YouTube creator DoodleChaos made a full-length music video utilizing AI. The video includes various digital camera angles, types, and themes; all made utilizing AI. DoodleChaos needed to information the AI instrument to provide one thing that made sense and aligned along with his conceptualization of the track.
DoodleChaos could be very happy with the output and appears like a graphic designer who spent days engaged on it. However, it was all performed by one particular person and an AI instrument simply utilizing the track’s lyrics.
How did DoodleChaos create a music video utilizing AI?
DALL-E 2 is among the many strongest picture technology instruments on the planet. MKBHD lately made a video that demonstrated the instrument’s energy to provide photographs based mostly on pure language enter. In less complicated phrases, you should enter textual content and describe the picture you need it to provide.
DoodleChaos doesn’t have entry to DALL-E 2 and wished to strive it out. But it isn’t obtainable to most of the people but, so he resorted to different obtainable instruments.
DoodleChaos first stumbled upon VQGAN + CLIP, which may generate AI movies however finally settled for Disco Diffusion V5.2 Turbo. The end result is a beautiful music video he made by commanding AI utilizing pure language enter.
Was it straightforward?
Disco Diffusion V5.2 Turbo, like DALL-E 2, can produce photographs based mostly in your wildest imaginations. But you can not anticipate it to create a video by merely inputting the lyrics and letting it give you one thing. DoodleChaos inputted the lyrics and even added handbook inputs to assist it perceive the necessities higher.
DoodleChaos defined in his video description, “I added keyframes for digital camera movement all through the generated world. These keyframes have been manually synchronized to the beat by me. I additionally specified adjustments to the artwork fashion at completely different moments of the track. Since lots of the lyrics are fairly non-specific, even a human illustrator would have a tough time making visible representations. To make the lyrics extra digestible by the AI, I typically modified the phrase to be extra coherent, equivalent to specifying a setting or environment.“
It was a collaborative effort between DoodleChaos and the AI instrument which led to the manufacturing of a full-fledged music video. However, it nonetheless depends on crisp language enter to provide the photographs and clips you need.
Graphic designers should not fret about their jobs as AI instruments might have many vital pre-visualization functions. These instruments will nonetheless require somebody to information them via the entire content material creation course of.