Voice sectional song search method

Abstract

The invention discloses a voice sectional song search method, which is realized through interaction of a client side and a server side; and a song database is required to be established at the server side. The method comprises the following steps of: during search, the client side prompts a user to speak out song information to be searched in sections and transmits the voice spoken by the user to the server side after the user finishes speaking; after receiving a voice signal, the server side automatically identifies a character corresponding to the voice signal, then searches in the song database in layers and transmits a search result to the client side; and finally, the client side provides the song search result to the user. According to the method, a rational voice interaction flow is designed at the client side, so that a super large song search space can be decomposed into combination of a plurality of smaller song search spaces by the server side; and thus, search efficiency and automatic voice recognition accuracy in the super large song search space are improved, user experience is improved, meanwhile, hardware cost is saved for service providers.

Description

technical field [0001] The invention relates to a voice segmentation type song retrieval method. Background technique [0002] Song download is a very active business category in the current mobile and Internet fields. According to the user's search requirements provided by the client, the service finds songs that meet the search conditions in the server's music library for the user to download. The most common example is that the user inputs the name of a singer or song on the terminal, and the server finds one or a series of songs corresponding to the singer or song name and returns it to the user. [0003] On a mobile terminal, due to the limitation of the physical size of the terminal, the efficiency of the input method is usually relatively low, and it will be very time-consuming for the user to input a complete song name through the keyboard or screen. In the mainstream position, it is very common for users to input homonyms and typos. For example, entering Faye Wong...

AI Technical Summary

Problems solved by technology

[0004] However, speech recognition technology has a natural weakness, similar to the human ear experience, speech recognition is not 100% accurate, and the accuracy of machine speech recognition decreases with the increase of the database, for example, with the song title database The increase of the number of songs with similar pronunciation will also increase, such as Wilber Pan's "Turning Turn" and Faye Wong's "Legend", which will seriously reduce the accuracy of speech recognition

In order to cover the needs of most users, a medium-sized song database usually contains hundreds of thousands of independent tracks. Speech recognition in such a large range is a very big challenge for recognition accuracy.

At the same time, speech recognition algorithms are usually computationally intensive, and their computational load increases rapidly with the increase of the database, which is a very big challenge to the computing power of the server. In order to deal with the concurrent needs of multiple users in a timely manner, the service Providers often need to provide more hardware resources, increasing costs

Images