Familiar: a video and live audio digitizing tool, a live audio synchronizing A multilingual translation to generate sound and mouth

Familiar It is a video and live sound scoring tool that produces sound and oral at the same time, and preserves the original sound and the scene. Supports the translation of 25 languages in real time. High-precision translation and face reduction, giving the global audience a natural mother tongue experience。

Familiar function

1. Sound is synchronized with mouth: the same model generates sound and mouth at the same time, avoids dissynchronization of sound drawings caused by separation tools, and preserves the sound quality of the original speaker. The entire face is translated to perform in the new language, rather than to act on the lips。
2. Live-to-live voice-to-peer: provide real-time voice-to-peer to support the roll-out of video from each language to an independent channel, with a delay of only a few seconds to allow the global audience to view the mother tongue version immediately。
3. The audio of the scene is preserved: the sound of the scene like laughter, music, sound of the environment is preserved in its entirety to avoid background sound errors. In the baseline test, background sound error is 270% less than Eleven Labs。
4. Quality translation: 54.7% is lower than the most recent version of Eleven Labs, with sound similar to original speaker in all 11 test languages。
5. Multilingual translation: Supports an arbitrary exchange of 25 languages, covering major market languages such as English, Spanish, Chinese and Hindi, to meet global distribution needs。
6. Video review editor: Provides pre-rendering review interface for video, which can be edited and generated by the user after any subtitle, to ensure that the content is accurate。
7. Personalized non-translation list: allows users to specify non-translation phrases, verbs or person names, and rewrites jokes in the target language to fit the local culture。

Official website link:https://thefamiliarlab.com