Winners of the Generative Music Prize

The Generative Music Prize is an international competition that highlights the creative potential of music generated with algorithms. This first edition is organized by IRCAM to mark the 30th anniversary of the IRCAM Forum. This page presents the award winners.

First Prize

The first prize of 2500 euros is awarded to Simon Colton for his achievements in the field of generative music, by an international jury composed of Agoria, Greg Beller, Berrak Nil Boya, Andrea Faroppa, Georg Hajdu, Holly Herndon, Philippe Langlois, Frank Madlener, Miller Puckette and Anna Xambó, at the first edition of the generative music prize, organized by the IRCAM forum.

The Parley Classical Music Composition System

Parley is an AI system that we hope will be taken seriously as a classical music composer in its own right, one day. This involves solving numerous technical problems, but also raises societal issues around creative personhood, which we have begun to explore in a philosophical manner. The technical approach is neuro-symbolic, where we have combined the power of deep learning with the control afforded by rule-based systems. In particular, Parley can listen to the music it produces using neural listening models which tell it how each bar sounds in terms of genre, emotion, mood and situation. It can use the information about the sound of the music to both communicate particular elements as talking points and to improve the music to sound better in a particular respect. Parley celebrates and communicates the musical composition process, by producing various drafts of a composition as it is being made, and describing generative and editing processes along the way. It produces a video performance of each piece with talking points in margin notes, to stimulate discussion and user involvement in the music. Parley has been ported to a Colab notebook available at tinyurl.com/parleyV2 and is available for anyone to use.

Simon Colton

I am an AI researcher with 27 years of experience in generative artificial intelligence. I’m most well known for the AI systems that I’ve built, which have had applications to pure mathematics (the HR series of programs); the visual arts (The Painting Fool software); creative ideation (The WhatIf Machine); video game design (the Wevva iOS app) and classical music (Parley). I’ve co-authored around 250 papers on generative AI, covering practical as well as philosophical elements of computational creativity. In particular, I’ve researched what it would mean for an AI system itself to be considered creative, and I’ve introduced philosophical notions such as the Machine Condition and Creative Personhood in this context. I’ve also published on theoretical issues around the assessment of progress towards creativity in AI systems. I’ve held grants totalling more than £20m, including an EPSRC leadership fellowship and an ERA Chair in Digital Games Technologies. I was a lecturer, senior lecturer and reader at Imperial College, before taking up a professorship at Goldsmiths College. I’m now a professor of computational creativity at Queen Mary University of London. I’ve led large and small research groups at Imperial, Goldsmiths and Falmouth University, and engaged in numerous high-level public engagement event, with more than 150 articles/interviews/documentaries covering my work. I was involved in staging the “World’s First Computer Generated Musical Theatre Show” called Beyond the Fence, which ran in London’s West End in 2016 and was covered by a Sky Arts documentary. Music was my first love, and I took A-level music as a teenager, but was ultimately seduced by pure mathematics and computer science at university level. I spent a sabbatical year in the Sony Computer Science laboratory in Paris, working with François Pachet and colleagues, and I ended up living in Paris for five years. I’ve devoted my research to generative deep learning for the past five years, and have spent the last two years building a neuro-symbolic system called Parley which generates classical music compositions. Our aim with the system is for it to one day be taken seriously as a classical music composer in its own right, and for it to contribute to musical culture, benefitting both novice and established composers.

Second Prize

The second prize of 1000 euros is awarded to Martin Heinze for his achievements in the field of generative music, by an international jury composed of Agoria, Greg Beller, Berrak Nil Boya, Andrea Faroppa, Georg Hajdu, Holly Herndon, Philippe Langlois, Frank Madlener, Miller Puckette and Anna Xambó, at the first edition of the generative music prize, organized by the IRCAM forum.

Fibonacci Jungle. A generative framework for Jungle and Drum & Bass based on the Fibonacci number sequence.

