AI dubbing

Modern AI dubbing combines three components: (1) speech recognition transcribes the original audio, (2) translation rewrites it into the target language while keeping length and tone, (3) a cloned or neural voice reads the translation. A lip-sync step can then re-animate the face so the mouth matches the new words.

AI dubbing makes a single source video available to audiences in other languages — useful for course creators, YouTubers and SaaS marketing.

Dubbing, subtitles and voice-over

Subtitles keep the original voice and add translated text, which suits viewers who read along. A voice-over lays a new narrator over the quieted original, as in documentaries. Dubbing replaces the speech itself, ideally in a voice that sounds like the original speaker, so the video feels as if it had been made in the new language. It is the most natural for viewers and the most demanding: timing, names and tone all have to survive the translation.

AI dubbing in VlogMe

DUB translates the speech in a finished video (MP4, MOV or M4V) into one of 28 languages and voices it to sound like the original speaker; the picture itself stays unchanged. Before the dub starts, you confirm that you hold the rights to the video and to every voice in it. The price follows the length of the source video, charged per started minute, and it is shown before you start.

A worked example: a cooking creator has a 6-minute English recipe video that does well. She uploads it to DUB, picks Spanish and then Portuguese, and gets two dubbed MP4s in which the voice still sounds like her. Before publishing, she watches both for dish names and quantities the translation may have changed. For the short vertical cut where she speaks to camera, she runs the Spanish track through Lipsync 2 Pro in Video Studio, so that her lips follow the new words.

Limits and risks

• Recognition and translation are automated: names, figures, jargon and jokes can come out wrong. Watch the dubbed file before you publish it.
• Overlapping speakers, loud music under the speech and a poor microphone lower the voice quality.
• Every voice you dub needs its owner's permission, including for the languages and places where the result appears, and you need the rights to the video (Terms, section 11).
• Each dubbing project, with the voices cloned for it, stays with the dubbing provider until you delete the dub (its DUB project, its dubbed video or its source video), your DUB data or your account, and we then delete it there, as the Privacy Policy explains.

Questions

Does DUB change the speaker's lips?

No. DUB translates and re-voices the speech; the picture is unchanged. Lip sync is a separate step in Video Studio.

How many languages can one video go into?

Any of the 28 target languages DUB supports; each dub is made from the same original, one language at a time.

Should I translate the script myself instead?

For a presenter video made in VlogMe, you can: translate the script, voice it in Audio Studio and render the same portrait again. For filmed footage, DUB is faster.