Exploring the generative AI art technologies in vocal training for increasing aesthetic and emotional perceptions

Abstract

The objective of this article is associated with exploring the approaches to employing generative artificial intelligence artistic technologies in vocal training to enhance aesthetic and emotional perception. Throughout the investigation, the potential utilization of various intelligent technologies for vocal education, including ChatGPT, LANDR, DeepSinger, and VOCALOID, was identified. It was determined that the VOCALOID (97.3) and DeepSinger (95.0) applications exhibit more significant advantages in the development of vocal skills. The research findings revealed that the application of DeepSinger facilitated the enhancement of emotional expressiveness in the respondents' singing (attributed to its capability for natural sound processing). Additionally, VOCALOID contributed to shaping the aesthetics of singing by enabling voice generation and comparison of different performance variations. The study also identified challenges associated with the use of artificial intelligence technologies in education, including a lack of individualized consultations (32%), imprecise interpretation of musical compositions (28%), a deficit of creativity in teaching (23%), and the availability of moderately high-quality instructional materials (17%). At a higher level, the emotional and aesthetic perception of singing by audiences was achieved among respondents who utilized the DeepSinger application during their training (average value: 0.83). Conversely, the least expressive singing was observed among respondents who utilized the ChatGPT application (average value: 0.5) in their training, primarily focused on developing theoretical rather than practical skills. The practical significance of this article lies in cultivating vocal singing skills through the utilization of artificial intelligence technologies such as ChatGPT, LANDR, DeepSinger, and VOCALOID