The Contribution of Lyrics and Acoustics to Collaborative Understanding of Mood

TL;DR

This study investigates how song lyrics and acoustics contribute to understanding musical mood, demonstrating that transformer models effectively predict mood associations and that lyrics and acoustics vary in importance depending on the mood.

Contribution

It introduces a large-scale data-driven analysis using transformer models to assess the roles of lyrics and acoustics in mood prediction, highlighting their relative importance.

Findings

Pretrained transformer models effectively capture song-mood associations.

Training on song-mood data improves prediction accuracy for unseen songs.

The importance of lyrics versus acoustics varies with the specific mood.

Abstract

In this work, we study the association between song lyrics and mood through a data-driven analysis. Our data set consists of nearly one million songs, with song-mood associations derived from user playlists on the Spotify streaming platform. We take advantage of state-of-the-art natural language processing models based on transformers to learn the association between the lyrics and moods. We find that a pretrained transformer-based language model in a zero-shot setting -- i.e., out of the box with no further training on our data -- is powerful for capturing song-mood associations. Moreover, we illustrate that training on song-mood associations results in a highly accurate model that predicts these associations for unseen songs. Furthermore, by comparing the prediction of a model using lyrics with one using acoustic features, we observe that the relative importance of lyrics for mood prediction in comparison with acoustics depends on the specific mood. Finally, we verify if the models are capturing the same information about lyrics and acoustics as humans through an annotation task where we obtain human judgments of mood-song relevance based on lyrics and acoustics.