As I continue to explore the fascinating world of Artificial Intelligence image and video creation, I’ve been pulling some of my original music productions out of the vault. My style of music production seems to fit neatly with AI imagery. In fact, one piece in my collection is named, Artificial Intelligence. I’m sure I’ll get to that piece at some point.
Yesterday I finished my latest video which came from a piece of oddly developed music I called Spock’s Trombone. I wanted to construct a piece of music made from samples of the trombone, percussion, mouth sounds, and other noises. I also played the trombone in the traditional manner to create other samples, but many of the sounds within the piece came from sampling the hitting of the bell with a mallet or flicking it with my finger, clicking the slide lock and water key, popping the mouthpiece, and various flutter and air sounds mic’d from within the horn.
After collecting these sounds, I assigned them to keys and ran then through a rhythm sequencer for a kind of hip hop beat. To give you a better idea of the process, below is a short TikTok video I made shortly thereafter that shows the making of the music called Spock’s Trombone.
This music seemed like a perfect track for an AI video. As I thought of a theme of a point to the video, it seemed obvious that the music was about rhythm and therefore, so should the visuals. Hence, the name, Rhythm of Life.
I wanted to highlight in the video various elements of rhythm within our world–macro examples such as stars orbiting within galaxies, micro examples of atoms and sound waves, and everything in between.
Most of these rhythms go on around us without giving them a second thought. Even the idea of aging, which I portray in the video, is a sort of rhythm of life. Sound is rhythm and so is light. The odd angle of our planet’s rotation provides us with seasons, mating rituals, and vast animal migrations.
But the continuity of the video is dance. I generated many images of dance and did so using the style of a silhouette.
As with my other videos, the images were created in both Midjourney and Dali, the image creation element of Chat GPT-4. Both platforms have their personalities. It’s difficult to make broad distinctions between them since they are both capable of terrific imagery and surprises.
The advantage of Dali is that you can have a conversation with Chat GPT as you tell it how to modify and edit the images until it produces something you envisioned. With Midjourney, the image you get is it. You can go back and re-write the prompt or create a variation of a previous image, but you can’t have a conversation (yet) to mold the image to your taste.
The other AI tool in my arsenal is called Gen-3 by Runway. Gen-3 is a recent image-to-video and text-to-video creator that produces four second spurts of video based on a text prompt or an image or both. AI video creation is a fascinating new technology that is still in its infancy, but has recently come a long way in a short time.
Part of the charm of Gen-3 and the other AI video creation platforms is its less than perfect output. As these tools progress, this curious feature will be lost (unless they add a legacy output aspect to it) but at their current state, the results are full of small and odd imperfections. For example, as the video of a woman progresses, she might grow more fingers or another arm. Sometimes these abnormalities are super subtle and then sometimes it makes the final product unusable, forcing you to try again and wait to see the new result.
I am fascinated by this technology and will be creating much more from it using my original music. I recently wrote a post arguing against the AI doomsayer’s here. But without further ado, here is Rhythm of Life: