Research And Design Of Digital Audio Content Monitoring System Based On FPGA

With the progress and development of science and technology, radio and televisiontechnology is also in constant progress, the number of radio and television programsalso increases a lot. The traditional way of monitoring under great challenge, so neednew type of automatic monitoring system to adapt to the development of radio andtelevision technology.At the same time, the market now mainly in the use of audiomonitoring system is based the text library monitor, this way of monitoring is limited bythe size and content of the text library, it is not commonly used.Therefore, following acontent-based audio monitoring technology has been widely studied. It is different fromthe way of the previous text library, we need to extract the characteristic parameters thatreflect the characteristics of the audio signal information, and could make the system tohave a stronger ability to adapt. At the same time,the content-based audio monitoringtechnology is the use of real-time monitoring method,and could quickly give the resultsof monitoring.In this paper, we study the issues of content based audio monitoring. The purposeof this study is to monitor whether there is a change of content is played, mainly relatedto two aspects with the broadcast program and advertising. This will not only protect theintellectual property of the radio and television producers, and also let people can seethe safe radio and television programs. The system uses FPGA chip, which finallymakes the system has the advantages of small volume, and is convenient for systemupgrading.According to the requirements of the monitoring system, we particular analyzethe main characteristics of the speech signal parameters, introduces a time-domaincharacteristic parameters of the speech signal. Finally according to the actual use,weselected short-time average zero-crossing rate and short-term average level these twocharacteristic parameters together as the characteristic parameters of the monitoringsystem.These two characteristic parameters are based on the short-time characteristic ofthe speech signal.This requires before extracting feature data have to pretreatment ofspeech signal as the window frame. Considering as a television sound monitoring system, so we choose the ODD signal in the video signal as a frame signal.So that thetwo signals can do coarse synchronization requirements. The two characteristicparameters have a small amount of data, can facilitate remote extracted the features ofdata transmission through the network to the local monitoring terminal. After a lot ofdifferent transmission path, the two signals could not synchronize, so thesynchronization of signals is important to work out the issue. Based on the analysis ofdelay information, we put forward to solve the problem of signal searching and trackingtechnology, fixed integer and fractional frame delay in the local to approximate thecharacteristics of remote data. In order to accelerate the speed of search and tracking,the use of an illumination system, namely using a continuous multi-frame of the localfeature data while comparing with the remote feature data. Finally,we do simulation ofeach module of the system.