AI dubbing

Redub

Every actor keeps their own voice.

Redub takes a scene into another language. It finds who speaks each line, samples a voice for every actor from their own lines, adapts the script to fit the time on screen, and mixes the new dialogue under the original music and effects.

Redub screens

Redub scene player on a Tears of Steel scene with the Urdu track selected and a quality panel showing loudness, sync, overruns and voice match. Demo data.

Switch language while the scene runs, with quality figures for every track.

Built for
Film, Drama and Broadcast Media
Status
demoDemo
Runs on
WebUses AI: Yes
Built with
Next.js, React, TypeScript, Python, ffmpeg

Video

Redub: one scene, two new languages

Video coming soon

Overview

Operational context

A dubbed scene usually means a script adaptation, a casting session, a recording booth and a mix, one language at a time. Redub runs the same stages on one clip: spotting, adaptation, casting, recording, fit, mix and checks, and shows what it measured before anyone listens.

What it covers

Capabilities

  1. A cloned voice for every actor, sampled from that actor's own lines.

  2. Script adaptation that fits each line to the time on screen.

  3. The dub mixed under the original music and effects.

  4. One picture with an Original and Dub switch, subtitles and an MP4.

10 features

Features

How it works

How the workflow connects

Send a clip, choose the languages, review each track on the original picture, and take the MP4 and subtitles.

  1. Spot

    the lines, the speakers and their timing are marked from the original.

  2. Cast

    a voice is sampled for every actor from their own lines.

  3. Adapt and record

    each line is rewritten to fit and spoken in that actor's voice.

  4. Mix and check

    the dub sits on the original music and effects and is measured track by track.

Status

Implementation and status

demoDemo

A working demo of our own dubbing studio, shown on request. Pilots run on our studio pipeline, one clip at a time, and no track is released until it passes the release gate and a listen. Dubbed voices are AI-generated and labelled as such.

The discussion starts with the intended users, existing systems and delivery requirements. Confirm which functions can be demonstrated, what would need adaptation and how access, hosting and support would be arranged for the proposed use.

Built with

  • Next.js
  • React
  • TypeScript
  • Python
  • ffmpeg

Runs on

Web

Questions buyers ask

What is Redub?

Redub is an AI dubbing demo that takes a scene into another language. It finds who speaks each line and samples a voice for every actor from their own lines. It then mixes the new dialogue under the original music and effects.

Who is Redub for?

Redub is for dubbing houses, production companies and channels that want to test a scene in a new language before a full dub. It is also for teams that localise documentary narration.

How does AI dubbing work in Redub?

Redub spots the lines, speakers and timing, then samples a voice for every actor from their own lines. Each line is rewritten to fit the time on screen and spoken in that actor's voice. The dub is then mixed and measured track by track.

Is Redub ready to use?

Redub is a demo: a working demo of our own dubbing studio, shown on request. Pilots run one clip at a time, and no track is released until it passes the release gate and a listen. Dubbed voices are AI-generated and labelled as such.

Build something like this

Have a scene to take into another language? Send us a short clip.

Request a Redub demo

See all products