Analysis of Speech Enhancement Incorporating Speech Recognition

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 09 Issue: 09 | Sep 2022 www.irjet.net p-ISSN: 2395-0072

Analysis of Speech Enhancement Incorporating Speech Recognition

Dept. Of Electronics and Telecommunication Engineering, TPCT’s collage of engineering , Osmanabad, Maharashtra, India. ***

Abstract In this paper review work, we have tried to develop a novel way to improve the structure of a speech test during a speech, and to measure clear speech will be removedfromtheconfusingspeechlimit.Automaticspeech recognitionconsistsofasoundspeechspectrumusingFast FourierTransformaswellasathoughtfulreductioninthe volumeoftheaudiosoundfromtheaudiosound.Usingthe MATLAB software the audio reduction algorithm is developed to store audio speech data in half-time and calculate the spectrum of various sizes using Fourier Transform, reduce noise in audio speech, and recreate retrospectivespeechduringhighperformance.FastFourier (FFT).

Keywords –SpeechandvoiceEnhancement,Spectral noise Reduction, FFT, Sound Measurement, IFFT, Spectrograms.

Introduction

Acleanandclearspeechsignalislinkedtotheamountof soundinspeechdevelopment.Therearemanywaystomake a speech signal without a clamor signal. In many places Phones and cell phones get noisy air Domains like cars, airports, roads, trains, stations. Therefore, we attempt to eliminate the clamor signal by using a spectral reduction method. The main purpose of this paper is for real-time devicetoreduceordecreasebackgroundsoundwithavisual speechsignal,thisiscalledspeechdevelopment.Varietyof languages of speech are present, in that background noise loweringspeech.Applicationslikemobilecommunicationcan be learned a lot in recent years, speech improvement is required. The purpose of this speech development has to reversethenoisewhichinputfromanoisyspeech,likeasthe speechphaseoraccessibility.Itisoftenhardtoremovethe background sound without interrupting the speech, therefore,theexitofthespeechenhancementsystemisnot allowedbetweenspeechcontradictionandnoisereduction.

There may be other techniques same as to Wiener filtering,wavelet-based,dynamicfilteringandopticaloutput remainausefulmethod.Inordertoreducethespectral,we mustmeasuretheclamorspectrumandreduceitfromthe clamorous acoustic spectrum. Completely this approach, there are the following three scenarios to consider: sound adds speech signal and sound is not related to a single channel in the market.During thispaper,weattemptedto reduce the audio spectrum in order to improve distorted speech by using spectral output. We have described the

methodtestedintheactualspeechdataframeintheMATLAB area.ThesignalswereceivefromRealspeechsignalsarea website used for various tests. Then we suggest how to remove noise between the average noise level and the spectrumnoise.

Ingeneral,onlyonemediumsystemissetupbasedona varietyofspeechdataandunwantedscreamingthat,itworks indifficultsituationswherenopreviousclamorintelligenceis available.Genresoftenassumethatsoundisstablewhenever thespeechisalert.Theyusuallyallowfordisturbedsound during speech operations but in actual manner, when the soundisnotmoving,theperformanceofthespeechsignalsis greatlyincreased

I. WAYS TO REDUCE NOISE

A. How to Remove the Spectral

Manyadditionalvariationsofthevariousconclusionsare designedtoimprovespeech.Theoneweareusingisthought tobetheendofthephantomrelease.Thistypeworkswithin thescopeofthespreadandcreatesthehopethatthehelp cycle is said to be due to the phantom additon sound and clamor phantom. The action is specified within the image belowandcontainstwomaincomponents.

B. Convolutional Denoising Autoencoder (CDAE)

ConvolutionalDenoisingAutoencoder(CDAE)Promotes the same function of default encoders by adding convolutionalencodinganddecodinglayers.CDAEhasa2D layoutinthedialogandadjuststheinputtothe2Dalignment structure using the embedded space and the desktop as showninthepicture.Thus,inthestudyweproposed,Sofrom thismethodwecanproposethattheCDAEisbest suitedfor speechdevelopmentwhichis showninEq.

