Makeout Magic Feels Right

Makeout Magic Feels Right

"Make Out Magic Feels Right" is an AI-Generated Music video. Every component of audio and video was generated and brought to life through AI. The Entire project is explained in this video!

Original Images created on Krea.ai
Music generated at Suno.com
Images animated and Lip-sync'd at Runway.ml

The song's lyrics were original, and not generated by AI because this was a gift with a very specific message. But AI technology allows us to upload our own lyrics, define a musical style, and generate multiple versions of our song.

After generating a dozen or so versions, I decided on the version I liked the best, and upon which the video will be synchronized.

This video will be created by using image to video technology, which brings movement that we direct via prompting to still images.
Because this was designed to be a music video, I needed to generate multiple images of me, on stage, as a country singer.

The required that I create a model of my face, and Tracey's face, so that I could easily generate multiple images with our likeness.

I generated several dozen images based on my vision for how the video would look.

Because I wanted to have total control of the facial animation, I chose RunwayML's "Act-1" techology which allows you to define the facial animation on a video based on a driving video that you create.

The facial movement on the driving video is transferred to the target video providing a great deal of control.

1. First I recorded myself lip-syncing the entire song just how I wanted it to appear on the target video.
2. Because there is a 20-second limit to the length of the video you can lip-sync at one time, I cut my recording into 15-20 second clips within a video editor, an rendered each clip separately.
3. The resulting videos are the exact same length as the input videos, so they will be easy to line up in the final editing step.

Finally, it's a relatively simple process to drop the videos that have had lip-sync applied, back onto the timeline in your editor for a rough cut.

Then, there is a fine-tuning process where I may take bits and pieces from the clips or different takes of the clip to get the best results.

Muting all but the original audio track, also on the timeline, I export the final video.