ISO_IEC_1449614_20031115

ISO/EC14496-14:2003(E Contents Page Foreword Introduction 0.1 Derivation ...a vvvv 0.2nterchange...…. 0.3Contentcreation ∴V 0.4Streamedpresentation…….….….….….….….….….….…yi Scope..... Normativereferences 3 StorageofMPEG~4.…,… 3.1ElementaryStreamTracks….,,…,… 日面日 3.2 TrackIdentifiers 3.3Synchronizationofstreams 3.4Composition… 3.5HandlingofFlexMux…..........….……… Fileidentification Additionstothebasemediaformat 5.1ObjectDescriptorBox 345566778 5.2TrackReferenceTypes 5.3Trackheaderbox 5.4HandlerReferenceTypes 5.5 MPEG-4Mediaheaderboxes 5.6SampleDescriptionBoxes 5.7DegradationPriorityValues 10 Templatefieldsused.... ∴10 AnnexA(informative)Patentstatements…… oISO/EC2003-Allrightsreserved ISO/EC14496-14:2003(E) Foreword iso(theInternationalOrganizationforStandardizationandlEc(theInternationalElectrotechnical Commission)formthespecializedsystemforworldwidestandardization.Nationalbodiesthataremembersof ISoorIEcparticipateinthedevelopmentofInternationalStandardsthroughtechnicalcommittees establishedbytherespectiveorganizationtodealwithparticularfieldsoftechnicalactivity.ISOandIEC technicalcommitteescollaborateinfieldsofmutualinterest.Otherinternationalorganizations,governmental andnon-governmental,inliaisonwithISoandIEC,alsotakepartinthework.Inthefieldofinformation technology:ISOandIEChaveestablishedajointtechnicalcommittee,ISO/ECJTC1 InternationalStandardsaredraftedinaccordancewiththerulesgivenintheISo/ECDirectives,Part2 ThemaintaskofthejointtechnicalcommitteeistoprepareInternationalStandards.DraftInternational Standardsadoptedbythejointtechnicalcommitteearecirculatedtonationalbodiesforvoting.Publicationas anInternationalStandardrequiresapprovalbyatleast75ofthenationalbodiescastingavote So/EC14496-14waspreparedbyJointTechnicalCommitteeISO/ECJTC1,Informationtechnology, SubcommitteeSC29,Codingofaudio,picture,multimediaandhypermediainformation ISo/EC14496consistsofthefollowingparts,underthegeneraltitleInformationtechnology-Codingof audio-visualobjects Part1:Systems Part2:Visual Part3:Audio Part4:Conformancetesting Part5:Referencesoftware Part6:DeliveryMultimediaIntegrationFramework(DMIF Part7:Optimizedreferencesoftwareforcodingofaudio-visualobjects Part8:Carriageof/SO/EC14496contentsover/Pnetworks Part9:Referencehardwaredescription Part10:AdvancedvideoCoding(AVC) Part11:Scenedescriptionandapplicationengine Part12/sobasemediafileformat Part13:IntellectualPropertyManagementandProtection(IPMP)extensions Part14:MP4fileformat Part15:AdvancedVideoCoding(Avc)fileformat Part16:AnimationFrameworkeXtension(AFX OISo/EC2003-Allrightsreserved ISO/EC14496-14:2003(E Introduction 0.1Derivation ThisspecificationdefinesMP4asaninstanceoftheISoMediaFileformat[Iso/EC14496-12andISonEC 15444-12 ThegeneralnatureoftheIsoMediaFileformatisfullyexercisedbyMP4.MPEG-4presentationscanbe highlydynamic,andthereisaninfrastructure-theobjectDescriptorFramework-whichservestomanage theobjectsandstreamsinapresentation.AnInitialObjectDescriptorservesasthestartingpointforthis framework.IntheusagemodesdocumentedintheISoMediaFile,anInitialobjectDescriptorwouldnormall bepresent,asshowninthefollowingdiagrams 0.2Interchange Thefollowingdiagramgivesanexampleofasimpleinterchangefile,containingtwostreams Figure1-Simpleinterchangefile 0.3Contentcreation nthefollowingdiagram,asetoffilesbeingusedintheprocessofcontentcreationisshown oISO/EC2003-Allrightsreserved ISO/EC14496-14:2003(E) Figure2-ContentCreationFile 0.4Streamedpresentation Thefollowingdiagramshowsapresentationpreparedforstreamingoveramultiplexingprotocol,onlyonehint trackisrequired Figure3-HintedPresentationforStreaming OISo/EC2003-Allrightsreserved