Optimization of Diversified Teaching Mode of Vocal Singing in Colleges and Universities by Integrating Audio-visual and Multi-sensory Senses

Abstract To cultivate students’ perception of music culture through audio-visual multi-sensory methods in college vocal singing teaching is a new way to understand music works and cultivate music culture consciousness. The flipped classroom is the basis for this paper, which aims to innovate the teaching of vocal singing in colleges and universities. The linear prediction of Mel’s cepstral coefficient is used to extract the auditory features of the music symbols of the vocal singing video in the flipped classroom so that the students can master the relevant skills and pronunciation of vocal singing. To help students understand vocal singing in its specific context, the time-frequency diagram is utilized to analyze the singer’s movements in the vocal singing video. The self-attention mechanism is used to fuse auditory and visual features, which is validated and analyzed for its effectiveness, teaching effect, and satisfaction. The results show that when the added noise signal-to-noise ratio reaches 50dB, the recognition rate of the LPMFCC algorithm for vocal singing music signals is above 80%. Utilizing the flipped classroom for vocal singing teaching, the average score of students’ vocal singing increased to 9.71, and more than 80% of the students believed that the flipped classroom teaching mode could enhance their independent learning ability. The flipped classroom teaching mode, which integrates audio-visual and multi-sensory teaching methods, can enrich the teaching methods of vocal singing in colleges and universities and further enhance students’ interest in vocal singing.