While singular generative techniques have already become an established part of the creative process in music writing, holistic approaches to generative music production in traditional electronic dance music genres yet seem under-represented both in theory and practice. Fibonacci Jungle provides a simple to use generative framework for Jungle and Drum & Bass built on the Fibonacci number sequence as structural alternative to conventional meters and track build-up. This framework is implemented in Pure Data; it uses probability and randomization within a pre defined set of genre typical parameter settings (tempo, harmonics, sample selection). Fibonacci Jungle allows creating stand alone tracks in a Jungle and Drum & Bass aesthetics with only a few clicks and can be individually customized.

Martin Heinze

Martin Heinze is a sound artist, composer and electronic musician working in the field of experimental electronic dance music and generative art. He holds a Master’s Degree in Media Arts, Art Theory and Media Sciences from the University of Arts and Design (HfG) Karlsruhe, Germany. Heinze writes and performs under artist names Martsman, Anthone, Bokeh (in collaboration with Katsunori Sawa) and //no (in partnership with Felix Krone). His works have been published since 2005 on numerous electronic music labels, such as Hidden Hawaii, Offshore, Hospital and his self-established The Weevil Neighbourhood. From 2005 to 2007, he engaged in collaborative projects with Centre for Media Arts and Design (ZKM) and Karlsruhe Institute for Technology (KIT) with a primary emphasis on the artistic utilization of location-based services. In 2010, Heinze graduated with holistic record label concept The Weevil Series which was supervised by Anna Jermolaewa and Paul Modler. Since then, the concept has been realized as The Weevil Neighbourhood, a creative outlet for music and media art showcasing a diverse range of artists from Germany, Japan and the United States. In 2018, he started collaborating with artist duo IRIS-A-MAZ (Iris Holstein, Matthias Hederer) leading to joint venture media art installation Animalis Prismatis, which has been on display at multiple venues across Germany since 2020 and was awarded at SwingBy young art award 2023 hosted by Poolhaus Blankenese, Hamburg. From 2019 on, Heinze has been conceptualizing various approaches to generative electronic dance music. These efforts began to materialize in late 2022 with empty x on Solana blockchain based Nina protocol. Subsequently, he delved into algorithmic Jungle music production leading to releases Fibonacci Jungle Versions and Edouard & Leonardo in 2023. Heinze's current focus revolves around experimenting with generative AI in the realm of audio, employing models like RAVE, MSPrior, and vschaos2.

Third Prize

The third prize of 500 euros is awarded to Axel Chemla--Romeu-Santos for his achievements in the field of generative music, by an international jury composed of Agoria, Greg Beller, Berrak Nil Boya, Andrea Faroppa, Georg Hajdu, Holly Herndon, Philippe Langlois, Frank Madlener, Miller Puckette and Anna Xambó, at the first edition of the generative music prize, organized by the IRCAM forum.

genesis v1 - patch version

Genesis v1 - generative version is a first example of no-input generation with neural networks. Instead of relying of data extracted from human practices, genesis rather intends to amplify and overdrive the inner materiality of several neural networks, providing a strange encounter with a self-listening system of four generative agents that have been designed without any external data.

Axel Chemla--Romeu-Santos

Based in Paris, France, Axel Chemla—Romeu-Santos works a researcher, composer, and performer in various fields such as music, theater, and artificial intelligence. After a double undergraduate degree in Engineering Sciences & Music Theory, he specialized in acoustics and computer music at IRCAM. Always looking for creativity through technology, he initiated a PhD between IRCAM (Paris) and LIM (Milano) on the creative uses of generative artificial intelligence for sound synthesis. After graduation, he continued a research & creation approach to artificial intelligence, working both on scientific papers on AI creativity, and experimental musical pieces exploring diverse aspects of these technologies (such as network bending, real-time improvisation, and composition). Besides institutional works, he also work as musician and composer for the company Théâtre de la Suspension, is co-founder of the w.lfg.ng collective, member of the maximalist electronic music band Daim™, and has his personal project Kenoma.