01What “own-voice” audio dubbing means
When you choose Keep my own voice, we pick the cleanest 10 to 60 seconds of speech from the file you uploaded — the stretch with the most talking and the least background noise — and use it to build a voice clone. That clone then reads your approved translation, sentence by sentence.
The result is not a perfect copy. Your timbre and general delivery carry over well; accent, rhythm and emphasis naturally change in another language. For many podcasters, teachers and coaches that is exactly the point: the audience gets their language, and you remain recognisably you.
02Consent comes first, every time
Cloning only happens after you confirm that the voice in the file is yours or that the speaker gave you explicit permission. We store the time of that confirmation with the job. If you would rather not clone at all, pick Use a standard voice and a stock narrator for the target language reads the script instead.
If your recording has guests, every person who speaks must agree. The current version treats a file as one main speaker; multi-speaker dubbing with separate voices is on the roadmap.
03Dubbing that follows the original timeline
Each translated sentence is placed at the timecode where the original sentence began. If a translation runs long, the voice is sped up slightly (within 0.85–1.15×) rather than pushing everything after it out of sync; the gaps between sentences are kept close to the original pauses.
That makes the dub easy to drop back into a video edit, a course module or a podcast feed with the same chapter markers. Note that this is audio only: we replace the soundtrack and do not alter video or lip movements.
04Fix the words before the voice is made
Machine translation is good, not infallible. After transcription and translation, the job pauses so you can read every sentence next to its source. Numbers, percentages and dates are checked automatically, and any line where a figure is missing is flagged.
Edit what you like; your changes save as you type. When you approve, only then is audio generated — and if you change a sentence later, only that sentence is re-voiced.
05Who uses it
- Podcasters publishing a second-language feed of their own show
- Course creators and lecturers localising recorded lessons
- Coaches, founders and authors sharing talks with a new audience
- Researchers and journalists who need a listenable version of their own interviews
QQuestions
Q.01Is audio dubbing the same as translating an audio file?
Translating gives you the words in another language; dubbing turns those words back into speech that replaces the original track. We do both in one job: you get the bilingual transcript and the dubbed MP3.
Q.02Do you keep my voice clone?
The clone is used for your job only. Uploaded audio, transcripts and outputs are deleted after 7 days. The clone itself is held by our speech provider; email us and we will request its deletion.
Q.03Can I dub a video?
Export the audio from your video, dub it here, then replace the audio track in your editor. Timing follows the original, but we do not do lip-sync.