Collaborative Song Dataset (CoSoD): An annotated dataset of multi-artist collaborations in popular music

TL;DR

The paper introduces CoSoD, a detailed dataset of 331 multi-artist collaborations from Billboard charts, annotated with musical and vocal production features, enabling studies on song structure, vocal production, and gender differences.

Contribution

This work provides a new annotated dataset focusing on multi-artist collaborations, including vocal and production details, and demonstrates its use in analyzing gender-related audio features.

Findings

Men's voices have less reverberation than women's.

Men's vocals are positioned more narrowly in stereo mixes.

The dataset supports research on musical collaboration and gender analysis.

Abstract

The Collaborative Song Dataset (CoSoD) is a corpus of 331 multi-artist collaborations from the 2010-2019 Billboard "Hot 100" year-end charts. The corpus is annotated with formal sections, aspects of vocal production (including reverberation, layering, panning, and gender of the performers), and relevant metadata. CoSoD complements other popular music datasets by focusing exclusively on musical collaborations between independent acts. In addition to facilitating the study of song form and vocal production, CoSoD allows for the in-depth study of gender as it relates to various timbral, pitch, and formal parameters in musical collaborations. In this paper, we detail the contents of the dataset and outline the annotation process. We also present an experiment using CoSoD that examines how the use of reverberation, layering, and panning are related to the gender of the artist. In this experiment, we find that men's voices are on average treated with less reverberation and occupy a more narrow position in the stereo mix than women's voices.