In networked audio-visual teaching environments, the processing and transmission of vocal and instrumental performance signals require reliable feature extraction, low-latency communication, and adaptive decision mechanisms that are consistent with engineering approaches used in intelligent sensing and wireless communication systems. This study addresses the substantial differences in skill level, learning style, and development direction among vocationalcollege music performance students by constructing a personalized teaching recommendation system that integrates collaborative filtering with content-based filtering. S tudent p erformance b ehavior, v ocal p ractice r ecords, a nd instrumental audio features are collected and transformed into multidimensional learner profiles, enabling the system to generate differentiated teaching-resource allocation schemes and dynamic practice paths. A quasi-experimental design was implemented with two parallel classes in a vocational-college music performance program, and teaching effectiveness was evaluated through skill achievement, learning autonomy, and resource-utilization efficiency. The results show that the proposed model significantly improves students’ vocal and performance skill acquisition, enhances learning motivation, and increases the matching accuracy of teaching resources. The study provides an operational paradigm for intelligent art-teaching transformation and offers a data-driven reference for adaptive signal-based learning systems.
Cao et al. (2026) studied this question.