CrEATE: Crowd-sourced Evaluation and Annotation of Tonal Entities.


The computational processing of music information has become a crucial need for music industry, education, and research over the last decade. Over the last decades vast amounts of musical data and user annotations have become available through the massive digitalisation of music content and user interaction with music services. However, exploiting these big data repositories for developing effective classification, content-based retrieval, and novel interaction methods to enable users to engage with big data collections of music, remains a major challenge. The CrEATE project will establish a long-term Public-Private Partnership (PPP) between Utrecht University (UU) and the Dutch music start-up Chordify with the purpose of addressing this challenge in the concrete context of automatic chord recognition.

Chordify is a Dutch start-up that offers a free online service that gamifies music e-learning. The service heavily relies on Automatic Chord Recognition (ACR) technology and is currently used by three million users each month. UU has a long standing tradition in Music Information Retrieval (MIR) and gaming research. The CrEATE PPP has two goals: (1) applying for a follow-up research grant in the context of NWO/STW/CLICK or Horizon2020 and (2) a pilot music annotation project. In this pilot project we will investigate how multiple crowd-sourced harmonic annotations can be combined into a high-quality harmonic representation that can be used to evaluate and train chord recognition algorithms. The proposed PPP combines unique MIR expertise of UU with big data collected by Chordify and will propel research into music search engines, and valorise music transcription technologies.





Dr. A. Volk

Verbonden aan

Koninklijke Nederlandse Akademie van Wetenschappen, Meertens Instituut


31/01/2015 tot 18/04/2016