BUBLEDUB / Product guides
AI dubbing with a voice for every speaker.
Dub Studio is the web workflow for interviews, lessons and longer videos with more than one speaker.
Dub Studio identifies speakers, translates the dialogue, assigns separate voices and fits the generated speech to the source timing. It accepts video files up to three hours and 5 GB, with one target language per job. These are upload limits, not a guarantee of the processing time or quality of every long recording.
- Voice layout
- Separate voice for each speaker
- Target
- 1 language per job
- Rate
- 1 credit / source second
- Timing
- Dialogue fitted to the video
How it works
Upload a video
Choose a supported video file and select Dub Studio. Audio-only files use Single voice or Transcript instead.
Set the target and review
Choose one target language and the available speaker settings. Check the estimate before starting the job.
Review dialogue in context
Follow the stages in the queue. Play the result, check speaker assignments and listen to difficult or overlapping passages before publishing.
Why separate voices matter
In an interview, a single narrator can make it harder to tell when a question ends and an answer begins. Dub Studio keeps speakers distinct through separate voices. It is useful for a lesson with discussion, a recorded interview or a scene built around dialogue. The outcome still depends on the source recording.
Timing without editing the picture
Generated dialogue is fitted to when the original speech occurs. This helps translated speech follow the action in the video. The process does not alter a person’s face or mouth, and it should not be described as lip-sync. Translations may be adapted to fit the available speaking time.
Example: a ten-minute interview
A 600-second video translated into one language in Dub Studio uses 600 credits. It is a flat rate per source second for this mode, rather than the Fast+ or Pro multiplier from Single voice. The studio shows the estimate before launch. Longer files take more work; no fixed completion time is promised.
Choose between the three workflows
Use Single voice for a short explanation with one narrator. Use Dub Studio when the conversation itself needs separate voices. Use Transcript when the main deliverable is a written record, a summary or subtitles rather than a new audio track. These modes produce different results, even when they begin with the same file.
Prepare the source and review the output
Clear speech and limited overlap make the task easier. Music separation and speaker detection are not perfect: overlapping voices, distant speakers, proper names and background noise can require closer review. You must have permission to use the source material and any voices involved. Do not treat an upload limit as a promise of broadcast-ready output.
Questions and answers
Can I upload audio without a video track?
Dub Studio currently requires video. Use Single voice for translated audio, or Transcript for audio-to-text.
Can I choose several languages in one job?
Dub Studio uses one target language per job. Single voice can handle up to four, subject to its duration and 20 target-minute limit.
Is this voice cloning by default?
The defining feature is separate voices for speakers and timed dialogue. Single voice Fast uses presets; own-voice features in Fast+ and Pro are opt-in and require consent.