Music chord sequence recognition method based on audio feature extraction algorithm

With the continuous development of music information retrieval, the traditional music retrieval method based on keywords is more and more limited in the application of massive song search on the Internet. Therefore, more and more researchers begin to turn to the research of content-based music retrieval technology. Chord is a typical middle-level feature in music information. It can fully express all the attributes of music. Chord recognition is the basis of automatic music annotation. It plays a very important role in the fields of computer music accompaniment, massive search of cover songs, audio signal segmentation and matching and so on. In this paper, the key technologies of content-based music retrieval are deeply studied: Piano single note recognition method, music chord feature representation method and recognition algorithm are proposed. The characteristic of PCP with higher frequency and its calculation method are used, and the specific calculation formula is given. When using the features of this paper for recognition, the recognition rate of the methods proposed in this paper is higher than that of the current recognition methods.