Two songs from the test set were randomly selected. Tags predicted by the GBT + features model (traditional method set as baseline) and by the proposed deep learning model (with a musically-motivated spectrogram front-end) are compared to human annotated tags.
female vocals, triple meter, acoustic, classical music, baroque period, lead vocals, string ensemble, major, compositional dominance of: lead vocals and melody.
acoustic, triple meter, string ensemble, classical music, baroque period, classic period, string solo, major, compositional dominance of: melody and form.
acoustic, string ensemble, classical music, period baroque, major, compositional dominance of: the arrangement, form, performance, rhythm and lead vocals.
'baroque period' and 'classic period' are mutually exlusive tags predicted with high confidence for the GBT + features model.
'triple meter' is not in the deep learning list.
English, male vocals, rap, East Coast, breathy vocal, joyful lyrics and compositional dominance of: lyrics, melody, rhythm, accompanying vocals.
East Coast, West Coast, rap, hardcore, lead vocals, funk, old school, drums, electronic drums, party.
English, lead vocals, male vocals, rap, accompanying vocals, danceable and compositional dominance of: accompanying vocals, lead vocals, rhythm, lyrics.
The deep learning model is biased towards popular tags, and a relevant (but less popular) tag is not in the list: 'East Coast'.
'East Coast' and 'West Coast' are mutually exlusive tags that are predicted with high confidence by the GBT model: 0.98 and 0.97, respectively.
Interestingly, although Kendrick Lamar is a 'West Coast' rapper, this particular song has an 'East Coast' sound.