資源描述:
《自適應(yīng)噪聲對(duì)消在語(yǔ)音信號(hào)處理中的應(yīng)用研究 畢業(yè)論文》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。
1、本科生畢業(yè)論文(設(shè)計(jì))中文題目自適應(yīng)噪聲對(duì)消在語(yǔ)音信號(hào)處理中的應(yīng)用研究英文題目Adaptivenoisecancellationinspeechsignalprocessingresearch學(xué)生姓名班級(jí)650603班學(xué)號(hào)學(xué)院儀器科學(xué)與電氣工程學(xué)院專業(yè)測(cè)控技術(shù)與儀器指導(dǎo)教師職稱摘要在實(shí)際生活中,任何語(yǔ)音信號(hào)都不可避免的受到噪聲信號(hào)的影響,如何有效的抑制和去除噪聲,提高語(yǔ)音的可懂度是近年來(lái)的熱門(mén)研究課題,文中介紹了自適應(yīng)濾波器的基本原理,結(jié)構(gòu)和應(yīng)用,應(yīng)用matlab軟件,對(duì)自適應(yīng)算法在噪聲對(duì)消中的應(yīng)用進(jìn)行了
2、仿真研究,并完成了語(yǔ)音信號(hào)噪聲消除實(shí)例。本文對(duì)自適應(yīng)濾波算法在語(yǔ)音信號(hào)去噪方面進(jìn)行了研究,對(duì)自適應(yīng)對(duì)消系統(tǒng)進(jìn)行了深入的學(xué)習(xí)與研究,在固定步長(zhǎng)的基礎(chǔ)上,建立了步長(zhǎng)因子u與信噪比及噪聲幅度之間的一種非線性函數(shù)關(guān)系,使步長(zhǎng)隨誤差信號(hào)e(n)的變化而變化,從而提高了收斂速度,能夠有效的濾除實(shí)際生活中語(yǔ)音信號(hào)中的高斯白噪聲,工頻干擾,以及其他講話者的干擾,大幅度提高輸出語(yǔ)音信號(hào)的信噪比,有效的提高語(yǔ)音的可懂度。通過(guò)實(shí)驗(yàn)證明,該算法在收斂速度,消噪性能,信噪比提高方面與常規(guī)的自適應(yīng)算法相比均有一定的提高。關(guān)鍵詞:自適
3、應(yīng)濾波變步長(zhǎng)LMS算法語(yǔ)音降噪AdaptivenoisecancellationinspeechsignalprocessingresearchAbstract:Inourdailylife,allspeechsignalwillbeinfluencedbynoise,Howtoeffectivelyeliminatethenoiseisoneofhotsubjectsforyears.Thepaperbeginswiththeprincipleofadaptivefilter,structureanda
4、pplication.BasedontheMATLABplatform,simulationiscarriedoutfortheapplicationsofadaptivealgorithmsinnoisecancelling,andcompletedtheinstanceofvoicesignalnoisereduction.Thispaperdiscussaboutadaptivefilteringalgorithminthespeechsignaldenoisingaspectsoftheresea
5、rchontheadaptivecancellationsystemin-depthstudyandresearchandestablishesanotherstepfactoruandtheerrorsignale(n)betweenthenon-linearfunctionofanewrelationship,thealgorithmusingvariablestepsize,thestepwiththemagnitudeofthenoisesignaltonoiseratioandthechange
6、,toimprovetheconvergencerate,caneffectivelyfilteroutinreallifespeechsignalGaussianwhitenoise,frequencyinterference,andinterferencewithotherspeakers.Cangreatlyenhancetheoutputspeechsignaltonoiseratio,experimentsshowthatthealgorithmconvergencerate,noiseredu
7、ctionperformance,improvingsignaltonoiseratiowiththeconventionaladaptivealgorithmshaveimprovedtosomeextentcompared.Inthispaper,thecoefficientoftheformulafittinganalysis,andfindtheoptimalsolution,theeffectofnoisecorrelationcoefficienthasdonesomeanalysisshow
8、sthatthemainchanneloftheinputnoiseandthenoisereferencechannelcorrelationcoefficienthigherlearning,filterthebetter.Keywords:AdaptiveFiltering;variablestepsize;LMSalgorithm;SpeechEnhancement目錄1緒論…………………………………………………………