While much of the AI-powered tech for music creation spits out replicas or facsimiles, DataMind Audio does something radically different. Founded by musicians for musicians, the UK-based startup is launching a first-of-its-kind virtual instrument, the Combobulator, a real-time generative AI instrument powered by custom models called Artist Brains, created in collaboration with a wide range of donor musicians who receive a 50% of the profits.
Backed by the Parisian giant IRCAM, Innovate UK, and Edinburgh University because of its financial and creative benefit to musicians, this AI-crafted instrument offers access to an artist’s sonic mind and musical style, while providing an ethical and sustainable way for artists and producers to experiment with and benefit from the massively augmented capacities of AI.
The latest artist to partner with DataMind Audio and contribute music to generate an Artist Brain is Mr. Bill–the Sydney-based DJ and producer known for packing huge venues and setting new standards in sound design.
“This is the coolest opportunity to share my musical style with the world for other artists to experiment with and play without sacrificing my identity,” says Mr. Bill. “I’m really excited to see what people create with these brand new sounds and textures inspired by my music. This is a genuinely creative ethical AI tool we’ve all been waiting to use.”
DataMind Audio has cleared another hurdle for generative tech, one that is important to many musicians and composers: the massive computing power required to run AI models and this power’s environmental impact.
Committed to being a green company, DataMind Audio has developed a solution that allows a CPU to process rapid computations like a GPU, the graphical processing units that typically run large AI models. This means music-producers can use The Combobulator for real time audio synthesis in their DAW.
‘We don’t always know where our ideas and creative drive come from. Similarly, when we map an artist’s artificial Brain, there is a mysterious realm called latent space inside it,” reflects music producer and technologist Ben Cantil, DataMind Audio co-founder.
“The Combobulator, allows sound designers and music producers to explore this space in 64 dimensions, discovering newly generated, never-before-heard sounds. They can then manipulate them by means of an easy-to-use interface, to make entirely new, original music.
The concepts for Artist Brains and the Combobulator originated in 2022 at DEFCON, the annual hacker convention. Entrepreneur Catherine Stewart, and celebrated electronic artist- programmers Encanti (Ben Cantil) and Rob Clouth collided with CJ Carr and Zack Zukowski from Dadabots, the duo who pioneered generative metal music via neural networks way back in 2012.
Together, they realized they had the combined skill sets to create an AI-powered software ecosystem of new instruments to discover entire universes of sounds: A new music, for a new era.
To support the development of the Artist Brains and Combobulator VST plugin, DataMind Audio saw the need for a new kind of music engineer for a new era. They coined the phrase “Model Reliability Engineers.” These genius hybrid musician-coders have the delicate and difficult task of tuning their fellow musicians’ brains: Creating a Brain requires a marriage of art and science.
“My job is collaborating with world-class artists to make a completely new product that makes never-before heard sounds and music in the AI universe,” says MRE Yashique Chalil. “I’m well paid and so are they, and we don’t hurt the planet. That’s a dream job.”     Â
“What is fascinating is that there is so much more to learn about the capabilities of neural networks,” says Dr. Martin Parker, Head of The Reid School of Music at The University of Edinburgh.
‘We are using our research to invent tools that serve artists first, and which build a creative community that is empowered financially and creatively by our application of AI. Datamind Audio is one of very few innovators out there who genuinely nurture artists and promote creativity using AI.”
The DataMind supercomputer is on an electrical grid that uses solar power as a primary energy source which, combined with a custom liquid-cooling system, results in a lower-energy footprint. All other cloud GPU services that the company uses for machine learning applications produce zero emissions.
“There is a lot of valid concern about the future of creatives in the AI era, but in the area of sound design and music production, we’ve created a solution by artists, for artists: They are paid for their original work, get to use new tools and instruments to create original music with new products,” says Catherine Stewart, Co-Founder and Managing Director at DataMind Audio.
“Word is spreading in the music community – and artists are approaching us to train and sell their Brains for them. We launched with a group of globally renowned artists who know and trust us, but now that the word is out, we’re expanding our sound palette and Artist Brain marketplaces with some seriously big names. I’m proud we’ve come up with a strong business model for Artist-AI thrivability – where everyone genuinely benefits.”