Video Dubbing · Quick Start

Video Dubbing Quick Start

If you are running through NativeVid dubbing for the first time, this page guides you from assets to final output in practical order.

On this page

  • Confirm the video, subtitles, and media folder first, then move into role and voice setup.
  • Pay special attention to review, retries, composition, and audio-video alignment.
  • Run a short sample video first, then move on to larger batches once the workflow feels clear.
Keep this one sentence in mind first: The NativeVid dubbing flow is “asset matching → translation / terminology → role assignment → voice selection → dubbing review → alignment / video composition”. If the subtitles are already in the target language, you can skip translation and shorten the workflow.

What to prepare before you begin

  • A folder that contains both videos and subtitles, preferably with the same batch grouped together
  • Keep video and subtitle filenames aligned as much as possible, for example episode01.mp4 with episode01.srt
  • Common video formats: .mp4, .mkv, .avi, .mov, .webm
  • Common subtitle formats: .srt, .ass, .vtt
Recommended approach: For batch processing, place only videos from the same series, course, or style in one folder so role assignment and overall management stay easier.

Step 1: Choose the media folder and check matching results

Open the “Video Dubbing” page in NativeVid, click “Choose Folder”, and select the directory that contains your videos and subtitles.

  1. The system automatically scans video and subtitle files
  2. It matches them by filename automatically, and matched items appear right away
  3. For videos without a matched subtitle, you can manually assign the correct subtitle in the list
  4. If duplicate files are detected, the system moves the duplicates into the _duplicates folder
Where people often get stuck: If video and subtitle filenames differ too much, auto-matching may fail. The safest approach is still “same name, different extension”.

After task creation, the default output directory is automatically set to the dubbed subfolder of the current folder, so no extra setup is needed.

Step 2: Set translation and audio-track options

Before submitting the task, decide whether this run needs translation and whether you want to keep background music.

1. Choose source and target languages

If the subtitles are still in the source language, choose the correct source and target languages; translation, voice loading, and output naming will all use this language pair.

2. Whether to skip translation

If the uploaded subtitles are already in the target language, for example you already have English subtitles and only want English dubbing, enable “Skip Translation” so the system uses the uploaded subtitles directly as dubbing text.

Workflow change: With “Skip Translation” enabled, task creation goes straight into “Role Assignment”; without it, the task enters terminology review or translation first.

3. Whether to enable terminology review

When the video contains names, brands, or technical terms, enable terminology review. This lets you confirm high-frequency terms before translation actually begins, reducing line-by-line rework later.

4. Whether to enable background music separation

If you want to keep parts of the original accompaniment, drums, or ambient bed beneath the new dub, enable “Background Music Separation”. After enabling it, select at least one track to preserve, such as drums, bass, other accompaniment, guitar, or piano.

How to choose more safely: For explainers, courses, and interviews, it is usually safest to keep only “other accompaniment” or a lighter bed first; if the original piece is highly rhythmic, then consider adding drums or bass. On your first try, keep fewer tracks, confirm speech clarity first, and then add more preserved content gradually.
Troubleshooting tip: If you cannot continue when submitting the task, first check whether at least one preserved track has been selected; if the background music feels too dominant after composition and hurts speech clarity, rerun with separation turned off or preserve fewer track types.

Step 3: Create roles and assign subtitles

Once the task enters “Role Assignment”, do not rush into voice selection. First clarify the speaker structure.

  1. Create roles in the left-side role library, such as “Narrator”, “Host”, or “Guest A”
  2. Assign a color to each role so it is easy to spot in the subtitle list
  3. Click subtitle lines to assign roles. Single-line selection and drag-based continuous labeling are both supported
  4. Frequently used roles can be switched quickly with shortcuts
  5. For clearly unassigned lines that belong to the same speaker, use “Assign All Unlabeled” for bulk processing
Requirement to move on: Every subtitle line in the current task must have a role assigned. Otherwise, later dubbing pages may have missing roles or block progress.

Step 4: Choose a voice for each role and start dubbing

On the “Role Voices” page, the system only shows roles that already have subtitle lines assigned.

  1. Click a role on the left to view the available voices for that role in the target language
  2. Use the search box plus gender and age filters to narrow the list
  3. Preview suitable voices and confirm whether the tone fits the content
  4. Save the configuration after selecting voices for all roles
  5. Click “Start Dubbing”, and the system generates speech in batch by subtitle line
If preview fails: This usually means the current TTS preview request is affected by network limits or service availability. You can still choose the voice and continue, and use the review page results as the final indicator of whether batch dubbing succeeded.

Step 5: Fix problematic lines on the dubbing review page

Once you enter “Dubbing Review”, you can see both file-level progress and line-level status.

  1. Check dubbing progress, completed count, and failure state for each video file
  2. Use single-line preview to check whether a line sounds natural
  3. If a line is too long, too short, or sounds unnatural, edit the translated text directly
  4. Click “Regenerate” to redo only that line instead of rerunning the whole section
  5. If there are failures, use “Batch Retry Failed Items” to continue generating
Pay special attention to: Openings, proper nouns, long sentences, lines with strong emotional shifts, and places where speakers switch frequently.

Step 6: Compose the video and review the final output

When all files finish dubbing, you can start “Compose Video”.

  1. Click “Compose Video” on the review page
  2. Wait for the system to perform audio-video alignment based on subtitle timing, dub duration, and preserved tracks
  3. Open “Video Preview” to inspect the composed files
  4. Use the preview page to toggle subtitles, adjust playback speed, and drag subtitle position for checks
  5. Once everything looks right, open the output folder and collect the final video and subtitle files
What does alignment do? NativeVid tries to keep the new dub aligned with the original shot rhythm. When needed, it synchronizes against subtitle timing and clip duration to reduce obvious audio-picture mismatch.
Review suggestion: After composition, focus on opening lines, long sentences, places with big speed changes, and speaker switches. If those areas feel unnatural, go back to review, adjust the related lines, and regenerate them first.

The default output is usually located in the dubbed folder under the original media directory, with common filenames such as original-video.target-language.mp4 and original-subtitle.target-language.srt.

Want to improve the result further?

If you have already completed your first task, the next best step is to read the FAQ, especially the parts about preview failures, missing roles, retries, and alignment or video composition.

View FAQ