Multiple Chord Coding in China’s Minority Dance Music

Abstract

Abstract In this paper, we first use CNNS to collect rich samples of folk dance music, establish a model framework and a functional system framework, and conduct a comprehensive process analysis of the music. Then, the instrumental features, frequency features, and timbre features are extracted to obtain the spectral information. In the stage of chord analysis and encoding, a multivariate chord encoding model is established based on the acquired spectral information, including two parts: chord representation preprocessing and chord encoding. By utilizing this model, the chord structure of music was successfully and accurately encoded, allowing for analysis with up to 98% accuracy. Furthermore, significant recall results were achieved, reaching over 0.9, which suggests that the extracted chord features are highly reliable and accurate in recognizing musical chord information.