DoReMIR · Project

Cloud Software That Converts Live Music Audio Into Written Sheet Music Automatically

digitalPilotedTRL 7

Imagine humming a melody or playing a chord on piano and having a computer instantly write out the sheet music for you — like Google Translate, but for music. That's what DoReMIR built. Their ScoreCloud software already handled single notes, but this project tackled the much harder problem of multiple instruments playing at once. They combined music theory, audio engineering, and machine learning to teach a computer to "listen" to complex music and figure out every note being played simultaneously.

By the numbers

Total project deliverables produced

Demo deliverables including final application and AMT engine

Commercial SME partner with existing products

100%

Industry participation ratio in consortium

API repository releases for integration

The business problem

What needed solving

Converting complex, multi-instrument audio recordings into accurate written sheet music is still a manual, time-consuming process. Music teachers, publishers, and producers spend hours transcribing what a computer should be able to hear and notate automatically. While single-note transcription existed, handling polyphonic audio — where multiple notes and instruments overlap — remained an unsolved commercial problem.

The solution

What was built

The project delivered a cloud-based polyphonic audio transcription engine (final AMT engine iteration), a consumer-facing application (final application iteration), and two API repository releases for third-party integration — totaling 13 deliverables. The solution extends the existing ScoreCloud product suite from monophonic to polyphonic transcription.

Audience

Who needs this

Music education platforms and digital learning tool providersMusic publishers needing fast notation from audio recordingsMedia production studios requiring sheet music from audio tracksWorship and church music teams transcribing live performancesIndependent musicians and songwriters who compose by playing rather than writing

Business applications

Who can put this to work

Music Education Technology

SME

Target: EdTech companies and music schools offering digital learning platforms

If you are a music education platform struggling to make learning interactive and accessible — this project developed a cloud-based engine that automatically transcribes polyphonic audio into sheet music in real time. Students can play or sing, and the system writes the notation for them. This was built with end-user participation in music teaching contexts and delivered as a final application iteration with API access.

Music Publishing & Licensing

mid-size

Target: Music publishers and rights management companies needing fast notation from recordings

If you are a music publisher spending hours manually transcribing recordings into notation for licensing and distribution — this project delivered an automatic music transcription engine with 13 deliverables including API repositories. The cloud-based solution processes polyphonic audio and outputs musically meaningful, high-level notation, cutting transcription time dramatically.

Media & Entertainment Production

any

Target: Production studios and content creators needing quick music notation from audio tracks

If you are a production company that needs sheet music from audio recordings for arranging, scoring, or archiving — this project built a polyphonic transcription engine that uses machine learning and musicology to analyze complex audio. The solution was delivered as cloud-based APIs, meaning it can integrate directly into existing production workflows without hardware investment.

Frequently asked

Quick answers

What would this technology cost to integrate?

The project delivered cloud-based APIs (two API repository releases), which typically means a subscription or per-use pricing model. As an SME Instrument Phase 2 project by a single commercial company (DoReMIR Music Research AB), the solution was designed for market deployment. Specific pricing is not available in the project data — contact the coordinator for current licensing terms.

Can this handle industrial-scale music transcription workloads?

The system was built as a cloud-based solution, which inherently supports scaling. The ScoreCloud platform already had a large worldwide user base for monophonic analysis before this project. The polyphonic extension was designed for the same cloud infrastructure, suggesting it can handle significant throughput.

What is the IP situation — can we license this?

DoReMIR Music Research AB is a Swedish private commercial company (SME) and sole consortium partner. As the only participant, they hold full IP rights over the 13 deliverables produced. Licensing would be negotiated directly with DoReMIR. The product suite is branded ScoreCloud.

How accurate is the polyphonic transcription?

The project used an interdisciplinary approach combining musicology, acoustics, audio engineering, cognitive science, and machine learning to model high-level musical knowledge. Based on available project data, the system was designed to communicate results in musically meaningful terms, but specific accuracy benchmarks are not published in the deliverable descriptions.

Is this still actively maintained and available?

The project ran from April 2015 to September 2017 and is now closed. The project website was x-score.eu and the product suite is ScoreCloud. Based on available project data, the company had existing commercial products before the project, suggesting continued development is likely — but current status should be verified directly.

What file formats and instruments does it support?

The project focused on polyphonic audio transcription — meaning multiple simultaneous notes and instruments. It extends their existing monophonic (single-note) capability. Based on available project data, specific supported formats and instrument types are not detailed in the deliverable descriptions.

Consortium

Who built it

This is a single-company project: DoReMIR Music Research AB, a Swedish SME that is 100% of the consortium. With 1 partner from 1 country and a fully industrial composition, this is a classic SME Instrument Phase 2 setup — one commercial company taking its own technology to market with EU co-funding. There are no university or research partners, which means all IP stays with one entity, simplifying any licensing conversation. The company already had commercial products and a worldwide user base before the project, so this is not a startup experiment — it is an established music-tech company expanding its product line.

DOREMIR MUSIC RESEARCH ABCoordinator · SE

How to reach the team

DoReMIR Music Research AB is a Swedish music technology SME — reach out via their ScoreCloud product channels or company website for licensing and partnership inquiries.

Order a report on this consortium See DOREMIR MUSIC RESEARCH AB profile →

Next steps

Talk to the team behind this work.

Want to explore how automatic music transcription could fit your business? SciTransfer can connect you directly with the DoReMIR team and help evaluate the technology for your specific use case.

Order a report on this topic More in Digital & ICT