Research on Music Style Similarity Detection Algorithm Based on Relative Features

Music is a kind of important audio data, and traditional music retrieval is based on keywords, and its use is limited. With the rapid development of the Internet and the popularization of digital equipment, content-based audio retrieval research has received extensive attention. The music information on the Internet has been massive, and the traditional way of obtaining music can no longer meet people’s search needs. Therefore, this paper conducts in-depth research on the key technologies of content-based music retrieval, and proposes a feature extraction method suitable for music retrieval and similarity measure. The method and retrieval algorithm, and the endpoint detection and continuous word recognition technology in speech recognition are studied, and good retrieval results are obtained. The related results can also be applied to retrieval systems for time series data with similar structures.