My research is motivated by my background in both engineering and music, and is within the Music Information Retrieval (MIR) field. I try to understand the way humans describe music and emulate these descriptions by computational models dealing with big music data. My research benefits from disciplines such as signal processing, music theory, cognition and artificial intelligence.
I have developed methods to automatically describe music audio signals in terms of melody, tonality and rhythm; to measure similarity between pieces and automatically classify music according to style, emotion or culture. Over the last years, I have coordinated the PHENICX FP7 STREP EU Project, trying to innovate the way we experience classical music concerts. In addition, I have researched on the computational modeling of flamenco music, improving current techniques for automatic singing voice description. I am currently coordinating the CASAS project, dealing with technologies for choir singing.
You can find an updated list of my publications, including my PhD thesis, at the MTG web site. You can check my citations at google scholar or researcherID. My work is also referenced in wikipedia: HPCP.
Recent research projects:
- CASAS (PI – Spanish Government): Community Assisted Singing Analysis and Synthesis 2016-2018.
- Gathering and Modelling Large-scale Crowdsourced Multimedia Music Data.
2016-2017, co-PI with Gloria Haro of an internal project part of the María de Maeztu program funded by the Spanish Government.
- PHENICX (PI – European Commission FP7 STREP): Changing the way we experience classical music concerts. In this project we want to use ICTs in the context of classical music concerts, in particular in large ensemble setups (symphonic orquestras). In addition to coordinating it, I work on music description (predominant melody, tonality, emotion) and visualization.
- COFLA (Researcher – Junta de Andalucía, Spain): Computational analysis of flamenco music. I research on automatic transcription, similarity and genre/style characterization in flamenco singing.
Technologies (software) available for download:
- MELODIA: Melody extraction vamp plugin (with Justin Salamon)
- HPCP: Harmonic Pitch Class Profiles vamp plugin (with Jordi Bonada)
- Matlab toolbox for multi-scale set-class analysis (with Agustín Martorell)
- CANTE: Automatic transcription of flamenco singing, open source (with Nadine Kroher)
Datasets available for download
- MTG-Query By Humming, Orchset (melody in symphonic music) available at the MTG web site.
- Flamenco datasets (corspus) available at the web site of COFLA project.
I was also a co-founder of BMAT.