole new voice or the music and effects up or down, or remove the background entirely

Background by section

A different background level for each part of a long file, such as a loud intro

Leftover speech cleanup

Off, normal or strong removal of faint original speech left in the background

Music ducking

How far music dips while the new voice speaks: off, light, normal or strong

Loudness

As mixed, or normalized to -14 LUFS for the web, -16 LUFS or -23 LUFS for broadcast

Subtitles

Subtitles in the dub's language or the original, from your corrected script, with style presets, a live preview and burn-in for videos

Each person keeps their own copied voice; there is no voice library to swap in.

Hear it before you pay

The editor plays your dub with the new voice and the background as separate layers, so voice, background and line levels change as you drag the slider, and switching between Dub and Original lets you compare any moment with what was said. Waveforms for the original, the voice and the background sit on the timeline, so loud music and quiet lines are visible before you listen. Changed words, delivery, timing, leftover-speech cleanup, ducking and loudness are applied when you re-process; the editor lists which of your changes the preview can't play yet.

Guided fixes, so one re-process is enough

Re-processing is where credits go, so the Review tab checks the script first and points at each line that needs attention:

• Lines with no words, and lines longer than 600 characters, must be fixed before you can re-process.
• Lines that will have to be sped up too much to fit their slot, with how many characters to cut or how much room to add.
• Lines that look untranslated, because they match the original words.
• Lines that never got a voice, which are voiced again at no cost.
• Subtitles that are too fast to read.

Each line in the script is marked when it will be re-voiced (and charged), voiced free, kept in the original voice, or edited without re-voicing, and the top bar shows the cost of everything you've changed. Undo and redo cover every change, and unsaved work is kept in your browser if the tab closes.

What re-processing costs

Mixing is free: timing, line and background levels, leftover-speech cleanup, ducking, loudness, keeping the original voice and subtitles never use credits. Only lines whose words, take or delivery changed are voiced again, at 50 credits per minute of those lines, the same rate as dubbing, with a 25-credit minimum per re-process. Before you confirm, the exact price is shown and nothing is charged if it changed.

If a re-process fails on our side, its credits are refunded automatically. A dub can be re-processed up to 40 times, and you can make several fixes in one go to pay the minimum once.

How to use it

1. Dub a video or audio file with "Full translation output" on. When it finishes, the editor opens on the same page, just below the tool; you can also reopen it from My files while the dub's files are kept.
2. Watch with subtitles on and press A to compare any line with the original.
3. Fix lines in the script: retype words, try another take, move or nudge the timing, or keep the original voice.
4. Set the overall sound on the Sound tab, and the subtitle style on the Subtitles tab.
5. Open Review, clear anything marked as an error, and press Re-process. The new dub replaces the old one in your downloads, with subtitles that follow your corrected script.
6. Happy with the dub? Skip re-processing and use Download for the translated video, the audio or the subtitles.

Keyboard shortcuts: Space plays and pauses, the arrow keys move between lines and seek, and Ctrl or Cmd with Z undoes a change.

Your files while you edit

Your video still never leaves your device; the editor works on the dub's audio, which stays on our servers so lines can be voiced again in the same voice. A dub's audio, script and voice files stay on our servers for 30 minutes after it finishes or after you last use the editor, so you can fine-tune it, and are then deleted; the editor's "Delete files" button removes them at once. Nothing is used to train models.

What it doesn't do

• It edits the dub mydubly made: you can't bring your own script or SRT to voice, or add new lines.
• No lip-sync and no voice library.
• No pronunciation dictionary; to fix a name, respell it the way it sounds or try another take.
• One language per dub. For another language, dub the file again.

Using the tool on this page

• The spoken language is detected automatically; you select the output language — the spoken language itself for an untranslated transcript, or another language for a translation. Files can be up to 2 hours long.
• Only the audio is sent for processing — the video stays on your device. The finished dub opens in the dub editor, where you can correct words, timing, levels, background and subtitles and re-process only what changed. Its files are deleted 30 minutes after you last use the editor (how files are protected).
• You can download a translated video that keeps your music and effects, the translated audio as M4A, a transcript and subtitles (SRT and VTT). Every output, step and limit is explained on the AI dubbing page.