Mathematical modelling and a greedy heuristic for harmonic mixing and popularity-based playlist generation problem

Year 2024, , 487 - 496, 31.12.2024

Zülkar Karakaya Zeliha Ergül Aydın

https://doi.org/10.56554/jtom.1387264

Abstract

The rise of the electronic music industry has led to a need for creative playlist-generation methods, particularly for DJs aiming to deliver seamless and harmonically enhanced performances. Harmonic mixing, a crucial process of DJing, involves synchronizing and aligning songs based on musical harmony, making the mix sound soft and clear. In harmonic mixing, the DJ has to select songs from the extensive music archive, considering notes, tempo, length, and popularity of the songs. However, manually generating playlists that adhere to harmonic mixing principles can be time-consuming. This paper introduces a mixed-integer mathematical model and a novel greedy heuristic to automate playlist generation, considering factors like popularity, tempo, and harmonic mixing rules. We compare the novel greedy heuristic's performance to the mathematical model on 16 test problems created with Spotify's API, incorporating real-world data on song characteristics. The results show that the heuristic method generates playlists at least seven times faster and has an average gap of 13.84% with the mathematical model.

Keywords

Harmonic mixing, Playlist generation, Greedy algorithm

Ethical Statement

The authors declared that there is no conflict of interest.

Thanks

This study is supported by Eskisehir Technical University Scientific Research Projects Committee (ESTUBAP-22LÖP394).

References

• Bittner, R. M., Gu, M., Hernandez, G., Humphrey, E. J., Jehan, T., McCurry, H., & Montecchio, N. (2017, October). Automatic Playlist Sequencing and Transitions. In ISMIR (pp. 442-448).
• Bonnin, G., & Jannach, D. (2014). Automated generation of music playlists: Survey and experiments. ACM Computing Surveys (CSUR), 47(2), 1-35. https://dl.acm.org/doi/10.1145/2652481
• Dias, R., Gonçalves, D., & Fonseca, M. J. (2017). From manual to assisted playlist creation: a survey. Multimedia Tools and Applications, 76, 14375-14403. https://doi.org/10.1007/s11042-016-3836-x
• Fields, B., Lamere, P., & Hornby, N. (2010, August). Finding a path through the juke box: The playlist tutorial. In 11th International Society for Music Information Retrieval Conference (ISMIR).
• Gabbolini, G., & Bridge, D. (2024). Surveying More Than Two Decades of Music Information Retrieval Research on Playlists. ACM Transactions on Intelligent Systems and Technology. https://doi.org/10.1145/3688398
• Gebhardt, R., Davies, M., & Seeber, B. (2016). Psychoacoustic Approaches for Harmonic Music Mixing. Applied Sciences, 6(5), 123. https://doi.org/10.3390/app6050123
• Hartono, P., & Yoshitake, R. (2013). Automatic playlist generation from self-organizing music map. Journal of Signal Processing, 17(1), 11-19. https://doi.org/10.2299/jsp.17.11
• Hsu, J. L., & Lai, Y. C. (2014). Automatic playlist generation by applying tabu search. International Journal of Machine Learning and Cybernetics, 5, 553-568. https://doi.org/10.1007/s13042-013-0151-y
• Kahanda, I., & Kanewala, U. (2007) PlayGen: A HYBRID PLAYLIST GENERATOR, in Annual Technical Conference 2007 of IET-YMS
• Mocholi, J. A., Martinez, V., Jaen, J., & Catala, A. (2012). A multicriteria ant colony algorithm for generating music playlists. Expert Systems with Applications, 39(3), 2270-2278. https://doi.org/10.1016/j.eswa.2011.07.131
• Pauws, S., Verhaegh, W., & Vossen, M. (2008). Music playlist generation by adapted simulated annealing. Information Sciences, 178(3), 647-662. https://doi.org/10.1016/j.ins.2007.08.019
• Pohle, T., Pampalk, E., & Widmer, G. (2005, September). Generating similarity-based playlists using traveling salesman algorithms. In Proceedings of the 8th International Conference on Digital Audio Effects (DAFx-05) (pp.220-225).
• Pohle, T., Knees, P., Schedl, M., Pampalk, E., & Widmer, G. (2007). “Reinventing the wheel”: a novel approach to music player interfaces. IEEE Transactions on Multimedia, 9(3), 567-575. https://doi.org/10.1109/TMM.2006.887991
• Shuhendler, R., & Rabin, N. (2024). Dynamic artist-based embeddings with application to playlist generation. Engineering Applications of Artificial Intelligence, 129, 107604. https://doi.org/10.1016/j.engappai.2023.107604 SpotifyWebAPI, Spotify for developers, 2023. Available a https://developer.spotify.com/documentation/web-api