Participants 2025
Artist in the Loop
AI Song Contest 2025 / participantsTEAM / Artist in the Loop
SONG / UU_SMT forAIver
About the TEAM
We are a team of university students, artist Florian, and professor Anja, working together to explore how AI can inspire new approaches to sound and music. Our collaboration brings together diverse perspectives: Florian’s artistic vision and experience, Anja’s academic guidance, and our curiosity and technical skills as students. Together, we design and experiment with AI systems that generate and transform sounds, approaching them as creative tools rather than replacements for human expression. The project combines research and artistic exploration, providing a space where technology and art intersect to spark new ideas. Through experimentation, learning, and collaboration, our team aims to expand the boundaries of sound design and uncover fresh possibilities for artistic expression.
About the SONG
"UU_SMT forAIver" by Ryan Fontane is an experimental track built entirely from sounds generated by artificial intelligence. Created in collaboration with master's students from the University of Utrecht, the project uses custom AI audio samples as its foundation. Rather than focusing on a traditional melody, the song layers interesting textures and rhythmic patterns. A distinctive sansula sample opens the piece, which is driven by a foundational beat made from AI sounds like "hyperbass." The bassline itself is a reinforced kick drum. The harmony was shaped by the tones already present in the AI samples, resulting in sections using D-Phrygian and D-Dorian modes. The result is a unique composition exploring the creative potential of human-AI musical partnership.
About the HUMAN-AI PROCESS
The song was created using a unique, AI-driven process. The journey began by feeding the artist's personal sound library into various AI models, which then generated a collection of completely new audio samples.
The foundation of the entire song was a drum beat built from two specific AI-generated samples: 2_full_hyperbass_new1 and output_beep_gsAna_intro_1. Because the original kick drum sound wasn't powerful enough, the artist reinforced it with a deeper, more resonant sine wave.
The melody wasn't composed in a traditional sense. Instead, the artist used melodic fragments and textures already present in the AI samples, particularly a prominent "sansula" sound from the output_sansula_gsRcd_full sample. They adjusted these sounds to fit the song's key. Through trial and error, they arranged the samples to create a harmonic landscape that shifts between different musical modes.
Finally, the artist mixed and mastered the track themselves, blending all the AI-generated samples in a digital audio workstation (DAW) to achieve the final sound. The entire creation was an ephemeral, collaborative experience, with the artist and their team working remotely to bring the AI-generated sounds to life.