Recently I stumbled upon a Tweet from Fabian Stelzer where he was exploring what is possible with Midjourney and animation.
Seeing this gave me some ideas that I wanted to experiment with.
I knew I could generate dialogue from an AI using Copy.ai.
I knew I could generate characters using Midjourney.
I also generated a city landscape and an interior of a house for the short intro.
I knew I could come up with simple yet effective picture animations using Capcut ( a great free app to edit videos).
And, I knew I could create fake voices from text with some other apps.
I was not sure how to do a deepfake video, to animate a picture and make it speak but I wanted to give it a try.
I discovered this amazing tool called MyHeritage That is focusing on helping you make animated videos from your ancestors? ( Bit weird but why not ) but as I could upload some Midjourney artwork, it was ok.
It did not work with robot faces, it needs to look like a human to work.
I managed to get the woman dialogue part, (with a french accent, I guess it figured out where I was connecting from).
I used a text-to-voice generator for the robot, after all, it could be that a robot speaks without moving its lips.
I then added some FX to the city pics to give them life a bit using Capcut.
Finally, I made the robot smile on the first 2 slides, giving him a bit of personality.
( Actually, I should say “ her “ as I gave it a female voice, the male voice in another app was terrible and the only good free app I found had only a female voice ^^ but anyway it’s a robot so it’s an “ it “).
I think I can get better results with paid plans for these tools but it’s already quite crazy.
It took me less than an hour this time but the next time will probably be 20–30 minutes.
I think this is a game changer and soon people will use it for visual novels or short films or even other things.
I cannot wait to see what people create mixing all these extraordinary AI tech coming!