Music is an important part of our lives. With the rise of streaming services, it has become even easier for us to legally access the desired music contents. Music enhances our experiences of entertainment such as film, television, and advertisement. The frequent exposure to a particular piece of music builds our association with that music often linked to certain aspects of the music. Often, it is not within the means of small scale creators to use the music their audiences have the most significant association with due to limited funding, and fear of legal action. The solution is to generate new within the desired genre, based on a model of the genre, with some specific inputs which are weighted more highly (the song we want our generated music to be most similar to).
This project aims to generate music by only dealing with film scores to avoid lyrics-based complexity by break up the pieces of music into discrete baskets, characterizing the connections and associations that are made with that piece of music. Once the input piece has been identified as belonging to one of these baskets, pre-loaded musical motifs (fillers) will be used in conjunction with the elementary rules and principles to generate a music that resembles the source music but still distinct.
Research Components:
Investigate methods and techniques for accurate identification of music with multiple instruments.
Explore different tools and platforms suitable for the implementations music generation application
Undergraduate
A software application for generating the music.
None
Lab allocations have not been finalised