Hi-Audio Online Platform: Bridging Music and Research
By J. Gil, G. Richard
The Hi-Audio research team is excited to unveil its new online platform hiaudio.fr tailored for musicians and researchers specializing in audio and Music Information Retrieval (MIR).
A Hub for Innovation:
This web application is a key initiative among several projects housed under the Hi-Audio banner. Supported by an ERC Advanced Grant since 2022, Hi-Audio aims to address the critical shortage of multitrack music data in terms of genre, annotations, and accessibility. Learn more at: https://hi-audio.imt.fr
A Novel Digital Audio Workstation (DAW):
The platform features a groundbreaking online DAW that allows musicians to record tracks asynchronously from any location, ensuring that each track is captured without interference from other instruments, a boon for tasks like audio source separation and generative audio.
Spotlight on Research and Open Science:
First introduced at the ISMIR 2023 congress in Milan during the Late Breaking Demos session, Hi-Audio stands out for its commitment to a public and open-source ecosystem, adhering to the FAIR principles. This commitment fosters a collaborative environment where the research community can share knowledge and leverage the data generated. Notably, the team has developed innovative signal processing methods to mitigate latency in web-based audio recordings, with publications on these advancements expected in the coming months.

Community Engagement and Legal Considerations:
At ISMIR 2024 in San Francisco, the team showcased a new project collecting various renditions of the “Happy Birthday” song, now in the public domain. All materials on the platform are published under a Creative Commons license to address copyright and intellectual property concerns adequately.
Looking Ahead to 2025:
The platform has begun integrating Machine Learning models from the Essentia library by MTG (Music Technology Group), Barcelona, facilitating the first automatic annotation tasks on audio tracks—an essential function in data science.
One of the significant challenges this year is enhancing AI (Artificial Intelligence) integration and promoting community use to foster new data creation. To this end, a practical workshop is scheduled for April at Télécom Paris, where participants will learn to use the application and have the opportunity to record their own tracks.
Join Us:
Interested individuals can register for the workshop here or contact the team for more information at the following link: https://hi-audio.imt.fr/contact/