Search

Tools and methodologies for distinguishing computer generated melodies from human-composed melodies have a broad range of applications from detecting copyright infringement through the evaluation of generative music systems to facilitating transparent and explainable AI. This paper reviews a data challenge on distinguishing computergenerated melodies from human-composed melodies held in association with the Conference on Sound and Music Technology (CSMT) in 2020. An investigation of the submitted systems and the results are presented first. Besides the structure of the proposed models, the paper investigates two important factors that were identified as contributors to good model performance: the specific music features and the music representation used. Through an analysis of the submissions, important melody-related music features have been identified. Encoding or representation of the music in the context of neural network modes are found noticeably impacting system performance through an experiment where the top-ranked system was re-implemented with different input representations for comparison purposes. Besides demonstrating the feasibility of developing an objective music composition evaluation system, the investigation presented in this paper also reveals some important limitations of current music composition systems opening opportunities for future work in the community. [ABSTRACT FROM AUTHOR]

Music is a powerful medium for expressing emotions and feelings acquired during childhood; we are used to improvise rhythms and melodies, memorize and imagine harmonies and even full songs. However, due to the complexity involved in the personal expression through an instrument, it becomes necessary not only to completely master it, but also understand the theoretical concepts that dictate the rules to which melodies must obey. Assuming that most people with little or no musical knowledge whatsoever can create melodies using only their voice, this project's aim is to identify and promote these intuitions using the design of an interactive system. Ultimately, it will become possible to convey melodic ideas by changing the way an instrument's perception is later conducted and thus offering the possibility of musical expression to anyone. [ABSTRACT FROM AUTHOR]

Recently, recommendation technology, aimed at providing information suited to user preference, has become a very active research topic. Previously, we developed a recommendation system using rough sets taking sequential data as their object. We have attempted to apply this system to music recommendations based on chord progression patterns. However, because this method used fixed length subsequences as attributes for the rough sets, the musical meaning of the subsequences was completely ignored. In our current research, we extend this method to allow handling of variable length subsequences to generate musically meaningful subsequences. This paper evaluates the effectiveness of the method through music melody recommendations based on chord progression patterns. The relevance of the extracted patterns to musical meaning is examined using Japanese melodies. We show that this method improves upon the recommendation accuracy of a traditional method and that it extracts meaningful chord progressions that reflect the lyrics of a song. [ABSTRACT FROM AUTHOR]