How to Translate a Video: Subtitles, Dubbing and Lip Sync
Choose between translated subtitles, dubbed audio and lip sync. Learn how to review English or Spanish video translations, speaker turns and timing.
To translate a video, decide whether viewers need translated text, translated speech or both. A subtitle track, a dubbed soundtrack and adjusted mouth movements are different outputs. Choosing the right one before processing makes the result easier to review.
The AI Video Translator workspace dubs videos up to 5 minutes long into 29 languages. Choose a preset voice, or clone voices from authorized source speech, and optionally enable lip sync. Only speech is translated: subtitles and other on-screen text stay in the original language. This guide explains how to prepare and check a translation workflow.
1. Choose the output your audience needs
| Output | What changes | What to review |
|---|---|---|
| Translated subtitles | Timed on-screen text | Meaning, reading speed and line timing |
| Dubbed audio | Spoken words in another language | Pronunciation, pacing and speaker turns |
| Lip-synced video | Visible mouth movement as well as translated speech | Mouth shapes, expressions and audiovisual timing |
Subtitles can preserve the original performance while making the words accessible in another language. Dubbing lets viewers listen without reading. Lip sync adds a visual processing step; selecting a target language alone does not imply it is included. Adobe's video translation guide treats lip sync as a separate option.
2. Check the source speech before translating
Listen to a short section with headphones. Note the spoken language, unclear words, interruptions and any music competing with speech. If the source includes several languages, check whether the selected processing service supports that situation.
A video does not need existing subtitles to be translated from speech. A suitable workflow can first transcribe the audio, then translate that transcript. Transcription answers what was said in the source language; translation expresses its meaning in another language.
Correct names, numbers and technical terms before generating a dub when the service allows editing. An early transcription mistake can otherwise appear in both the translated text and the new voice track. Keep a short glossary for recurring product names and specialist vocabulary.
3. Translate English or Spanish for the intended listener
Choose the actual source language and the intended target language. For an English-to-Spanish clip, “Let's get started” might become “Empecemos.” For Spanish-to-English, “Nos vemos mañana” becomes “See you tomorrow.” These are illustrative phrases; a full sentence's context determines the right translation.
Review the whole idea rather than matching every word. Check regional vocabulary, tone, units and how numbers are spoken. A translation may require a different amount of time to say, so listen for rushed endings or pauses that feel misplaced. YouTube's automatic dubbing guidance notes possible errors with names, idioms, accents and background noise.
4. Review speaker turns without assuming identity
For an interview or conversation, verify who speaks each sentence before assigning translated voices. Speaker separation, often called diarization, groups speech into turns such as Speaker 1 and Speaker 2. It does not establish a person's real-world identity.
Microsoft's diarization documentation describes these labels as generic identifiers. Short interruptions and people talking over one another deserve a manual check. Listen for sentences assigned to the wrong voice or a speaker label changing halfway through the clip.
5. Distinguish watching a translation from publishing one
When watching YouTube, use an available subtitle language or Auto-translate in the caption settings where offered. This changes the viewing experience; it does not create a translated video file. YouTube's caption help explains the viewer controls. For your own uploads, YouTube Studio supports adding subtitle languages through a separate creator workflow.
Google Translate can help with transcript text or supported live speech features, but its document upload is not a finished video-dubbing workflow. Its document guide lists document formats rather than video files. Translating words still leaves subtitle timing, speech generation and video assembly to other tools.
6. Check the complete translated version
Before keeping a result, compare its meaning against the source, then watch at normal speed. Check caption readability, pronunciation, speaker changes and the balance between speech and background sound. For lip sync, inspect mouth movement and expression at cuts and turns. Ask a fluent target-language speaker to review important wording.
Open the AI Video Translator to upload an MP4, MOV or WebM up to 2 GiB and 5 minutes, choose a voice and target language, and create a paid translation task. Local preview is free. Completed tasks provide a downloadable MP4; generated illustrations and sample translations on the page are examples. To speak a new script without translating the source, use the separate Lip Sync tool with an MP4 and a preset voice.