(3),onanyothermapincluded, hi=f(Wi*α+ai)(3) where, the process of conversion and the value of bias is Measuringthebackgroundofinthespectrum.

Reducetospectrumnoisefromthesoundspeechspectrum

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 09 Issue: 09 | Sep 2022 www.irjet.net p-ISSN: 2395-0072

Here,

Theproposedmethodofnoisereductionisprovidedfor visualremovalthatreliedonthereductionofthebackground sound level and the development of the speech signal. Evaluate the volume of sound enhanced by reduced visual acuityandthetypeofsoundreductionrecommended.

II. Sound spectrum measurement:

Fig1:SpectralSubtractionMethod

Assumingthatx(n)canbesoundandthemodifiedspeech signalisconsideredtobex(w)isasoundsignalspectrum willbe,N(w)thespectrumofratedintervalsandtherangeof speech processed by Y (w)and the pure speech signal the first is y (n). therefore, the processed spectrum as to the unpollutedspeechsignalwillbeprovidedasfollows:

Y(w)=X(w)-N(w)

Next Drawing Speech Block Impression By Spectral Removal.

Thespectrumsoundcannotbecalculatedinadvance,but it will be about the time for some instance speech is not present within the sound speech. When the people are speaking,oneshoulddefinitelypausetotakeadeepbreath. Wecanincludethesegapswithinspeechsignaltobalancethe background.

Wecancalculatetheaveragesizewhichiscalculatedfor anyframeinthelastfewseconds

III. Removing Noise Spectrum:

After removing, all data of spectral signal which is appear unlikelyin positivevalues Somechancesareavailablefor removingunwantedparts.Fourierconversions,usingsection segments directly from the Fourier conversion unit, and additionaladditionsaremadetoreconstructthespeechrate withinthetimezone.

The basic idea will be to reduce the noise from the given signali.e-inputsignal:

Y(w)=X(w)-N(w)

Suggestworkflowandsimulationresults

Figure2:Speechdevelopmentblockdiagram

Here,wecancreateacontinuousoutputsignal.Every frameis50%overwhelmed.FFT(FiniteFouriertransform) isusedineveryframe.

FFTsareoftenusefulindeterminingasignalhiddenin anoisytimezoneinpartsoffrequency.ThefunctionsY= FFT(x)andy=IFFT(X)suggestthealternatingrotation giventhevectorslengthNby

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 09 Issue: 09 | Sep 2022 www.irjet.net p-ISSN: 2395-0072

1. FLOWCHAR

Fig.3FlowChart

TheOverlapadd-onmethodisusedtointerruptandkeep thesignalsintosmallersegmentstomaketheprocesseasier. Thescatteringmethoddependson:(1)signaldecodeinbasic parts,(2)theprocedureofeverypart,(3)reassemblingthe cleared parts in an additional speech. In order to execute domain of frequency process to reduce visibility, it’s importanttodividetheconstantwisdom signalspeech into compact fragments which is called by frames. After completingprocedure,theframesarethenreassembled.

Theflowchartaboveshowstheproposedmethod,which involves collecting audio speech information and passing throughawindowtoextractexistingartobjectswithinthe ClamoroussignalandusetheFFTalgorithmthatdetectsthe phaseandmagnitudeoftheaudiosignal,duringwhichthe Technology focuses onthesize of the clamor to producea different refined speech signal. By using the type of visual reductionthenoiseismeasuredandreducedbythevalueof therequiredmagnitude.ThenusetheIFFTalgorithmandthe overlapaddstheprocesstowishforarefinedexpressionin thetimezone.

I. Simulation Results

Thefollowingtestsareusedusingscreaming:Computer, trainsound,andcaraudiousingthisAWGNsoundproduced. Everyframeis50%spacedFFTisapplied.Inparticular,the quality of speech that is consistent with the test of comprehension test is held to judge the quality of sound speechthatisimprovedbystandardvisualremovalandthe proposednoiseremovalmethod.Audio-typeaudiobasedon visualdeletionandspeechmodificationcreatedbyactivated

waveforms and spectrogram. Different sound level with soundspeechsignalisgeneratedusing3typesofshouting.

