How to Dub a Video Into Five Languages With Altered Without Re-Shooting
Altered’s dubbing path takes one source video, transcribes the speech, translates it, and re-voices it in a target language while keeping the original timing. Five languages means five passes over the same timeline.
The tool does the heavy lifting. The parts that fail are usually upstream — bad transcript, wrong number of syllables, or a source clip with overlapping speakers.
What is the correct order of operations for dubbing?
- Export the source audio as 48 kHz WAV, not the video file. Altered works better with clean audio than with a multiplexed MP4.
- Let Altered transcribe, then read the transcript. Fix names, product terms and numbers before translating. Errors here multiply by five.
- Lock the glossary. Add brand names to the term list so they are never translated.
- Translate. Pick the target voice from the matching language library, not a converted English voice.
- Re-voice and check syllable alignment on the longest line.
- Recombine with the original video in your editor, keeping the original music and effects track under the dubbed voice.
Why does dubbing drift out of sync?
Because languages have different syllable budgets. A German line for an English sentence runs about 30% longer; Japanese runs shorter. When the target voice cannot compress, it either speeds up unnaturally or overruns the shot.
Altered handles moderate drift by adjusting pacing. Past roughly 15% overrun, it sounds rushed.
Fix it at the script stage. Ask the translator for “dub-friendly” output: same meaning, same approximate syllable count, no idioms. That one instruction removes most sync work.
What does a five-language dub cost in practice?
Assume a 10-minute video, about 1,400 words of speech.
| Item | Rough cost |
|---|---|
| Transcription review, 1 hour | your time |
| Altered processing, 5 languages | plan minutes, roughly 50 min of audio |
| Human spot-check per language, 30 min each | 2.5 hours |
| Realistic total | 4-5 hours of work |
The machine part is fast. The review part is not, and skipping it is how brands end up with a Spanish dub that says the wrong price.
Always have a native speaker check the number-heavy segments. Currency, dates and product SKUs are where machine dubs fail.
Where does Altered’s dubbing stop being enough?
Lip sync. It does not re-animate mouths. For talking-head marketing with a tight face shot, viewers notice within two seconds. Pair it with a lip-sync tool or use subtitles instead.
Multiple speakers. Diarisation handles two voices okay. Three or more overlapping speakers in a panel recording gets messy — split the audio per speaker first.
Heavy accents in the source. Convert from accent to neutral before translating, or the target-language voice inherits the confusion.
Legal. Synthetic voice disclosure is required in several jurisdictions. Put a line in the description.
When is a human dub still the right call?
Premium launch films, anything shown on TV, and content where a celebrity voice is the product. Also dubbing into languages with strong formality rules — Korean and Japanese honorifics do not survive a machine pass reliably.
The honest framing: use Altered for volume, use humans for the five clips that matter most. Most teams do both, in that order.
If your bottleneck is actually the video edit that wraps the dub, Blizzy handles assembly. If the source is a recorded meeting or interview, Avoma gives you transcripts to translate from.
Quick Summary
- Export clean 48 kHz audio first, review the transcript, then translate — in that order.
- Glossary entries for brands and SKUs save five passes worth of fixes.
- Languages differ 20-30% in syllable length; ask translators for dub-friendly scripts.
- Altered re-voices but does not re-animate lips. For face-forward shots, add lip sync or subtitles.
- Budget 4-5 hours of human review for a 10-minute video across five languages.
