PopMash: an automatic musical-mashup system using computation of musical and lyrical agreement for transitions

Musical-mashup is a popular form of music re-creation, aiming at combining multiple pieces of music to create new music artworks. Presently, it is also a challenge in the field of music information study. In this work, an effective framework for harmonious musical-mashup generation is provided. In the experiment, lyrics, melody, and rhythm of music were synthetically analyzed. The “harmony” of mashup transition was evaluated in view of the similarity scores of rhythm, melody and lyrics rhyme. “Mashupable” song segments were selected based on the transition harmony evaluation. Then, the musical-mashup output was carried out by adjusting the rhythm, tone and loudness of each segment. Finally, we created PopMash based on the proposed method, an automatic musical-mashup system that can make smooth and harmony transitions from multiple perspectives, which can efficiently reduce the manual work of music recreation.