The Lyricator - deriving emotion from lyrics
Discovering the mood of a song is important for may reasons. Mood can be used to help make good playlists, to help classify music, and to make sure that a good party doesn't go bad.
The mood of a song can be derived from the audio, or from the lyrics. The Lyricator derives the mood from the lyrics. The Lyricator uses the PAD Emotion scale to
position a song's lyrics in a 3 Dimensional mood space. The three
dimensions are pleasure-displeasure, arousal-nonarousal, and
dominance-submissiveness axes. The Lyricator calcuates the PAD for each
song and then uses this to classify the song as engaging, soothing, boring or annoying/angry.
The Lyricator was written by Owen Meyers at the Mit Media lab. It is written in Python and the source/executable can be downloaded from Owen's Lyricator page.
Here are some lyricator results
Song Title | Pleasure | Arousal | Dominance | Classification |
Eric Clapton's Tears in Heaven | 0.16 | 0.02 | 0.07 | Engaging |
Beatle's Come Together | 0.07 | 0.007 | 0.03 | Boring |
Mama's & the Papa's California Dreamin | .09 | -.02 | .02 | Soothing |
Christina Aguilera Beautiful | .12 | .02 | .05 | Engaging |
Coolio Gangsta's Paradise | .075 | .023 | .032 | Angry |
John Lennon Imagine | .223 | .058 | .115 | Engaging |
Sarah Brightman Everything's Alright | .07 | .005 | .02 | Boring |
Bob Marley I Shot The Sheriff | .06 | .005 | .022 | Boring |
Death Left To Die | .042 | .025 | .0034 | Angry |
Janet Jackson The Pleasure Principle | .188 | .029 | .069 | Engaging |