Study On Computer Aided Algorithmic Composition Approaches And System Designs

This thesis concentrates on computer composition techniques based on Markov model and rule-based systems. It also discusses the important melodies and drawbacks associated with these two algorithmic composition approaches. The corresponding solutions are proposed to address the issues. By the adoption of these two techniques, we can study how the fundamental elements in classic music deducts in music composition, the way that music themes developing in the deduction process, the structural role of the forming logic in music development and the reasoning process of the music style under different structures. The key points in this work lie in the way how musical parameters can be quantized described in computer language and automatically produce a partial or entire music works. We also designed and developed an independent software system, called "Sound in a System", which combines the functionalities of editing, score input, algorithmic composition and other easy-using functions. This is a multidiscipline research field of computer and composition techniques, which may include the content of computer programming, software engineering, database techniques, data analysis, artificial intelligence, traditional and modern composition techniques. In the software platform, the composition system surpasses the previous algorithmic composition structure and adopts a hybrid system which uses multiple approaches as a whole, we believe this will be the trend in algorithmic composition practice. The system also provides a flexible interactive interface between user and computer. This will form the underpinnings of the next generation algorithmic composition systems and have extensive value in serious electronic music and game music composition.