A Case Study of AI Application in Singing Research

In the interpretation and performance style of the song "Chun Hua Qiu Yue He Shi Liao"(虞美人•春花秋月何时了) the four singers, YiJie Shi, Sen Ren, AI for YiJie Shi, and AI for Sen Ren, exhibit various differences in their approach of singing the same piece due to their diverse educational backgrounds, vocal training methods, and aesthetic perspective. We try to find the differences about different ages, educational background, singing experience between people and people, people and machines on the work of the second creation by analyzing the singing of the two singers and their robot AI, which is extracted and imitated by their timbre, and highlights the difference in the treatment of the works of the two singers. The audio recordings of these four singers (two of them are AI) were selected and imported into music visualization analysis software. Using pitch processing, tempo processing, dynamics processing, and timbre processing as entry points, the different performance approaches of these four singers were quantitatively analyzed and compared. Based on quantitative analysis, a scientific interpretation of their singing styles is provided, which opens new research methods and perspectives for sing research. It is a typical case study of AI technology enhancing learning and teaching.