Archive
IJMSTA - Vol. 8 - Issue 1 - January 2026
ISSN 2612-2146
Pages: 15
Aesthetic Evaluation of AI-Generated Text-Free Music
Authors: Lei Gao, Guodong Wang, Jian Hou, Yi Li, Chongrui Sun
Categories: Journal
Abstract - Music generation technology is in a phase of rapid development, with AI-generated text-free music widely applied across multiple fields such as music creation assistance, independent scene scoring, and music education. As the volume of generated text-free music continues to grow, there are increasingly higher aesthetic demands for the generation outcomes. As a key interdisciplinary topic at the intersection of music aesthetics and music generation technology, the aesthetic evaluation of text-free music holds significant importance for improving the generation quality of models. To enable generative models to produce more text-free music that meets scenario-specific requirements and possesses inherent aesthetic value, it is necessary to further refine the seven dimensions of aesthetic evaluation, optimize the model's generation logic, mitigate the sense of mechanicality and homogenization in generated works, and strengthen their humanistic expressiveness and emotional resonance. This constitutes both the core task and arduous challenge in optimizing current text-free music generation outcomes.
Keywords: Text-free music, Music generation, Music aesthetics, Aesthetic evaluation
Download paper
