Introduction

Music is said to be the language of emotions and the activity of listening to music is indeed a part of everyday life.Musical aspects definitely play an important role in deciding the emotion of a song. Research into music and emotion seeks to understand the psychological relationship between human affect and music. The field, a branch of music psychology, covers numerous areas of study, including the nature of emotional reactions to music, how characteristics of the listener may determine which emotions are felt, and which components of a musical composition or performance may elicit certain reactions. The research draws upon, and has significant implications for, such areas as philosophy, musicology, music therapy, music theory and aesthetics, as well as the acts of musical composition and of musical performance.

Label Prediction

Rule based method is used to detect the emotions of the whole song by using the predictive feature of the arousal and valence values. The arousal and valence prediction values are representing with matrices of frequency for audio and lyrics Emotion label are Predicted By Arousal and Valence values

Feature Variables from Music .mp3/.wav files

Structural features are divided into two parts, segmental features and suprasegmental features. Segmental features are the individual sounds or tones that make up the music; this includes acoustic structures such as duration, amplitude, and pitch. Suprasegmental features are the foundational structures of a piece, such as melody, tempo and rhythm. There are a number of specific musical features that are highly associated with particular emotions. Within the factors affecting emotional expression in music, tempo is typically regarded as the most important, but a number of other factors, such as mode, loudness, and melody, also influence the emotional valence of the piece.

Refrences:

A video explanation of the repo and research problem domain: https://youtu.be/5b5unjdikPo