Font Size: a A A

Real-Time Analysis System Based On Audio For Network Radio And Television

Posted on:2011-12-11Degree:MasterType:Thesis
Country:ChinaCandidate:S D SunFull Text:PDF
GTID:2178330338479968Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
In recent years, with the development of computer and network technology, the way people get information from Internet is turning from text to audio and video program. The rapid development of network and growing demand of audiovisual experience have provided traditional media a broad space for development. Based on this background, radio and television on internet emerged and rapid develop. The radio and television programs are with rich content and contain great audio and video information, however, faced with such huge information, there is lack of technology that can be effective for processing this information.Based on the above background, this paper develops a feasible system for automatic information processing of radio and television programs on Internet. This system accesses those programs on Internet directly, and classifies them based on the audio information extracted from them. The main work is summarized as follows.1. This thesis studies the detailed requirements of the system, and based on this, gives the design scheme of system frame. Then the system is divided into several function modules reasonably, and the specific function and interface definition of each module are given.2. The technology of programs acquisition and compressed audio decoding is studied then. It includes the realization of several popular streaming protocol, data buffer support scheduling, analysis of popular streaming media file format and technology of decoding and re-sampling compressed audio.3. The classification of radio and television program based on special audio scene is also studied. Firstly, the features of target audio with special scene are analyzed, and choose the best features which can differ those target audio from the others. Then the SVM (Support Vector Machine) method is used to classify short-time audio. Future more, the long-time features of the audio stream are also considered to verify the result. Finally, the correctness and effectiveness of this method is proved by some test.4. Finally, function and performance testing is given to the system. The aspects of the system which need improving are also discussed.
Keywords/Search Tags:Internet Protocol Television, Streaming Media Transmission Protocol, Streaming Media Decoding, Audio Scene Classification
PDF Full Text Request
Related items