Video Dubbing FAQ

No. Video files, subtitles, intermediate audio, and final exports are processed and stored locally. Only the translation or TTS requests you actively use go online, and original media assets are not uploaded to NativeVid servers.

Not always. If the uploaded subtitles are already in the target language, you can enable “Skip Translation” during task creation. The system will use the uploaded subtitles directly as dubbing text and enter “Role Assignment” right away.

Put the same batch of videos and subtitles in one folder and keep filenames aligned whenever possible, for example lesson01.mp4 with lesson01.srt. This gives the most stable auto-matching and makes batch work easier.

The current workflow targets common video formats such as .mp4, .mkv, .avi, .mov, and .webm, together with common subtitle formats such as .srt, .ass, and .vtt. Cleaner filenames also improve auto-matching success.

The “Role Voices” page only shows roles that already have subtitle lines assigned. If you created a role but assigned no subtitle lines to it yet, it will not appear in the voice setup list. Go back to “Role Assignment” and complete the assignment first.

Not necessarily. Preview failure usually means the current TTS preview request is affected by network conditions, service availability, or restrictions. You can still save that voice and continue. The final result should be judged by the “Dubbing Review” page.

Find the line on the “Dubbing Review” page, edit the translated text first, and then click “Regenerate”. This only redoes that one line, so you do not need to rerun the whole batch. It works especially well for long lines, proper nouns, and lines with strong emotional changes.

Yes. The review page provides “Batch Retry Failed Items”, which continues generation for failed subtitle lines only. You can also click “Regenerate” for a single problematic line.

There are three common reasons: first, some subtitle lines still have no assigned role; second, some roles still have no voice selected; third, some files in the current task have not finished dubbing or still contain failed items. Once those are resolved, composition and alignment can usually continue.

Audio-video alignment means that during final composition, the system combines subtitle timing, generated voice duration, and clip rhythm to reduce mismatch and make transitions feel more natural. NativeVid handles this during composition, so it is not a separate preliminary page but a key feature after dubbing finishes.

It is optional. When enabled, the system tries to preserve parts of the original accompaniment or background bed and layers them back with the new dub during final composition. If you want a cleaner final output, turn it off. If you enable it, select at least one preserved track. For courses, explainers, and interviews, it is usually best to keep only a small amount of background bed so speech stays clear.

There are three common cases: first, task submission fails because no preserved track was selected after enabling the feature; second, background music feels too dominant after composition, in which case keep fewer tracks, preserve only “other accompaniment”, or rerun with the feature off; third, the voice and ambience do not blend well, so verify the dub itself first and then compare versions with and without separation.

When the system detects duplicate content, it moves the duplicates into the _duplicates folder under the current directory and shows which file was kept and which one was moved. This reduces mismatches and duplicate generation in batch jobs.

The default output directory is the dubbed subfolder under your selected media folder. After composition, you will usually see the final video file and aligned subtitle file there. Both the preview page and the review page let you open that folder directly. When judging alignment quality, review the exported final video there rather than listening only to single-line audio.