Application of dynamic segmentation and weighted comprehensive matching in music style recognition

Content-based music application has brought great changes in the way of human-computer interaction in music. Humming retrieval system enables people to find songs by humming tunes, which expands the traditional text retrieval method; By analyzing users' music preferences, the music recommendation system actively pushes songs close to the style, which saves music fans the trouble of "searching for songs", and can also make new songs quickly appear around potential lovers; Various online karaoke systems and music teaching systems help users improve their singing and performance skills by analyzing the music they sing or play and giving evaluation and feedback. Music recognition technology is the hub connecting music applications and real music. Through music recognition technology, the computer can automatically recognize the melody, rhythm, type, style and other information of music. With the development of Internet and digital audio technology, music information retrieval (MIR) has gradually become a research hotspot. Among them, the recognition of music style is an important research content; In addition, the recognition of musical instruments, songs and emotions is also a research hotspot. In the field of pattern recognition, music retrieval algorithm has attracted more and more attention because of its easy and efficient characteristics. However, the imprecision of note segmentation and the inconsistency of matching template directly affect the accuracy of Mir algorithm. In order to solve the above problems, a dynamic segmentation and weighted comprehensive matching algorithm based on threshold is proposed. Dynamically set the amplitude difference step and segment the notes according to the change threshold to improve the accuracy of note segmentation. The standard score frequency and pitch template are used to realize input normalization, so as to enhance the accuracy of matching and the adaptability and robustness of the algorithm. Finally, experiments show the effectiveness of this method.