INTERNATIONALSTANDARD ISO/EC14496-14:2003(E) Informationtechnology-Codingofaudio-visualobjects Part14 MP4fileformat 1Scope ThisInternationalstandarddefinesthemp4fileformatasderivedfromtheisobasemediafileformat 2Normativereferences Thefollowingreferenceddocumentsareindispensablefortheapplicationofthisdocument.Fordated references,onlytheeditioncitedapplies.Forundatedreferences,thelatesteditionofthereferenced document(includinganyamendments)applies ISO/EC14496-1:2001,Informationtechnology-Codingofaudio-visualobjects-Part1:Systems ISO/EC14496-12:Informationtechnology-Codingofaudio-visualobjects-Part12:/sobasemediafile format(technicallyidenticaltoISO/EC15444-12) 3StorageofMPEG-4 3.1ElementaryStreamTracks 3.1.1ElementaryStreamData Tomaintainthegoalsofstreamingprotocolindependence,themediadataisstoredinitsmost'naturalformat andnotfragmented.Thisenableseasylocalmanipulationofthemediadata.Thereforemedia-dataisstored asaccessunits,arangeofcontiguousbytesforeachaccessunit(asingleaccessunitisthedefinitionofa sampleforanMPEG-4mediastream).Thisgreatlyfacilitatesthefragmentationprocessusedinhinttracks Thefileformatcandescribeandusemediadatastoredinotherfiles,howeverthisrestrictionstillapplies Thereforeifafileistobeusedwhichcontains'pre-fragmented'mediadata(egaFlexMuxstreamondisc themediadatawillneedtobecopiedtore-formtheaccessunits,inordertoimportthedataintothisfile format Thisistrueforallstreamtypesinthisspecification,includingsuch'meta-informationstreamsasObject DescriptorandtheClockReference.Theconsequencesofthisare,onthepositiveside,thatthefileformat treatsallstreamsequally;onthenegativeside,thismeansthatthereare'internalcross-linksbetweenthe streams.Thismeansthataddingandremovingstreamsfromapresentationwillinvolvemorethanaddingor deletingthetrackanditsassociatedmedia-data.Notonlymustthestreambeplacedinorremovedfrom,the scene,butalsotheobjectdescriptorstreammayneedupdating Foreachtrack,theentireES-descriptorisstoredasthesampledescriptionordescriptions.The SLConfigDescriptorforthemediatrackshallbestoredinthefileusingadefaultvalue(predefined=2),excep whentheElementaryStreamdescriptorreferstoastreamthroughaurl,i.e.thereferredstreamisoutside thescopeoftheMP4file.InthatcasetheSLConfigDescriptorisnotconstrainedtothispredefinedvalue oISO/EC2003-Allrightsreserved ISO/EC14496-14:2003(E) Inatransmittedbit-streamtheaccessunitsintheslpacketsaretransmittedonbyteboundaries.this meansthathinttrackswillconstructSLPacketheadersusingtheinformationinthemediatracks,andthehint trackswillreferencetheaccessunitsfromthemediatrack.Theplacementoftheheaderduringhintingis possiblewithoutbitshifting,aseachSLPacketandcorrespondingcontainedaccessunitwillbothstarton byteboundaries 3.1.2ElementaryStreamDescriptors TheESDescriptorforastreamwithinthescopeofthemP4fileasdescribedinthisdocumentisstoredinthe sampledescriptionandthefieldsandincludedstructuresarerestrictedasfollows EsID-settooasstoredwhenbuiltintoastreamthelower16bitsofthetrackidareused treamDependenceFlagsetto0asstored;ifadependencyexists,itisindicatedusingatrack referenceoftype'dpnd URLflag-keptuntouched,i.esettofalse,asthestreamisinthefile,notremote SLConfigDescriptor-ispredefinedtype2 OCRStreamFlag-settofalseinthefile TheESDescriptorforastreamreferencedthroughanESURLisstoredinthesampledescriptionandthe fieldsandincludedstructuresarerestrictedasfollows ESID-settooasstoredwhenbuiltintoastreamthelower16bitsofthetrackidareused streamDependenceFlag-settoOasstored;ifadependencyexists,itisindicatedusingatrack referenceoftype'dpnd URLflag-keptuntouched,i.e.settotrue,asthestreamisnotinthefile SLConfigDescriptor-keptuntouched OCRStreamFlagsettofalseinthefil