Music emotion analysis

This web page represents results for the following paper:

Wiafe A, Sieranoja S, Bhuiyan A, Fränti P, Emotify+: Music dataset with annotated emotions.

Music emotion recognition

A web-based music platform is developed to study the effects of music on emotion.

The Emotify music dataset (Aljanaki et al., 2016) which consists of 400 songs (44100 Hz, 128 kbps, one minute each) from four different genres: classical, rock, pop and electronic music (100 tracks per genre) is used.

A listener can listen to music, select, and rate the emotions they perceive.

Participants discovered new music based on the emotional content of music and assisted in the gathering of "ground truth" data.

Music data: Emotions perceived:
Emotions annotated results: