USE WHEN
Use this manual when
- Chinese-to-English localization or the reverse
- Multi-speaker productions with names and retained music/effects
- Teams delivering MP4, WAV, SRT or AAF packages
PRODUCTION MANUAL · Audio & localization
Bilingual short-drama dubbing: from locked dialogue to a deliverable audio master
Unify terminology, voice rights, speakers, timecodes, captions and stems in one dubbing QC process instead of publishing automatic output unchecked.
DO NOT USE AS
Do not use it as
- Cloning a real person's voice without authorization
- Automatic publication without a target-language reviewer
01
VERIFIED FACTS
What the evidence currently establishes
- 01
ElevenLabs currently documents 90+ languages, speaker separation and retained background audio; Automatic Dubbing v2 remains labelled Alpha and recommends no more than nine unique speakers for best results.
- 02
ElevenLabs documents exports including MP4, AAC, AAF, SRT and separated WAV tracks, allowing editable delivery instead of a single flattened video.
- 03
YouTube warns that automatic dubs can fail on proper nouns, idioms, accents, dialects, background noise and speaker matching. Auto dubs are not directly editable, so target-language review should precede publication.
02
REQUIRED INPUTS
Prepare before starting
- Locked source-language dialogue with timecodes
- Character, speaker and voice-rights matrix
- Names, places, terminology, catchphrases and do-not-translate list
- Lossless master audio, dialogue/music/effects stems and clean picture master
- Platform loudness, caption, aspect-ratio and credit requirements
03
WORKFLOW
Execution workflow
- 01
Verify source dialogue, speakers and timecodes before translation.
- 02
Build a bilingual glossary and flag names, wordplay and cultural references for human work.
- 03
Turn literal translation into performable dialogue while preserving meaning, relationships and duration.
- 04
Generate test segments only with authorized voices and check timbre, accent, emotion and separation.
- 05
Review meaning, pronunciation, pacing, lip tolerance and background-audio damage line by line.
- 06
Export video, dialogue stems, captions, glossary and revision log with a path back to the source master.
04
DELIVERABLES & QC
Deliverables and acceptance
Required deliverables
- Bilingual dialogue and glossary
- Voice-rights matrix
- Target-language video master
- Dialogue/music/effects stems
- SRT/captions and revision log
QC checklist
- Names, places and canon remain consistent
- Every character voice has documented permission
- Emotion, meaning and relationships are preserved
- Dialogue does not mask key information or damage the background
- Captions, track names and language tags match platform requirements
EVIDENCE BOUNDARY
What this manual does not prove
Language count, separation and export formats come from official documentation and do not guarantee broadcast quality for every language, accent or mix. Similarity does not replace voice rights; final dubs require target-language and audio QC.