lotethattheQoSDescriptoralsomayneedre-writingfortransmissionasitcontainsinformationaboutPDU sizesetc 3.1.3ObjectDescriptors TheinitialobjectdescriptorandobjectdescriptorstreamsarehandledspeciallywithinthefileformatObject descriptorscontainESdescriptors,whichinturncontainstreamspecificinformation.Inaddition,tofacilitate editing,theinformationaboutatrackisstoredasanESDescriptorinthesampledescriptionwithinthattrack Itmustbetakenfromthere,re-writtenasappropriate,andtransmittedaspartoftheODstreamwhenthe presentationisstreamed Asaconsequence,ESdescriptorsarenotstoredwithintheODtrackorinitialobjectdescriptor.Instead,the initialobjectdescriptorhasadescriptorusedonlyinthefile,containingsolelythetrackIDoftheelementary stream.Whenused,anappropriatelyre-writtenESDescriptorfromthereferencedtrackreplacesthis descriptorLikewise,odtracksarelinkedtoestracksbytrackreferencesWhereanesdescriptorwouldbe usedwithintheODtrack,anotherdescriptorisused,whichagainoccursonlyinthefile.Itcontainstheindex ntothesetofmpodtrackreferencesthatthisOdtrackownsAsuitablyre-writtenESDescriptorreplacesitby hehintingofthistrack OISo/EC2003-Allrightsreserved ISO/EC14496-14:2003(E TheEsIDIncisusedintheobjectDescriptorBo classESIDIncextendsBaseDescriptorbit(8)tagESIDIncTagf nsignedint(32)TrackID;/IDofthetracktouse 三S工D工ncTag=0x0Eisreservedforfi1eformatusage, Theesidrefisusedintheodstream classEsIdRefextendsBaseDescriptorbit(8)tag=ESIDRefTagI bit(16)refindex;/trackef.indexofthetracktouse SIDRefTag=0xoFisreservedforfileformatusage MP4IODTag-0xl0isreservedforfileformatusage MP4ODTag-0x11isreservedforfileformatusage IPIDescrpointerRefrag=0x12isreservedforfileformatusage EsDescrRemcveRefrag=0x07isreservedforfileformatusage(commandtag) NOTETheabovetagvaluesaredefinedin82.2.2Table1and8.2.3.2Table2oftheMPEG-4SystemsSpecification andtheactualvaluesshouldbereferencedfromthosetables ahintermayneedtosendmoreODeventsthanactuallyoccurintheOdtrack:forexample,ifthe EsdescriptionchangesatatimewhenthereisnoeventintheOdtrack.Ingeneral,anyODeventsexplicitly authoredintotheodtrackshouldbesentalongwiththosenecessarytoindicateotherchanges.TheES descriptorsentintheOdtrackistakenfromthedescriptionofthetemporallynextsampleintheEstrack(in decodingtime 3.2TrackIdentifiers Thetrackidentifiersusedinanmp4fileareuniquewithinthatfile;notwotracksmayusethesameidentifier. Eachelementarystreaminthefileisstoredasamediatrack.Inthecaseofanelementarystream,thelower wobytesofthefour-bytetrackIdshallbesettotheelementarystreamidentifier(ES_ID).theuppertwo bytesofthetrack_IDarezerointhiscase.Hinttracksmayusetrackidentifiervaluesinthesamerange,ifthis numberspaceisadequate(whichitgenerallyis).However,hinttrackidentifiersmayalsouselargervaluesof trackidentifier,astheiridentifiersarenotmappedtoelementarystreamidentifiers.Thusverylarge presentationsmayusetheentire16-bitnumberspaceforelementarystreamidentifiers Thenexttrackidentifiervaluefoundinnexttrackidinthemoviebox,asdefinedintheisobase MediaFormat,generallycontainsavalueonegreaterthanthelargesttrackidentifiervaluefoundinthefile Thisenableseasygenerationofatrackidentifierundermostcircumstances.However,ifthisvalueisequalto orlargerthan65535,andanewmediatrackistobeadded,thenasearchmustbemadeinthefileforafree trackidentifier.Ifthevalueisall1s(32-bitmaxint)thenthissearchisneededforalladditions Ifitisdesiredtoaddatrackwithaknowntrackidentifier(elementarystreamidentifier)thenthefilemustbe searchedtoensurethatthereisnoconflict.Notethathinttrackscanbere-numberedfairlyeasilywhilemore careshouldbetakenwithmediatracksastheremaybereferencestotheiresid(trackiD)inothertracks