Audio retrieval is based on waveform sound retrieval and an important part of multimedia retrieval technology. We all know that audio information, as an opaque bit stream, has external properties, but it lacks entities that can be matched with keywords. We can convert audio into text for retrieval through methods. But the disadvantages are: the amount of data is increasing, and the workload of manual labeling is increasing; People's perception of audio is inaccurate; Real-time audio data stream retrieval is not supported. In order to solve these problems, audio retrieval came into being. Today, according to the development of audio retrieval technology, we will discuss some key technologies in the research of audio retrieval technology at this stage: audio feature extraction, audio classification, speech recognition technology, feature similarity matching and so on.

Audio information retrieval technology has broad application prospects: (l) it is the core technology of intelligent voice system, and users can obtain the required information conveniently and quickly through this technology; (2) It can realize real-time retrieval, examination and effective monitoring of audio information of media in Internet TV, audio and video, and protect the copyright of various digital audio products, such as music, that is, unauthorized use. (3) It also plays an important role in the research of audio information classification and statistical technology. For example, radio and television news programs, digital libraries and other contents all contain a lot of information such as voice and music. Using audio information retrieval technology can effectively classify, count and retrieve these information, and better manage, store and use it efficiently.

These resources. In addition, it can also be used in network management, information security, military criminal investigation and many other fields.