Music generation in virtual reality based on grouping-combining algorithm model and multi-style chord music generation network

The quality of generated music is crucial for whether an artificial intelligence music generation model can be applied in VR scenarios. Regarding the above issues, the research proposes a virtual reality music generation model based on step-by-step combinatorial algorithm and multi-style chordal music generating network and finally sets up experiments for verification. The experiment outcomes show that in the simulation experiments, the average similarity between the works generated by the proposed step-by-step combination algorithm and the respective style templates are 94.11%, 93.74% and 93.65%, respectively. After further improvement, their average similarities were 98.19%, 97.17% and 95.28%, respectively. At this time, the similarity range of works generated by other algorithms is 70.13%-88.66%. In addition, in the actual performance experiments, the average rating of the works generated by the proposed step-by-step combination algorithm is 7.62, and the average absolute rating deviation is 1.00, while the average rating and deviation of the multi-style chordal music generation network are 8.14 and 0.76, respectively. To summarize, the proposed virtual reality music generation model is capable of producing high-quality musical compositions, thereby facilitating the broader application of virtual reality technology across various domains.