IfhinttrackshavetrackIDsoutsidetheallowedrangeforelementarystreamtrackS,thennexttrackID documentsthenextavailablehinttrackID.Sincethisislargerthan65535,asearchwillthenalwaysbe neededtofindavalidelementarystreamtrackiD Iftwopresentationsaremerged,thentheremaybeconflictbetweentheirtrackIDs.Inthatcase,oneormore trackswillhavetobere-numberedtherearetwoactionstobetakenhere Changingtheidofthetrackitself,whichiseasy(trackIDinthetrackheader) Changingpointerstoit oISO/EC2003-Allrightsreserved ISO/EC14496-14:2003(E) Thepointersmayonlyoccurinthefileformatstructureitself.thefileformatusestrackIDsonlythroughtrack references,whichareeasilyfoundandmodified.TrackIDsbecomeEsIDsintheMPEG-4data,andEs_IDs occurwithintheODStream.SinceallpointerstoESIDsintheOdstreamarereplacedbymeansoftrack references,thereisnoneedtoinspecttheOdstreamforcross-referenceswithinMPEG-4streams Inthefileformat,ESDescriptorRemovecommandandIPlDescrPointerdescriptorareconvertedto ESDescrRemoveRefandIPlDescrPointerRefby changingthetagvaluetoESDescrRemoveRefTagorIPIDescrPointerRefTagrespectively; changinganyEs_idtotheappropriatetrackreferenceindex(usingreferencesoftypempodandipr respectively-see5.2) Whenhintingorserving,thetagvalueandtrackreferenceindexchangesshallbereversed 3.3Synchronizationofstreams Intheabsenceofexplicitdeclarationstothecontrary,tracks(streams)comingfromthesamefileshallbe presentedsynchronized.Thismeansthathintersand/orserversmusteitherpickoneofthestreamstoserve astheocrsourcefortheothersoraddanocrstreamtoassociateallthestreamswithittrackreferences oftype'sync'maybeusedinthefiletodefeatthedefaultbehavior.InMPEG-4theOCRStreamFlagand OCRESIDfieldsintheESDescriptorgovernthesynchronizationrelationships.ThemappingofMP4 structuresintothosefieldsshallobeythefollowingrules TheMPEG-4ESDescriptor,asstoredinthefile,usuallycontainsOCRStreamFlagsettoFALSE,and noOCRESID.IfanOCRESIDisset,itisignored Ifatrack(stream)containsatrackreferenceoftype'sync'whosevalueis0,thenthehinterorserver shallsettheOCRStreamFlagfieldintheMPEG-4ESDescriptortoFaLsEandshallnotinsertany OCREsIDfield.Thismeansthatthisstreamisnotsynchronizedtoanotherbutotherstreamsmay besynchronizedtoit Ifatrack(stream)containsatrackreferenceoftype'syncwhosevalueisnot0,thenthehinteror servershallsettheOCRStreamFlagfieldintheMPEG-4ESDescriptortoTRUEandshallinsertan OCRESIDfieldwiththesamevaluecontainedinthe'trackreference.Thismeansthatthis streamissynchronizedtothestreamindicatedintheocrEsID.otherstreamsmayalsobe synchronizedtothesamestream,eitherexplicitlyorimplicitI Ifatrack(stream)doesnotcontainatrackreferenceoftype'sync,thenthedefaultbehaviorapplies ThehinterorservershallsettheOCRStreamFlagfieldintheMPEG-4ESDescriptortoTRUEand shallinsertanocresidfieldwithavalueselectedbasedontherulesbelowthismeansthatthis streamissynchronizedtothestreamindicatedintheOCRESID.Therulesforselectingthe OCRESDareasfollows Ifnotrack(stream)inthefile trackreferenceoftype'sync,thenthehinterpicksone TrackldandusesthatvaluefortheoCrESIdfieldofallESDescriptors.Thereisonepossible exceptionwheretheEsDescriptorofthestreamwhichcorrespondstothattrackld,forwhichthe OCRStreamFlagmaybesettoFALSE oIfoneormoretracks(streams)inthefilecontainatrackreferenceoftypesync,andallsuchtrack referencesindicateconsistentlyasingleTrackld,thenthehinterusesthatTrackld.Inatrack referenceoftype'sync'thevalue0isequivalenttotheTrackldofthetrackitself oIftwoormoretracks(streams)inthefilecontainatrackreferenceoftype'sync,andsuchtrack referencesdonotindicateasingletrackld,thenthehintercannotmakeadeterministicselection andthebehaviorisundefined.Inatrackreferenceoftype'sync'thevalue0isequivalenttothe Trackldofthetrackitself OISo/EC2003-Allrightsreserved