Research and development of VR system for training vocal music in Vietnam

Virtual Reality has seen growing adoption in vocal training, demonstrating notable practical advantages. However, most current approaches primarily focus on simulating spatial environments and external body expressions during singing, without adequately representing the internal anatomical mechanisms that play a crucial role in phonation. This study proposes and implements an anatomy-based VR-supported vocal training system, in which both internal and external components of the body during singing are visualized through VR technology. Furthermore, we introduce animation mechanisms and VR performance models tailored to the characteristics of Vietnamese phonation, including: multilayer anatomical singing visualization, a phoneme-aligned Vietnamese lip-sync and articulatory animation dictionary (PVL&AA), an automatic animation support system, and airflow control integrating particle and multi-path techniques. The system was practically deployed at the Hanoi College of Art. Post-deployment evaluation was conducted via surveys with 15 lecturers, 115 students, and 12 experts. The results show that most lecturers and students considered the system useful in vocal training, with the feasibility of course content and teaching highly rated across multiple criteria. Some challenges related to system operation and adaptation to technology among users in the arts field were also reported. This study contributes to the innovation of vocal training methodologies and suggests a novel approach to integrating virtual reality and anatomical modeling into vocal education, both in current practice and future applications.