AI Avatars From Video of Read people
Here is an early example of an AI avatar that was not made by "text to video" but from a video of an actual person shot on location with a professional film crew.
In this case an actress named Aubrey was hired to be recorded on video in her home. The resultant image and voice were then used to create this video.
Why is it more useful to convert the original video into an AI version rather than just playing the video?
Several reasons:
The original is static. The copy is fixed. But with the AI version, an endless number of new videos can be easily produced with new copy without reshooting.
New videos can be whatever length the copy and use requires. Short for social media, longer for a website,
The AI Avatar version can deliver the copy in any of 37 languages.
See the examples below of Aubrey / Grace speaking in fluent Spanish and Mandarin. The real Aubrey only speaks English.
Imagine your CEO spending a brief hour in his office or boardroom, being shot on video by a small video crew.
Then, have a new video with a new copy to send to social media or post on the website every week or even soon.
This video can be in the correct language spoken by his associates or clients in virtually any country in the world.