TheresultoftheaudiofilebuiltintoMATLAB

Fig 3: Results of speech enhancement: Noisefreespeech Noisyspeech EnhancedspeechusingSSmethod

3.1 Results of Spectrogram:

Fig(4)Cleanspeechsignalspectogram

Fig(5)Spectrogramofnoisyspeechsignal

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 09 Issue: 09 | Sep 2022 www.irjet.net p-ISSN: 2395-0072

III. Results of Spectrogram :

Fig(8)cleanspeechsignal

Fig(6)Enhancedspeechsignal

The above fig.6 shows the different colors of the spectrogram, while the red and yellow colors show two dissimilarenergylevel.Allofthesoundgiveninputsignals are full color of a very strong force due to the screaming patternassociatedwithbeinglargeyellow,redtomedium power,andblacktoasmallnumberwithalmostzeropower. Starting mainly from the effects at the exact time of the speechsignal,theeffectoftheshoutingcycle.Bycombining Fih.5withFig.6,theeffectastoclamorisreduced.

II. RESULTS OF DATABASE OF CAR NOISE AT 10DB

Fig 7: Results of speech enhancement: Cleanspeech Noisyspeech

EnhancedspeechbySSmethod

Fig(9)noisyspeechsignal

Fig(10)Enhancedspeechsignal

WhenmatchingFigure9withFigure10,thenoiseeffect decreased.Anotherinterestingtestoftheproposedalgorithm isseenincontrasttothespectrogramresultsfoundinFigure 9andhereanadditionalnumberofyellowlinesleftoutthe female speech signal suffering from high frequency. This provesthatawoman'svoicecanhaveaveryhighfrequency whenseparatedfromamaletone.

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 09 Issue: 09 | Sep 2022 www.irjet.net p-ISSN: 2395-0072

IV. RESULTS THAT WE GOT FROM TRAINNOISE AND COMPAREDWITHBACKGROUNDNOISE INDB USED FOR FEMALE AND MALE SPEECH SIGNALS

Signalsoffemale speech background Level(dB) Noise Levelin (dB) Reducedlevelof noise

TrainNoise(at0dB) 3.6448 -4.852 8.4968

TrainNoise(at5dB) -15.1722 -17.0717 1.8995

TrainNoise(at10dB) -25.5382 -41.2185 15.6803

Table-1:ReducedNoiselevelformale

Malespeechsignal NoiseLevelin (dB) NoiseLevelin (dB) NoiseReduced level

TrainNoise(at0dB) -3.9138 -6.2895 2.3757

TrainNoise(at5dB) -22.4948 -19.3090 3.1858

TrainNoise(at10dB) -26.1571 -31.5079 5.3508

Table-2:DecreasedNoiselevelformale

FromReducednoiselevelformale,Itwillbefound the amplitudestagesofmaleandfemalesoundgettinginputand outputdifferent.Thisindicatesthatithasasignificanteffect on sound speech. It promotes the clamor reduction phase. Thistypeofspectralreductioncreatesasignalofspeechbut also of shouting. Noise reduction was not properly guaranteedandcanbeguaranteedifseparatedbyinputand output.

TheClamorStageoftheinputandoutputspeechsignal andamplitudereductionupdatesaredifferentfromusingthe table.Table1showsthebackgroundstagedivisionforthe maleandfemalespeechsignalsite.

V. CONCLUSIONS

Inthispresentationpaperitishelpfultodevelopabetter spectralreductionalgorithm.Ithasrecentlybeenobservedin the fictional results that the suggested type significantly reduceslownoisecomparedtothealgorithmofmanyways toreducevisibility.Thistypeofspeechcausesthesignaltobe visible but is accompanied by a negative screaming. The soundisnotverylowanditisverylowcomparedtoinput andoutput.Theseforceswillalsobeadjustedandexpanded toaccommodatestaticnoise.Fromthistypeofdomaindesign wehavereached70%andthatwecanbuildthissystemfor embedded organizations that are subject to speech processing or communication purpose. For better comparisons, we showed as results and therefore

spectrogramsastoclear,soundandpreparedspeech.Delete thesite-preparedspeechsentence.

