音频信息隐藏算法的研究与实现-翻译.docx

音频信息隐藏算法的研究与实现-翻译.docx

《音频信息隐藏算法的研究与实现-翻译.docx》由会员分享，可在线阅读，更多相关《音频信息隐藏算法的研究与实现-翻译.docx（17页珍藏版）》请在冰豆网上搜索。

苏州大学本科生毕业设计（论文）附件:

外文文献资料与中文翻译稿

外文文献资料

收集：

苏州大学应用技术学院 11电子班（学号1116405046）汤珏

MULTI-CHANNELAUDIOINFORMATIONHIDINGABSTRACT

Weconsideramethodofhidingmanyaudiochannelsinonehostsignal.Thepurposeofthisistoprovidea‘mix’thatincorporatesinformationonallthechannelsusedtoproduceit,therebyallowingall,or,atleastsomechannelstobestoredinthemixforlateruse.Afterprovidinganoverviewofsomerecentlypublishedaudiowatermarkingschemesinthetimeandtransformdomains,wepresentamethodthatisbasedonusingafourleastsignificantbitsschemetoembedfiveMP3filesintoasingle16-bithoststereoWAVfilewithoutincurringanyperceptualaudiodistortionsinthehostdata.ThehostWAVfileistakentobethefinalmixassociatedwiththeoriginaldatabeforeapplying‘loseMP3’compressionoralternativelyanarbitraryhostaudiosignalintowhichothermultichannelaudiodataishidden.Further,theembeddedinformationcanbeencryptedand/ortheembeddinglocationsrandomizedonachannelbychannelbasisdependingontheprotocoldesiredbytheuser.Themethodisillustratedbyprovidingexamplem-codeforinterestedreaderstoinvestigateandreproducetheresultsobtainedtodateandasabasisforfurtherdevelopment.

1.INTRODUCTION

Theapproachtoinformationhidingdiscussedinthispaperrelatestoaninvestigationintothedesignof‘intelligent’codingalgorithmsforaudiopost-productionbasedonacurrentFP7researchproject.TheresearchhasandiscontinuingtobeundertakenjointlybetweentheAudioResearchGroup,theDigitalMediaCentreatDublinInstituteofTechnologyandTamborineProductionsLimited.Thisincludestheuseof

第0页

watermarkingmethodsforaudiosignalauthenticationandmultipleaudiosignalembeddingtechniquesusingcomputationalsolutionsbasedonimageinformationhidingandauthenticationtechniques,e.g.

Whilethereareanumberofdifferentapproachesforhidingasignalinanotherhostsignal,inclusiveofencryption,orotherwise,therearerelativelyfewpublishedmethodsthatconsideramulti-signaltosinglehostsignalapproach.InthispaperweconsidertheproblemofembeddingmanyaudiosignalsprovideasinMP3formatsintoasinglewaveofWAVfileunderthefollowingconditions:

（i）theembeddedMP3audiodatacanberecoveredwithoutanylossofinformation;（ii）thehostsignalhasnoperceptualchangeinitsaudiofidelity.AnalgorithmispresentedbasedonatimedomainmethodinwhichfiveMP3filesarehiddeninasingleWAVfileusingafourleastsignificantbitschemewhichsatisfiestheseconditions.

Themethodconsideredhasapplicationsincovertinformationtransmissionwithapplicationsinareassuchascommunicationssecurity,signalauthenticationandDRM（DigitalRightsManagement）,forexample.However,themethodreportedhasbeendevelopedinthecontextofanaudiopost-productionproducttoallowsastereo‘mix’toincorporateinformationonallorsomeofthechannelsusedtoproducethesamemix,therebyallowingallorselectivechannelstobeembeddedinthemixforremixingatalaterdateand/orfordatastorage.Currentscenariosusuallyinvolvesapost-productioncentrehavingtomaintainrecordsofallthedatausedtogenerateamixwiththeviewthatthissamedatamayberequiredatalaterdate,therebynecessitatingtheneedforadatabasemanagementinfrastructurewhichmayrequiresignificantoverheads.

Afterprovidinganoverviewofrecentlypublishedworksthatfocus

ontheareaofaudioinformationhiding（Section2）,wepresentatimedomainsolutiontotheprobleminSection3.Thisincludesprototypem-codeforinterestedreaderstotestthealgorithmsforthemselvesandasa

第16页

basisfordevelopingthemethodologyfurther.

2.AUDIOINFORMATIONHIDING

AudioinformationhidingorAudioSteganographyisbasedontwospecificanddistinctapproacheswhichconsiderthedevelopmentofalgorithmsusingtimedomaindataortransformdomaindatasuchastheconsinetransform,forexample.Inthissection,weprovideabriefoverviewofsomeofthemostrecentlypublishedmaterialinthisarea,publicationsthathavebeenstudiedbytheauthor’sinassociationwiththemulti-channelinformationhidingapproachconsideredinthispaper.

2.1.AudioHidingintheTimeDomain

Timedomainaudiohidinghastheadvantageofbeingrelativesimpletoimplementandcomputationallyefficient.Howevertheapproachdoesnotyieldalgorithmsthatareasdiverseandrobustasthosedevelopedforthetransformdomain.Thisisbecauseoftherelativelylimitednumberofwaysinwhichdatacanbemanipulatedinthetimedomainsubjecttothegenerationofanoutputthatisperceptuallycompatiblewiththeoriginalaudiosignal.Intheliteraturesurveythatfollows,itisclearthatallthemethodsconceivedarebasedonsomeformofdatamanipulationb