Comparison of Algorithmic Music Composition: Translational Models, Mathematical Models, and AI Tools

Algorithmic composition, the usage of computational processes to create musical works, has evolved significantly with advances in technology, offering composers new creative possibilities. This study presents a comparative analysis of three distinct approaches to algorithmic music composition, i.e., translational models, mathematical models, and AI tools. Translational models establish mappings between non-musical domains and musical parameters, creating compositions that reflect patterns from diverse sources. Mathematical models leverage formal structures like stochastic processes, fractals, and combinatorial techniques to generate music with distinctive organizational principles. AI tools, particularly neural networks and transformer models, learn patterns directly from musical data to generate compositions that capture nuanced stylistic elements. Through detailed examination of each method's principles, technical implementation, and applications, this research reveals their complementary nature rather than competing alternatives. Each approach offers unique advantages while facing specific limitations in areas such as control, accessibility, and aesthetic flexibility. It is concluded that the most promising direction for algorithmic composition lies in the integration of these different approaches, combining the cross-domain connections of translational models, the structural elegance of mathematical approaches, and the pattern-learning capabilities of AI tools. Such hybrid systems could leverage the strengths of each method while mitigating their individual limitations, expanding the creative possibilities available to composers and enriching the understanding of computational creativity.