Each video in this series presents a song from 'The Dark Side of the Moon,' blending traditional and modern techniques. The approach combines 3D animation and compositing with AI-generated imagery. All visuals are ultimately processed with AI to resemble hand-drawn, 12 fps frame-by-frame animation, echoing the psychedelic styles characteristic of the 1960s and 70s. This in order to capture the album's essence in a digital format, reflecting both the album's legacy and contemporary technology.
In "Brain Damage", I first generated a long list of prompts text-to-image generative AI using the lyrics, AI's interpretations of the lyrics (ChatGPT4), and my own custom prompts as needed. These prompts were then turned into images using genrative AI. I created a custom workflow for Stable Diffusion / Comfy UI to do this, which required coding custom nodes to parse the list of text prompts and combine them with user input to create the final prompts that waere fed to the video generation AI. The psychedelic style of the imagery allowed me to embrace the inherent weirdness that emerges from generative AI when pushed to the edge.
The TDSOTM prism is 3D animation made in LightWave 3D. I modeled and hand-animated the prism and the moon (elevation data from NASA), processed through AI to resemble hand-drawn animation.
The result was hundreds of short images that were based on the lyrics and also the overall arc of the album. These were then animated using AI, again with a custom image-to-video workflow in ComfyUI. I then manually selected and edited the resulting animations onto the final cut in Vegas Pro, which was once more processed through AI and compositing to finalize the look.