A music recognition method, system and device

Abstract

The embodiment of the invention discloses a music identifying method, system and device. The method includes the following steps: a client extracts audio information used for music identification from acquired audio data; the client uploads the audio information used for music identification to a data analysis server; the data analysis server matches the audio information used for music identification with a characteristic database of known music so that a matching similarity is obtained; and when the matching similarity does not meet a music identification success condition, the client continues to upload the audio information used for music identification to the data analysis server and when the matching similarity meets a music identification success condition, the client stops temporarily uploading the audio information used for music identification to the data analysis server. The music identifying method, system and device are capable of keeping balance between ensuring of an identification success rate and saving of data transmission resources.

Description

Technical field [0001] This application relates to the technical field of speech recognition and information processing, and in particular to a method, system and device for music recognition. Background technique [0002] Music recognition technology refers to the technology of collecting the characteristic data of the played music, matching the characteristic data with the characteristic data of each known music in the database, and determining which music the played music belongs to according to the matching result. [0003] Wherein, the music composition may be a singing voice without a soundtrack, a singing voice with a soundtrack, or pure music without a singing voice. The feature data of a music piece can also be called audio fingerprint information, which can express the characteristics of a piece of music such as pitch, loudness, and timbre. [0004] figure 1 It is a detailed flow chart of the current music recognition method. [0005] Such as figure 1 As shown, the process ...

AI Technical Summary

Problems solved by technology

[0015] If the amount of audio data to be recognized manually collected by the user is too small, for example, only 3 seconds of audio data is collected, it will make it difficult for the data analysis server to accurately identify which piece of music the audio data to be recognized belongs to. Low, if the data volume of the audio data to be recognized manually collected by the user is too large, for example, 2 minutes of audio data is collected, it will cause the client to upload the music pattern information of the audio data to be recognized to the data analysis server and consume more data transfer resources

It can be seen that the current music recognition method is difficult to strike a balance between ensuring the success rate of recognition and saving data transmission resources

[0016] Additionally, using figure 1 When performing music recognition by the method shown, if there are multiple pieces of music played in sequence, you need to manually input the command to start collecting and the command to end collecting for each piece of music, in order to obtain the audio data to be recognized, and then realize music recognition. Continuous automatic recognition of multiple songs played in sequence

Images