Design of a Singing Evaluation System of Heyuan Hua Chao Opera Based on Mel-Frequency Cepstral Coefficients

Abstract

Heyuan Hua Chao opera is affected by modern culture and is facing a limited number of teaching staff, little time available for teachers to be dispatched, a small scope of popularity, the inability for continuous tracking in teaching, unsystematic and coherent learning among students, and difficulty of independent learning. Therefore, the voice feature parameters required for the evaluation model were extracted through MFCC coefficient features and later input into a convolutional neural network to generate data sets and training sets. These feature-labeled sets are segmented to output singing feedback. With a comprehensive overview of the artistic characteristics of Hua Chao opera singing and the research and interviews conducted by the Hua Chao opera heritage development center and local people, a system is designed to evaluate the singing voice of Hua Chao opera singers based on the MFCC. The aim is to effectively help students participate in learning and intelligently evaluate their learning effects. The model can be applied to other opera repertoires to promote the preservation and dissemination of traditional opera culture.

Keywords

• Mel-Frequency Cepstral Coefficients
• Hua Chao opera
• vocal evaluation
• opera into the campus