Despite the difficulties in carrying out regular academic activities in 2021, at the MTG we have been able to continue our research quite succesfully. In this blog post we give some numbers and highlights of the year.

People

Our team continues to be a large one:

The size of the team has been quite stable over the past few years:

Number of MTG members 2016-2021

Funded projects

In 2021 we started 5 new public and private projects (out of 16 current funded projects). These projects, plus the software licenses (11 new ones in 2021 of Essentia and Freesound API) and the Freesound donations, allowed us to support the team and its research work.

Number of projects 2016-2021

Research outputs

In terms of outputs, we had quite a number of new publications:

• Articles: more than 45 (14 journals, >31 conferences) (cumulative > 1000, with a cumulative number of citations > 26500)
• PhD thesis: 7 (cumulative 68)
• Master thesis: 16 (cumulative 261)

and new or improved technologies:

• 10 new datasets in Zenodo (uploaded or updated in 2021). Total downloads (cumulative) >141K
• Mirdata, an open-source Python library for accessing and working with MIR datasets
• Soundata, a Python library for reproducible use of audio datasets
• Cantāmus, a singing rehearsal tool that allows choir singers to practice their part with vocal practice tracks synchronized to the score (by VoctroLabs - TROMPA project)
• SOURCE, an open-source hardware/software music sampler connected to Freesound.
• TROMPA Music Enthusiasts, a citizen science platform to gather and study of “ground truth” data for machine learning models of emotions in music.
• SkyNote app v.1 for iOS, an intelligent music learning app based on sound and motion real-time analysis
• Essentia.js, a JavaScript library that allows running music/audio processing and analysis algorithms/models on your web browser

Essentia continued growing and having a large impact (>10K downloads in 2021) and also Freesound (20.8 millions of sounds downloaded, 1.2 millions of new users in 2021). Specially relevant has been the new ML models of Essentia and the new interface for Freesound.

Outreach

In terms of communication of our research, we were involved in organizing and participating in several conferences. We organized DCASE Workshop, Web Audio Conference, and MML Workshop, and had a significant participation in MediaEval, ISMIR, ICASSP and SMC. Our researchers got some awards for their work:

• Correya, P.Alonso, J.Marcos, D.Bogdanov: Best Paper Award at WAC
• M. Won: Best Student Paper at ISMIR
• R. Caro Repetto: Best Reviewer at ISMIR
• E. Fonseca: Best Audio Representation Learning Paper at WASPAA 
• F. Navarro: Cruz Roja Awards to Humanitarian Technologies
• F. Font: Best Demo Award at the Audio Mostly Conference

We were also active in general dissemination activities, such as:

• New MTG video presenting the current research challenges
• Release of the video The Science Behind Music Technology as part of MIP-Frontiers Project
• Marius Miron at AI and Music Festival organized by Sonar
• Booth of the MTG at ISMIR 2021
• Lorenzo Porcaro participated in the European Researchers’ Night
• Lorenzo Porcaro participated in the MUTEK symposium: A future without gender
• Participation in Artificia sessions on Music and AI
• Emilia Gómez gave a keynote speech at Sound and Music Computing 2021.
• MML-lab participated in 4YFN Mobile World Congress 2021
• TV3 - CTRL (Xavier Serra), Telenotícies (Banda Sonora Vital project)
• Catalunya Radio - Tots els Matins del Mon (Emilia Gómez and Xavier Serra)
• RNE - Longitud de Onda (Emilia Gómez and Juan S. Gómez-Cañón)
• Several articles and interviews related to Gender Imbalance in Music Recommenders (Andrés Ferraro)
• Articles and interviews related to Music and Emotion research by Juan Gómez-Cañón

We kept updating and improving the website to reflect our current research and activities. A recent addition has been the section on demos, which showcases some interactive demos of our technologies (e.g.: Essentia, Source, MusicCritic). We were also very active on our social networks.

• Web: 55 news published on the web and 99K page views
• LinkedIn: 1436 followers (+437 vs previous year)
• Twitter: 2418 followers (+258 vs previous year)
• Youtube: 1384 subscribers (+196 vs previous year)

An important public dissemination effort was done through the Phonos Foundation. Through it we organized 17 concerts, and 8 freesound workshops. We also had 8 artists in residence.

Education

MTG's main social impact is accomplished through our graduating students. In 2021, 17 people finished their master with us and 7 their PhD. These alumni have taken advantage of their MTG experience to get relevant jobs in music technology. 

2021 has been a very strange year but this has not prevented us from advancing in our research and generating significant outputs.