Triangle Afrobeat Orchestra - Buena 4Tuna: Chroma Awards

A fast, music-driven visual experiment built with next-gen AI filmmaking tools based on the poem by Dr. Joan Santos and performed by the Triangle Afrobeat Orchestra.

Comment

Inspiration

We wanted to explore how far we could push AI-driven cinematics inside a music-centered project. Afrobeat has a bold, rhythmic visual identity, and we wanted to see if we could recreate that energy using a fully AI-assisted production pipeline while recreating a cinematic experience for Dr. Joan Santos poem.

What it does

This demo pairs a high-energy Afrobeat track with AI-generated cinematics, motion, and stylistic color work. It functions as both a creative experiment and a proof-of-concept, showing how modern AI tools can rapidly produce polished music visuals that feel handcrafted.

How we built it

We generated the core visuals using Hailuo for video, Imagen/Seedream for character and performance consistency, and custom prompt engineering to control lighting, movement, and camera behavior. After refining and iterating on each shot, we assembled everything in Adobe Premiere, smoothing transitions, syncing motion to beat changes, and stabilizing the visual flow.

Challenges we ran into

The biggest challenge was maintaining consistency across scenes while working with multiple AI models. Keeping color, lighting, and character design aligned across generations required careful prompt tuning and multiple iterations. I also had to work around occasional green-screen artifacts and motion inconsistencies that Seedream introduced.

Accomplishments that we're proud of

I’m proud of how cohesive the final piece feels, especially considering it was built entirely through AI tools. The color flow, rhythm matching, and overall cinematic tone demonstrate how quickly high-quality visuals can be produced with the right workflow. The major technical feat was getting Dr. Joan Santos to lip synch underwater to the lyrics!

What we learned

We learned how valuable tight prompt control is when mixing models. Small adjustments in tone, camera language, or lighting phrasing have a huge impact on consistency. We also gained insight into which models handle motion best versus which excel at texture, expression, or stylization.

What's next for Triangle Afrobeat Orchestra

We have two more music videos in the pipeline.

Built With

  • omnihuman
  • seedream
Share this project:

Updates