Home

Welcome to the musicGeneration wiki!

Researching the intersection between complex networks and music, the article called “Composing Music With Complex Networks” by Liu X. et al. [1] was one of the ones that attracted me the most. In this paper, the authors create a complex network from various classical pieces and a Chinese pop song, and create melodies using three different algorithms. Although it is very interesting, the generated melodies are not pleasant to the human ear. To quote, they suggest that a human aspect is necessary in music generation:

“The resulting reconstructed music are still very numerous and not all sound appealing. An optimization (selection) process is needed to pick the finalist, and it will be a challenging task to study how the human brain does the selection in the process of composing music.” [1]

What I hope to accomplish in this project is to add a human aspect to algorithmic music generation from complex networks. I plan to build a system that will generate melodies from complex networks, ask for its user to rank these melodies, then update the network according to their preferences and generate a new set of melodies for the user to rank.

For this project, I am imagining the user to be a musician who is looking for ideas and inspiration. So the system does not aim to produce entire songs or pieces, rather aims to come up with melodies that can influence musicians.