REFERENCES

[1] S. Kamath, and P. Loizou, A multi-band spectral subtraction method for enhancing speech corrupted by colorednoise,ProceedingsofICASSP-2002,Orlando,FL,May 2002. R. Nicole, “Title of paper with only first word capitalized,”J.NameStand.Abbrev.,inpress.

[2]S.F.Boll,Suppressionofacousticnoiseinspeechusing spectralsubtraction,IEEETrans.onAcoust.Speech&Signal Processing,Vol.ASSP-27,April1979,113-120.

[3]M.Berouti,R.Schwartz,andJ.Makhoul,Enhancementof speech corrupted by acoustic noise, Proc. IEEE ICASSP , WashingtonDC,April1979,208-211.

[4]H.Sameti,H.Sheikhzadeh,LiDeng,R.L.Brennan,HMMBased Strategies for Improvement of Speech Signals Embedded in Non-stationary Noise, IEEE Transactions on SpeechandAudioProcessing,Vol.6,No.5,September1998. International Research Journal of Engineering and Technology(IRJET)e-ISSN:2395-0056

[5] Yasser Ghanbari, Mohammad Reza Karami-Mollaei, Behnam Ameritard “Improved Multi-Band Spectral Subtraction Method For Speech Enhancement” IEEE InternationalConferenceOnSignalAndImageProcessing, august2004.

[6]Pinki,SahilGupta“SpeechEnhancementusingSpectral Subtraction-typeAlgorithms”IEEEInternationalJournalof Engineering,Oct2015.

[7]S.Kamath,Amulti-bandspectralsubtractionmethodfor speech enhancement, MSc Thesis in Electrical Eng., UniversityofTexasatDallas,December2001.

[8] Ganga Prasad, Surendar “A Review of Different ApproachesofSpectralSubtractionAlgorithmsforSpeech Enhancement”CurrentResearchinEngineering,Scienceand Technology(CREST)Journals,Vol01|Issue02|April2013| 57-64.

[9]LalchhandamiandRajatGupta“DifferentApproachesof Spectral Subtraction Method for Speech Enhancement” InternationalJournalofMathematicalSciences,Technology andHumanities95(2013)1056–1062ISSN22495460.

[10]EkaterinaVerteletskaya,BorisSimak“Enhancedspectral subtractionmethodfornoisereductionwithminimalspeech distortion”IWSSIP2010-17th InternationalConferenceon Systems,SignalsandImageProcessing.

[11]D.Deepa,A.Shanmugam“SpectralSubtractionMethod ofSpeechEnhancementusingAdaptiveEstimationofNoise

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 09 Issue: 09 | Sep 2022 www.irjet.net p-ISSN: 2395-0072

with PDE method as a preprocessing technique” ICTACT JournalOfCommunicationTechnology,March2010,Issue: 01.

[12]M.Berouti,R.Schwartz,&J.Makhoul,“Enhancementof SpeechCorruptedbyAcousticNoise,”Proc.ICASSP,pp.208211,1979.

[13] Prince Priya Malla, Amrit Mukherjee, Sidheswar Routary,GPalai“Designandanalysisofdirectionofarrival using hybrid expectation-maximization and MUSIC for wireless communication” IJLEO International Journal for LightandElectronOptics,Vol.170,October2018.

[14] Pavan Paikrao, Amit Mukherjee, Deepak kumar Jain “Smart emotion recognition framework: A secured IOVT perspective”, IEEE Consumer Electronics Society, March 2021.

[15] Amrit Mukhrerjee, Pratik Goswani, Lixia Yang “DAI basedwirelesssensornetworkformultimediaapplications” MultimediaTools&Applications, May2021,Vol.80Issue11, p16619-16633.15p.

[16] Sweta Alpana, Sagarika Choudhary, Amrit Mukhrjee, AmlanDatta“Analysisofdifferenterrorcorrectingmethods for high data rates cognitive radio applications based on energydetectiontechnique”,JournalofAdvancedResearch inDynamicandControlSystems,Volume:9|Issue:4.

© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page894