What song is next? Acoustic and musical factors involved in the order of album tracks
Main funder
Funder's project number: OPH-4688-2020
Funds granted by main funder (€)
- 18 000,00
Funding program
Project timetable
Project start date: 01/04/2021
Project end date: 31/05/2022
Summary
With the development of music streaming services such as Spotify, Deezer, and Apple Music, researchers have recently engaged in the task of developing intelligent music recommendation systems. Early algorithms were based on factors such as acoustic similarity, melodic style, and collaborative filtering. Leading- edge systems, however, started to account for temporal dynamics of music sequences, hoping that the enjoyability of a song could be maximized by accounting for the order in which it appears within a playlist. Until today, however, no study has taken advantage of the fact that, for nearly a century, the music industry has faced the task of music ordering as a human intelligence problem. Parallel to those interested in music recommendation algorithms, professionals responsible for conceiving music albums have been trying to organize tracks into a sequence that is most cohesive and interesting to the listener. In this work, the project proposes to leverage the existence of these large bodies of human-made music sequences (i.e. albums), and intends to look for patterns in the way that professional musicians and producers order tracks within their music albums.
