Research on Aesthetic Features in Computer Music Based on Spectrum Recognition and Matching Algorithm

Based on the spectrum recognition and matching, this paper conducts in-depth analysis and research on the aesthetic characteristics of computer music, and constructs a set of sound signal extraction methods based on the dynamic characteristics of the spectrum and an aesthetic feature recognition algorithm with Extreme Learning Machine (ELM) as the classifier. Adopt the feature extraction method based on Mel frequency cepstral coefficient (MFCC), the feature extraction method based on first-order differential Mel frequency cepstral coefficient (MFCC), the feature extraction method based on second-order differential Mel frequency cepstral coefficient (MFCC), and the dynamic feature based on spectrum In the sound signal extraction method, there are 71 classification results based on tags, with an accuracy of 97.52%.