Imagine you have millions of classical music scores and recordings scattered across dozens of websites, but no easy way to connect a Beethoven score to its best recordings, or to see what expert musicians think about a particular passage. TROMPA built smart tools that automatically link all this music data together — scores to recordings, recordings to annotations — and then let music lovers and professionals add their own knowledge on top. Think of it like Wikipedia meets Spotify, but specifically for public-domain classical music, where the crowd helps the AI get smarter over time.
Music platforms, archives, and education companies sit on vast amounts of public-domain classical music — scores, recordings, annotations — spread across disconnected repositories with inconsistent quality. Linking a score to its best recordings, or surfacing expert insights about a piece, still requires expensive manual curation. This fragmentation means users cannot get comprehensive access to Europe's musical heritage, and businesses cannot monetise or enrich this content efficiently.
TROMPA built 5 working prototypes tailored to different audiences: music scholars, orchestras, instrument players, singers, and music enthusiasts. These prototypes use deep learning to automatically analyse, link, and enrich music data across repositories and formats (scores to recordings), combined with crowd-sourced annotation workflows where users contribute expertise back to the platform. In total, the project produced 24 deliverables across its 3-year run.
If you are a digital music platform struggling to connect fragmented music content across formats — this project developed AI tools and crowd-sourced workflows that automatically link scores to recordings and enrich metadata. With 5 working prototypes tested across different user types, these tools could help you offer richer browsing experiences and keep users engaged longer.
If you are a music education company that needs structured, well-annotated practice material for students — this project built working prototypes for instrument players and singers that connect public-domain scores with recordings and expert annotations. The platform was designed to serve millions of users, giving learners context-rich practice tools built on top of freely available classical repertoire.
If you are managing digital music collections and spending too much time manually cataloguing and cross-referencing assets — TROMPA developed deep-learning-based tools that automatically analyse and link music data across repositories. With 9 consortium partners across 5 countries contributing to 24 deliverables, the technology was built to handle large-scale public-domain archives with varying data quality.
TROMPA followed an open innovation philosophy — all knowledge derived was released back to the community in reusable ways. This means the core tools and data enrichments are likely available as open-source or open-access resources. Licensing costs for commercial integration would need to be discussed directly with the consortium, led by Universidad Pompeu Fabra.
The project was designed to handle massive public-domain archives and link data across multiple repositories at scale. With deep learning at its core and crowd-sourced quality feedback loops, the architecture was built for volume. However, commercial deployment beyond public-domain content would require additional rights management.
The project was funded as a Research and Innovation Action (RIA) with EUR 3,054,930 in EU funding across 9 partners. IP ownership typically follows EU grant rules where each partner owns what they developed. The open innovation approach suggests much of the output is openly available, but specific commercial licensing terms should be clarified with the coordinator.
TROMPA delivered 5 distinct working prototypes: for scholars, orchestras, instrument players, singers, and music enthusiasts. These were demonstrated at months 24 and 34 of the project. The technology has been tested with real users but would likely need further engineering for full commercial deployment.
The project was built on open standards and designed to work across existing community-contributed repositories. Based on available project data, the tools link data across different modalities (scores to recordings) and repositories, suggesting good interoperability potential. Integration specifics would depend on your platform's architecture.
The project focuses specifically on public-domain classical repertoire, which minimises copyright risk. However, individual recordings or editions may carry separate rights. The consortium included 3 industry partners who likely addressed commercial use scenarios during the project.
The TROMPA consortium brings together 9 partners from 5 countries (Austria, Germany, Spain, Netherlands, UK), with a healthy mix of 4 universities and 3 industry players including 3 SMEs — giving it a 33% industry ratio. The project was coordinated by Universidad Pompeu Fabra in Spain, a leading music technology research university. The presence of 3 SMEs signals that commercial viability was considered from the start, though the 2 "other" partners (likely cultural institutions or foundations) reflect the project's public-interest mission. For a business looking to adopt this technology, the mix of academic depth and industry participation means the tools were developed with both rigour and practical usability in mind.
Universidad Pompeu Fabra (Barcelona, Spain) — look for the Music Technology Group, which leads music AI research there.
Want to explore how TROMPA's music enrichment tools could fit your platform? SciTransfer can connect you directly with the research team and help evaluate commercial fit.
