Once a virtual character has a stable face, the next step is a stable voice. A spokesperson who looks the same in every video but sounds different each time is not a consistent character.
From script to speaking video
- The team writes and approves a script.
- Speech synthesis turns the script into audio, with the voice assigned to the character.
- Lip synchronization and facial animation match the speech to the character’s video.
- The result is reviewed for visual consistency and audio quality before it enters the library.
Localization without a new shoot
The same video can be adapted for several languages, with matching voice and synchronization. For a brand, that means one approved character can deliver the same product message in each market. For a company, it means training or onboarding videos can be produced in several languages with the same presenter.
Typical uses
- Spoken marketing presentations and product explanations
- Multilingual promotional messages
- Corporate and educational videos
- Virtual brand representatives
- Narrated social media content
Whose voice is it?
For original virtual characters, voices come from licensed synthetic voice libraries. For a digital version of a real person, identity verification, consent and a specific authorization to use their voice are required. Cloning someone’s voice without that authorization is not something the platform supports. See our Responsible AI page for the full rules.