Analysis And Design Of Answering Alarm System Based On Speech Enhancement Algorithm

In our daily life, the answering almarm is one of the most important public securitywork.Especially, dealing with various enmergencies and people’s reports,its effects seem tobe more obvious and convient. However, on the actual circumstances, communication qualityis not stable on account of the speech singals which are influenec by different noises. If thenoise influence is serious enough,it will lead to a fact that speech systerm doesn’t work orreceive pronunciational signal.normally. More strikingly,while answering alarms, reportersusually are located in those places which are full of noises and unsafe elements,such as CBDand public amenities and other social activities.the voice of background is very strong,namelythe noise is so serious.Accoring to what have mentioned above,it is easy to understand whythose policemen who are responsible for receiving alarm find it difficult to record exactinformation which are relevetn to life and property.Every unexact information will result insevere aftermath.At this very monent, it is significant for us to improvethe quality of voicecommunication. Therefore,when this happens, the way of signal processing would be used toenhance the speech signal, weaken the impact of the environmental noises,and improve theclarity and intelligibility of speech,in these ways,the human doesn’t feel fatigued. So thestudy of speech enhancement technology is of great significance.For possible problems, such as white noise and music noise, etc, the paper introduces图the voice enhanced algorithm to deal with the influence of customer service noise. Based onthe speech processing, pretreatment, noise estimation and speech enhancement algorithm, theprocess of speech processing is understood. Then based on introducing the core of severalcommon speech enhancement algorithm, including the spectrum subtraction, wiener filteringmethod and minimum mean square error (MMSE) method, their respective advantages and shortcomings are analyzed. Via simulation, diverse analysis contrasts are shown directly bythe time domain waveform diagram and frequency domain waveform diagram.Through thecontrastive analysis of the simulation data, of the several speech enhancement algorithms, thespectral subtraction algorithm is simple, with low complexity, done more easily, maximummeeting the real-time requirements etc, so it has been widely used. But the traditional spectralsubtraction in the enhanced speech often accompanied by "music noise", and at the same time,a speech distortion. According to the defects of traditional spectral subtraction, this paperintroduces improved enhancement spectral subtraction algorithm which can dynamicallychange the parameters. Through the simulation test, the time domain waveform diagram andfrequency domain language spectra analysis verifies the correctness and rationality of thespectral subtraction. The simulation results show that the improved algorithm is designed toreduce the background noise and residual "music noise". Meanwhile it maintains speechintelligibility and improves the quality of speech.In this paper, response to the voice business of public security work, the system ofanswering alarm is designed, which is on the basis of business processes and businessrequirements. In this essay, the organization structure of public security system, networkstructure, main function module is designed and the voice quality problems are the focus ofthe discussion.Finally, combined with improved parameters dynamically changed speech enhancementalgorithm, answering alarm system is designed and implemented. Through the test applicationverification, the system can well improve the sound quality, makes the call more clearly andreduces the wrong fill information and better safeguards people’s property and life safety.