In the past week, I've been researching AI avatar generation. Basically, it works as follows:
- You generate an image or take an image of a real person
- You then define what this person should say and chose a voice.
- According to the platform, it either generates your talking AI Avatar automatically or you might need to generate the voice first and add it to the avatar after that.
I have tested several platforms with the idea in mind to see where I get the most natural results and the most authentic feeling from the avatar. In the video, I show you the end result that were generated through these different platforms.
Links to the websites I've used:
https://hedra.ai/
https://www.vidwud.com/
https://www.vidnoz.com/?a_aid=683466ae057fc (referral link)
https://www.heygen.com/
https://klingai.com/global/
https://www.sievedata.com/ (lip sync)
Why generate AI avatars?
You might ask yourself what the reason would be to generate AI avatars in the first place. They can be used for many reasons. You can create an AI avatar of yourself and then use this same avatar and make him say different things. That can be useful for the creation of videos, shorts, online courses, instruction videos. For people who don't want to be in front of a camera, this opens up the possibility to create videos with talking heads using an artificially made avatar and using a voice of your choice. There are no limits to it and many brands and businesses are already using such avatars to promote their brand or instruct their customers.
With @ph1102, I'm running the @liotes project.
Please consider supporting our Witness nodes: