`
`STANDARD
`
`ISOIIEC
`
`11172-2
`
`First edition
`1 993-08-01
`
`
`
`lnfonnation technology — Coding of
`moving pictures and associated audio for
`digital storage media at up to about
`1,5 Mbitls —
`
`Part 2:
`
`Video
`
`Technobgfes de finfo/mation — Codags de l'image animée et du son
`associé pour Ies supports 09 stcckaga numérique jusqu'3 environ
`7.5 Mbit/s —
`Pattie 2- Vidéo
`
`
`
`.
`
`Refeience number
`ISO/|EC11172—21993{E)
`
`Page 1 of 124
`
`Samsung Exhibit 1004
`
`
`
`ISD/IEO 11172-2: 1993 (E)
`
`Contents
`
`Page
`
`1.:
`
`L2
`
`Normative tduances..........;................................................................,l
`
`Secliun 2: Technical elumnIs._......-..-....-....-_.............._.._.....-_-........-.........3
`
`2.] Dcfinitiou....., ...........................................
`
`.................... ..3
`
`2.1
`
`Symbolrmd I.lIbravrnlions......
`
`II
`
`2.3 Melhod of ducxibing bitsu-cum syntax......
`
`.................................. ..... I3
`
`2.1 R=quiI=ucnt:...-.-.......-..-............-..................................................... I5
`
`Annexes
`
`A
`
`B
`
`C
`
`8 by8 Imcno disctclc cosinemmfotm
`
`Vnriabb lcngtboocle table: ................................................................40
`
`Video |ufl'etingvEif1er............ .-.. ...................
`
`D‘
`
`Guidebencodingvidao ......................................
`
`..........................S1
`
`E
`
`F
`
`Bibliography
`
`List of parent holders................
`
`O ISO/EC 1993
`All rights mservad. No put of Ibis publication may be mpmduaad or utilized in any form or by
`any meals. electronic or mechanical. includlngplntuoopying and microfilm without
`pennissian in wtiting from the publisher.
`
`ISOIIEC Copyright Office -Cue Pasule 56 - Cl-11211 Geneva 20 - Switzerland
`
`Ptinlzd in Swiutlcnd.
`
`Padé 2 of 124
`
`
`
`O I$G'lEC
`
`ISUIEC 11172-2: 19925 (E)
`
`Foreword
`
`ISO Ithe |ntema1:ionalO'ganizalion for Standardization) and IEC (the Inter-
`nationaf Electtotechnbd Commission! farm the specialized system for
`worldwide standardization. National bodies that are members of ISO or
`IEC participate it the development of lmematicnal Standards through
`technical committees established" by the respective organization to deal
`with particular fields oi techncal activity. ISO and IEC technical com-
`mittees collaborate in fields of mutual interest. Other international organ-
`izaticns, govemrnentd md non-gcvemmentnl, in liaison with ISO am IEC,
`also take part ‘n the work.
`'
`'
`
`In the field of infonnazion technology. ISO and IEC have establisheda ‘pint
`technical committee, '50/IEC JTC 1. Daft lmematbnal Standards adopted
`by the ioint technil committee are circulated to national bodes for vot-
`ing. Publication as an International Stmdard raquites-appxwal by at least
`75% o‘ the national bodies casting a vote.
`
`lmemationol Standard ISO/IEC 11172-2 was prepared by Joint Technical
`Committee lso/IEC J1’C 1, Informatian teclfindogy. $u_b-Commitee SC 29.
`Coded n3prasantatian"o.' audo, picture, mdtimedi: and hypermedia infcr-_
`matidn.
`'
`‘
`
`asorsc 11172 consists of the following parts. unzlar the‘general title In-
`formation technology —' Coding of moving pictures and associated audb
`for digitel storage media at up to about 1,6 Mbit/st
`
`— Part 1: Systems
`
`— Part 2: ‘Video
`
`— Pan 3: Audio
`
`— Part 4: Compliance testing
`
`Annexes A. B and C fawn" an integral part of this pa-t of ISOIIEC 11172.
`Annexes D. E and F are for information only.
`
`Page 3 of 124
`
`
`
`nsouec 11112-2: 1993 (E)
`
`Isonec
`
`Introduction
`
`Not: --
`
`prooonl to am-.xD,bctore remrnlnx uoclaurcs 1 and 2.
`
`imnrcsted in an overview ofthe MPEG Video layer sholldteadlhis Introduction and than
`
`0.1
`
`Purpose
`
`This partofISOI1EC l1172was developed in rcsponsclolhc gnawing Iced torn conxrnonfomm for
`representing omnpressed videoon variousdigital storage mediasuch as CDs,DAT§, Windeserdisks and
`optital dives. 'l‘hispm of ISOILEC 11172 spccifxesa oodedrepresenlation matcm be used ior
`compressing videosecpenccslobitrauzsarumd 1.5 Mlvitls. The uscoi'thispartofISOllEC11172. means
`Ihntmotim videocal bc nanipulaled as afonnofcnmputerdau andcan he Irnmmimdand modved over
`calming mdfuure networks. The coded reueaznalionan he used with bun 625-line and 525-line
`tclcviion and provides flexibility fa use with workstation and personal computer displays.
`
`This pnrlofISOlIEC 1] l72w:Ls developed In operal: principally from storage mediaoflamf aoonliunous
`muster nteof alaou 1.5 Mbitls. Nevenhelssit can be used morn widclythan this becansc Inc approach
`(aka) is generic.
`
`0. 1 . 1 coding parameters
`
`'1heimentim h_Iev¢1opixgIhis parrot ISOIlE'_'1.11'72 has been rodefine asouu: coding algorithm witha
`larp degree of flcxibiityflntcan be used in many diffcrent applicauons. To achicve this gm!,n number of
`the parnmetcrsdefining tlaecharaderistbsofooded bitslreamr and dncodzrsnre ooriained in1.he bitsneam
`iself. Thisnllnws forexnmple,Ihealgon'thm to be used forpicules witha variety ofsizzs and aspezt
`mlaa'mdon1:hannclsnrdovincsopu'atinga’awiIcrangcofbI1raxe4.
`
`Berauseoflheiargemngeoftlednractetistirzoibitsmamsflnuzn berepnesmtedbythispmtoflsollfic
`11172. asub-sat ot uletcooding parunetnrs known as the'ConsuainedPara_me1m" hasbccndefined. ‘me
`aimindefining Kite amsuainodparauetcrs is toolfcr guidance about: wicbly useful rnrgeofprramuers.
`Conforming tomissetoloonstrinls isnotareq11irem:ntoltlispnnot'lSOIlECl1172. Aflzgin the
`bitsmcam indicates whamrnrnot itisaconstrahed Parancters bimream.
`
`Summary of the constrained Parameters:
`
`Hodznnlal icuunsizc
` Lassthnnor altn768 ls
`L55 lhnnor
`~ alm575|ims
`Veaicnl 'oturcsize
`E
`Lessthanor aIto396ma:rnblocks
`Lrssdnnor .
`to396:(25mactoblocksIs
`lazsthanor to30Hz
`Less tlnm-6<Hn+63.Spels(nsing half-pd vecors)
`»-
`-xv - Loouhardiorward
`cod;-.<:=4 snetableb
`-- at butfersize inVBV model
`l..assthanorualtn3276&)biIs
`
`Biunt:
`Lessllnnos :
`tnl8560D0biIs/s (nonslanlbinte
`
`
`
`
`
`
`Pdral:
`Picnmnae
`Motion vncbnange
`
`0.2 Overview of the algoflthm
`
`Themded I:p1es'.nI:I‘nndefinedinIh's panolISOIIBC1II72achicves 2 hfgh ounprssion ralio while
`praetving good picture quality. 11:: algorimn isnotbssless as heenupel VIIJIICS are nmprcsaved
`during coding. Thechoice oflhctectnziqucsisbnsed on the need to halance a high picmreqmlity and
`ounlpvcssion ratio with the requimmnt tnmaka random moesslo the coded bilntrean. Obtaining good
`picture qnaity auhc bitntes of interest demands a very ligh compression min. which is nolachievaale
`will intrafnm: aiding alone. .1}: need for random access. however, isbst aalisfed with pure inuaftane
`coding. Thkreqliresacarefulbdanoebetwcen ima- at3din|::'fi'anIeoodingarrdbe.mee11recursiveand non-
`tecusiV¢ten1:nr-alrcdxmclarrcylednaion.
`
`Pagé'4 of 124
`
`
`
`o ISO/IEC
`
`uso/use 11172-2: 1993 (E)
`
`Annbaofuwmqoesmundmdniewaflghemptmimnxh. Iinfilavhichisalnmt
`indepmdeufroumispatoflKMEE1!171.i&bs:lcctmagnpfiaeqaaidrcsdufimtbrdnsigml.
`Imalgoflmmunnuscsmod-lnmdmaimwmpuuaionmmdwauucnnpanhedanducy. Motion
`mnpamflmisusuifmaumlpmdicflonoftheameupiunmfiomapuvbmgicme,tnruon-causal
`peliaioncfheanmtpicnIefiunamumpicu:e.o1fcrin1a-polaivupmdlcfim trompaandfmne
`giclnas. Mououvocmtsaahlnndftreath I6-pelbylsiincregionotlhepicmxe. Thedifferuucesigml,
`fiemdiofimuu.isfinlnrcanpmseduhgu:dsaenw§mmmnrm(D(J)wruwvespathl
`axnhfimbdueitkqtudndhmhmvuxiblepucmsmadswismelushpumlhfixlnim.
`Fmfll:.memoflmvanusucunbindfidlbDCI'infmunimuflodedmb;mhbhhg&wde&
`
`0. 2 .1 temporal processing
`
`Baamofthecumiuingmqficmauofmndmuwusmdhighlyemdmtmmpressimmnenmn
`picmruypsaedefincd. bu:-wdedpiamas(I-Picu:u)aeoodedwithoumfiaenoctoo1hapictms.
`Dnqpufimwcuummnmecuhdmqnnawhacdeuflhganbqmwtaemdwwimoflya
`uodeaxecnmpnssioomio Pmdiai1cwdedp'ctuu(P«Piuuxes)a1cw¢hdmo1efida1I1yusigmlion
`cmpumbdpredifimfionapmhnmpndhhewddpimmaduegmaanymcduudmamcfu
`futlcrptndlcfion. Bidixunhnnlly-ptediuive uoxhdpimIcs(B-Pimm) gunviicdnligheubgeeol
`cmmnsdonhxtreqnflrebmhputmdfimmzfueneepicuresbrnnuonaumpumim. Bidlmct'ualy-
`prediciveuxbdpimnesmnevetusedasneitxmcesfiotpnediction Ttnorgmixafionotumhruepiaxsn:
`typsinsuqmmcisvczyflexlble. 'IhechoioeisbfluoIheena>d¢randwilldcpa1donI21eIe¢piIeneusof
`the qzplicaflon. figure I illunmcs me uemionship luwecntlne thee rlfferempicnn types
`
`-’w’¢‘$ §V’v‘- M-..1,c.;.,.....
`
`Prediction
`
`(MIMI!
`
`mm.
`
`Figure I -- Example of temporal plot:-e strnctun
`
`The [omhpicurctypcdcfmcd in thispanoflS0l{EC 11172, the D-picunmispmviied Io allows simple,
`bullnitnd
`fast-fonratdylaybackuncb.
`'
`
`0.2.2 Motion toprouuainn - uacrobloelts
`
`neculm of16by lémaaroblocls forlhcnaotboconpusafionnnitisausuldmctmlooflbetwnn
`hlaeasmatlecodingdfuciacy pmvidedby u:immuiu1infomanonand1heove1headn«1zdedmsto1eiL
`Ead: tmcmblockcan ha one or ammberofdifferem types. Fot exxmplc. inn:-wdad. 1uwm-3Iediaiw-
`eodd.halvad»yadiaiwm(bdmdN&Iu1bully-peddwmdedmamhbcksmpmniundh
`bidheeinnaly-ptadhliveocxhdpicmnes I1-.pn1lingonthc1yp:o(Ihctnnatl:locls.mou'onvecInr
`hramnimunabusidcifamafimuumundwimthemmneswdpmdiaimcumsumlhmch
`Inaoroflock. Tbemodonvazwsuemeodcdaiflamfidlywihtupectmhclastwdcdnndmvuanr,
`using variablenlcnglh mdes. Thcumimnm length ofthe vecnra thaluuy-be repnzsenled cm In
`proyamned, magiclure-by-pia1nebas|s,sodn1lhemostdennnding:pplimiona umbcuctwilhout
`canptunising the parbnnmce ofthesystanin more normal situmims.
`
`ltlsthcluspnsifilityofmcawodarmmlcuhmappnoniwmotion vectors. Ihispmoflfi)/[EC [H71
`doesnaqaecifyhovthisslntldbedona
`
`Page 5 of 124
`
`
`
`BO/IEC 11172-2: 1993 (E)
`
`(D ISDIIEG
`
`0.2.3 spatial tcdundancy tadsction
`
`Both origin! picunes and radiation em: signals have high spalial rerlundmcy. Thispatt of ISOIIEC
`11172 mesa block-hmed DC!‘ memodvim visually weighted quantizalicm and run-length coding. Each 8
`by 8 Rock oflbeotigiml pietuxeforimm-coded macroblocts or oflleptedictionermrfor pmdictive-coded
`nambloch is transformed intoIheDC'I‘domain whereitis scaledbefote beingqmmimd. Afler
`quantization nany of meotefficieus arezwo in valueand so two-cimemioaal rm-length and variabk:
`length coding is used to encode hetunahlingcoefficiznle etficicntly.
`
`0.3 Encoding
`
`1hispanofISO!fl£l'll72docsmtspacifyaneneodingpmeess. Itspacfliesthesyntuumdsemmfics of
`Ihebhsueammdthesignalgxooessinginthedecodcr. Asaresnlt, manyoplions-arelaftopa1ncnno¢hrs
`nuade-offnostandspendagakrxqvicturequalityandwdixagefiieiancy. Tmisclmscisabriafdescripficmof
`Iheflmcfionstlnntnedtolaepetfonnedbyananodet. Fxguxezshowslbenuinfunaicnal blocks
`
`DC’! is disuete cnsineualsfum
`DC'.'1""is inverse flixrme ctsim uansfaln
`Q isqmmiznion
`Q1 is dequanizatiou
`VLC is variable length coding
`
`Figure 3 - Simplified video encoder block tiagnm
`
`Tlleinput vidcosigndinust he digfnzedand tepnwenbdasalunimnee and twoouluur diflemnoe signals
`(Y. Q). Cr). This may be followed by preproaessing and formal conversion In selectan appropriate
`window. msdulion and inputfmmt. Thispm of ISOIIEC ll 172 requi-es than the ooinur dilfeaeme
`signals (C5 and C.-)a1e subsanpled with nespct In the luminance by 2:! in both vcnica andhuzizomal
`rlimcimsmd amefomaued Ilm-.ccssay,asa non-imedwed sigml.
`
`I'be-encoder-mstchoose uhichpictutetype to use foreach yiclun. Having definedthc pictnretypes. Ihe
`encodcrestimalzs motion vectors breach 161:3 16 macmblocl: inn: picture.
`In?-Picuues onevecnr is
`needed breach non-inua nacobloekand in B-Piemrcsoncnuwo Vedas unneeded.
`
`If]!-Picunas uemed.somereon1ermgclI|epiamewqueIceisneoessaybeforeencodhg. BecamcB—
`Pictuesamoadedusiagbiairecfiolal motion wnpulsmndpledicfimythcytznuflybetbmdedafbt be
`subsequent re£e.remepictu’e(an I or]!-Picunc) hasbeendeuoded. Thenefosetbepimuesaxetuadered bytle
`
`vi
`Page 6 of 124
`
`
`
`(D ISCI/EC
`
`ISOIIEC 11172-2: I993 (E)
`
`ennodctsethat Ihcpictuvcsanivc atthchcodtarin titeotdcr tbrdctzoding. Th:mnactd'sp1ayortIetis
`tecuvundbyttedecoder.
`
`Tltebasictmitofcoditgwithinapiumeis thetmcmblodt. Wttttincachpictttm. mamblocksueencoded
`inseqttence. lefi to right, top to bottom; Eafiunacmblock consists ofsix8 by 8 blocks font bioclts of
`Iunimmemneblockofch chmminanoe. atdone block of Crd:rmnit1ancc.See figueafiotettnt the
`pictmcapacuvcratlby tttefourbloclts otltminanccisthcsamemtlnareacwwedbycadntfthe
`chmminance bloclts. This is due tostbsamplingofthe chrominanceinfotmatiun to tnntdt an snnsitivity of
`the htttmnvistnl system.
`
`EIE '5‘
`
`'1‘
`
`Y
`
`Cb
`
`C:
`
`Fgure 3 — ‘Mad-obloclt structure
`
`.
`Fttstly. fora given macrobiotic, thcooding meat: is chosen. Itdepends on the piclum type, the
`eflectivennssofmotion ootnpcmated pxaitctim ‘n that locatmgion, and the natun, of the signalwihin the
`block. Semndtg, depending an heootting node. a maion tnmpensated prediction of the contents ofthe
`blodtbasetl at past andlurtuutn: tefctcnna pictmes is turned. This ptediuiuzt is subtracted flan the actual
`daain Ixecunmtmatmbtod: tan roman emorsigmt. Ihirdly. tlisenorsignal is scparatedinb 8 by 8
`blocks (4 Imam: mdfz chrontinanoe btocksineatzhmacmblockj and sdiazeteoosinc hansfamis
`perforated on cachbloclt. Each tesulttng 8 by 8 block of DCT cocflicicnts is quantized and the two-
`dinensinna] block is seamed in a zigzag outer to convert it into a one-dimasjottal string 0! qtnntizal DC!‘
`otactftcicnts. Fmrthly. the sitb-hformatbn fat the tn_actd1loclt(Inode. motion vectors etc) and the
`qmntimd codficicnt data atccncodsd. Rx tnn efliiettcy, anunbet ofvarizble lalgh code tablesan-.
`defined [otthc dfferent Iata etanents. Run-length mding is used forthcquantized coefficimtdaa.
`
`A colsequcnoc of ufing tliffetattpictmetypesatxl mriabb hngtlteoding isttuttthcovemtl dataraeis
`variable In qaplicatitms that involve: fixed-utechttnnel. aFlJ-‘Obutfer may be used to tmtdt thcencode:
`otiput to the channel. Tbestatus oftlisbutfcr maybe tnonitoted tooonttol the nuniaetj qfbits generated
`by the cncodct. Conmling theqttantizatitn ptutaess is the tntst dimct way of oontrollingthe bitnte. This
`pan of ISOIJFL‘ 1 1172 specifas anabstnctmodcl of the bulfeting systnm (the Videonntlbrhg Vaiftcr) in
`order toctnsttain tttctmxinum variability in the number of his that are used (ora given picture. This
`emunesthaabitstrcam can be decoded with aburfct ottcnown sin
`
`Attbis stagmthecodedtegxestttatianofthepichuehasheen ganemted. The finalstopintheatoodetisto
`regatta:-mt-Picums and?-Pi¢(lIrBSbydwt1dlIIgIlCda(8S)lhaItlEyh‘.lB6da5 tcfetemcpichnasfu.
`subsaqttentettooding. 'I'h:quant‘medcoeffia‘etis tttedequtntizedsndanirtvetsestoyltbcfispcrfltrtnedon
`eachblock. 1hepredtctbnemI'ttigtta!producedisIttettaddedbaclt tolhepmdictiontigtal and limitzdto
`thetequimttnngctogiveadeuotbdrehuaoepicutt.
`
`0.4 Decoding
`
`Decoding is theinvetse of thcemcuiing opuation. It is cousidaably sitnplm thnnenootling as mate '3 In
`need to pafotmmotion astimation and them are many feveroptims. Thedeoodirtg paces is dnfined by
`this pan of ISOIIEC 1] 172. The dcsaipim that follows isa very brief ovmview ofone posibte way of
`deuotling abislteatn. Otter tboodets with diflcrmtamhineutues nrcposstblc. I-igutt 4sltows the main
`fimcfionat blocks.
`
`Page 7 of 124
`
`vii
`
`
`
`TSQIIEC 111722: 1 993 (B
`
`QISOIIE
`
`
`
`isaequannzaion
`Q-1
`Mux-1 isdanilfplexitlg
`VLD
`isvariable length deaading
`
`Figure 4 - Basic vileo decoder block diagram
`
`Formed-1-an applications, mcchannel fillsafll-D bulferalacousunuam with mecodedlilstream. The
`deooch mils Ibis blfltrandtficods lhedaladum-mshnhebilstnam according toxledefined syuax.
`
`Asthedecodaxmdsthebitsueam,-hidentiflslhestanolaeodedpicuueanl rhea tletypaoflhcpicmre.
`fldecodseachnacroblockimhelicuueln um. Themaunblocrkrypeaxudthcmotion vectors. ifplf-‘MEI.
`aeusedlocolxsuuclaplndictian oflhcunuumwroblodclnsodanpastmdfimnetcfexenoepimres that
`vluavebeeustoredinlnedecodet. lheooeflidezttdalaaxtdeoodedsnddequmized. I.-‘.ach8 lvy8 lnockof
`mefiiciamlanis Innsfonnedbyan inveIseDCI'(specified in mncxA). andmsmsultisaddedtothc
`pmdidionsigpalandlinbd Intlladefilednnge.
`
`Amznllthemauohlocksinlhcyicnua havebeenptocassed Ihepicurehasbeenraconstn|c1ed.Ititisanl-
`picIweoraP-pimncitisarefetmocpictmtfawubsequent pit1umsandissbmd.mplacing Ihcoldcstsxxed
`mfaucepictue. Bcfimetbepicmezaaedisplayed Iheymayneedzobete-orfleredfrmnmecodeduderm
`theiruaunl displayotder. Afiaweordzring. me picuns:rea\ailabb.in ¢ligiuifom.iorpost~pmccssing
`anddisphyinany manna’ than thcapplicalionchooscs.
`
`0.5 Structure of the coded video bitstream
`
`’l‘h's pano(ISOIII-3011172 specifics: syntax foracoded video billstn-.an. This syntax contains sixlayers,
`eadlof whicheixher supports a signal ptooessingota sjmeln fisnction:
`
`‘
`Sequeme layer
`Randomaccas uni: contnxl
`Gmupufpicnnes laya
`Random accssunilz video
`Picture layer
`Primary coding unit
`Slice layer
`Rssyuthronization In‘:
`Mannhlocklayet
`Motion cnmpemalion unit
`Blockla s-
`
`DCT unit
`
`3
`V
`
`0.6 Features supported by the aigolithm
`
`Applications uslugumpessedvldeoomligital slongemednaneedbletbbtoperkxm anunberof
`opelationsinndcilion Lononnalfntwasdplaybackomn sequence.
`'l'hecodedbitstreaml1sbemdesigned
`I3 supportanumberof these operations.
`
`viii
`
`Page 8 of 124
`
`
`
`Q ISOIIEC
`
`I50/IEC 11172-2: 1993 (E)
`
`0.6.! Random access
`
`Randemaoeessisawsssnlialfeatnreforvldeounastorasenecliun. Randnmaouessreqninsslhnany
`picture cnnbedcoodediuliniad amumtoftime. Itimpliesthcexistazccotaocsspoinlsinthe
`Murmu-mazksegnamofmfummmmuaeidmdfiabhmdanbedandedwmomnfamocmuhu‘
`aagmentsofdm. Aspnaiugoftwoumdunmuzsspaims Elma-Pichms)persc.cmdca1beaclicve:l
`without significant loss ofpicurc quality-
`
`0.8.2 Fast search
`
`Dependingon thcstnragcmalinn. iIispossibletoscantheacne$pointsh1aoodedbitstteam(wmchc
`helpofan quplication-specific citectorymolsaknowlcdgc beyondlhc soopcofIhispanoflSOIIEC
`11172 tnoblainafast-fnI'vraniandl':m-re\'e::e playbad:a‘Iect.
`
`0.6. 3 nevus: playback
`
`Samcawltationsmay taquixc thcvidoosignal‘Iobephyadinm.eascmIlcr. 'Ihisambeachiavedina
`decoder byusing memory IO stone mfiregzotpsofpictlns ametthayhaveheendeooded bufineheing
`displayedin nzvcmeoulet. Anencadercanmahc flfisfieamreesiahy nemanglhehngtlx ofgroupsol
`pictums.
`
`0 .84 Error robustness
`
`Most digital storage nediaani connmication channels axe notcn'ot—£Iat'. Appmpxiat: dzaunel noting
`schcmcsshodd beusedand uebeymdthc suopeofu1isparLofIS()IIEC 11172. Nevutleless the
`conpcssim scheme definadimhis panof ISOIIBC-11172 '3 tobusl totesidual errors. The slice structure
`allcwsadeeodermmaovezartuadanmnrand to rtsynclunnizcitsdeooaixg. 'I1Ieu-.fore.bitcnoasind:e
`oompiessedduawfllcauseeuorshdnedecodedpicutesbbeimimdhaxem Deoodezsmwbeablelalse
`oomealnentsmucgiesm disguise Ihcsecnms.
`
`0-6.5 Editing
`
`'1he:cis acacnflict hcwvaanlhe naquixumcnt for high coding cfficieucy andeasyedting. Ihecalfingsuuonnr.
`and syntax havenot been designed with the primary aim of sinplifying editing anmy picture. Nevertheless
`ammbetoffeaumzs havnbemiuclutbdthatulabluediting ofoodeddan.
`
`Page 9 of 124
`
`
`
`This page intentionally left blank
`
`Page 10 of 124
`
`
`
`
`
`
`
`NTERNATIONKL STANDARD 0 ISOIIEC ISDIIEC 1 1172-2: 1993 (E)
`
`Information technology —— Coding of moving
`pictures and associated audio for digital storage
`media at up to about 1,5 Mbit/s —
`
`Part 2:
`
`\fideo
`
`Section 1: General
`
`1.1 Scope
`
`T11ispanoflS()!IEClli7Zspccifi:sthcoodedrepvesemttionotvi1eo for-digiul storrtgemediannd
`specifies the deciding process. The representation supports normal speed forward playback. as well as
`special functions such as rmdnn access. fast forward playback, inst reverseplayhadc, normal speed reverse
`playback. pause and still pictures. ‘Dds pmofISOlIE(.' 11172 lscompatble with standard 525- and 625-
`line television formats. and it provides tlcaibility for use with personal computer and worltslalion displays.
`
`I80/IEC 1l172is primarily applicable to digital storage meciasuppottitg a continuous tnmfer me up to
`about 1.5 Mbitls, such: as Compact Disc. Digital Ando Tape, and magnetic haul dslrs. Nevertheless item
`be used Inorewiiely than his hecarscoflhc generic approact taken. The storage ncdiarmybedireaiy
`connected to the dccoder, or
`rneanssuchas bussw, LANs, or telecommunications
`links. This putoflSDfIEC 11172 is intended for non-interlaced video formats having apprmixnarely 88
`lines of 35?. pels and picture ratesamund 24Hz to 30112.
`
`I .2 Normative references
`
`‘I11: following International Standards contain mwisions which. thmglt reformer: in this tell. colntltule
`provisions of this part of ISOIIEC 11172. At the time of publication. the. editions indium! wen}: valid.
`Allstandmdsnresubjectto nwisiou. and panlestoagemtsulslramdonthis partoflSOf]EC11172 are
`tmoumgal to investigate lmpowiaility of applying Ihcmostreoent editions of the stmlardsindicated
`below- Membus of EC and ISO mirzrtain rcghtcrs of currently valid International Standavds.
`
`ISOIIEC 11 172-1 :1993 lnjonnaim technology Coding cfmotiru pu'crtm.r and associated audafor drgiul
`rtarazemedia at up to about 1.5 Mb-it/J - Part I: System.
`
`ISOIIEC tlt7z3:I993 Inpmuriau technology - Coding rimming pictures and arsaciared audbfordrgiral
`rrarazemcfia at up to about 1.5 Mbu/s - Part 3 Auaia.
`
`CCIR Recommendation 601-2 Enmding paramzrers ofdigital relevisionfinr sludiat.
`
`CCIR Report 624-4 Clurccterirtics ofsysleuu jar monochrome and cokmr l‘elevi.u'on.
`
`CCIR Raooumendatiou 648Rccanfing Qfandb signals.
`
`CCIR Report 955-2 Sound bmadcasfing by satdliteprporrable mdnwbie receivers. tnduding Annex IV
`Summary description ofAdvrutced lxgiral Sysum II.
`
`CCITI‘ Recotnmetdalion {.17 Pr!-emplultis‘ usedan Sound-Pmgranum Circuits.
`
`Page 11 of 124
`
`
`
`ISOIEC 11 172-2: 1993 (E)
`
`D SO/IEC
`
`IEEEDIBI1 Slandsrd P1180/D2 1990 Speaficafioujbr the implen-enralion af&: 8 inverse discrete m.n'nc
`1rau.vanu".
`
`nzc publicatbn 9418:1937 cv Digilrzl Audio System.
`
`2
`
`Page 12 of 124
`
`
`
`O ISOIIEC
`
`ISOREC 11172-2: 19% [E]
`
`seclion 2: Technical elements
`
`2.1 Definitions
`
`fixnhc purposcsoi ISOIIZEC 11172, IIW following dztiniuons apply. Ilspecifi: In a put. this is noted in
`squuebiactetx.
`
`2.1.1 nc cocfllclun lvldoolz Any DCT coefficient fa which the frequency in one or both dimensions
`is non-zao.
`
`2.1.2 accesu unit lsgstcml: In the caseofoompressedaudinanaooessuiitisanandioaocess mi. In
`Itemseofamtyessedvidcoanaoemmitisunoncledrepresallztimotaplanxn
`
`2.1.3 adaptive segmentation [audio]: A subdivision of the digital reprwamlion of an indie signal
`in variable spgmenls of lime.
`
`2.1.1 adnplivc bit allocation [audio]; The assignment of hits to subhands in a time and fwqnelcy
`varying fashion acconling In a psychaamuslic model.
`]
`
`2.1.5 adaptive noise dlocation [antic]: The assigmnam of coding noise In frcqnency hands in a
`line and frequency vaying faslfnn ammiilg ma psychoaconsfic mood.
`
`2.1.6 alits [audio]: I/[irmred signal component resulting from suh—NyquisI sampling.
`
`2.1.1 analysis filterlalk [audio]: Fihetbank in tin encode: that transforms a broadband PCM audio
`signal into a set of subsanpled subband samples.
`
`21.3 nullo access will [audio]: For—Laycrs I and II an audio access unit is defined as the smallest
`pattoflte wooded bistreann whidacm be decodedby itself, wheredecodetl mam "hilly ueeonsrucled
`sound‘. Forlayartllan audio acoessmitispartoflsebistreamunl ‘sdecodahle with lheusnof
`
`2.1.9 uulin lmfler [audio]: A bulfer in thesystan target decoder for slonge of cnmpiessed audio dam.
`
`2.1.10 audio sequence [audio]: A non-intern-oped series of audio francs in which we following
`
`-&mpling!‘1equenqr
`- ForLayer lam! II: Biuac index
`
`2.1.11 backward motion wclor [video]: A motion vaclor mat is used for niolion canpensatim
`fiomardcaencepiclmo atalata-timeindisplayouder.
`
`2.1.1.2 Burk [nudio]: Unit of clinical hand me. 1116 Bath scale is anon-linear mapping of the frequency
`seal.-. ova‘ the anrlio range closely conesponding with the inaqualq selectivity of Lhe banana’ acrosslhe
`ham
`
`2.1.13 hiiireclionally predictive-coded picture; B-picture [video]: A picmre that is coded
`using motim compensated pmdictioo h-oma past zndlot fume teferenoc picture.
`
`11.14 bin-ate: 1'11: tall: at which the compressed bilstream isdclivetcd mm the storage medium to do
`inputofa tkcoder.
`
`2.1.15 block callpamding [nudb]: Nonmlizing of the dgial representation oi an audio signal
`within a certain (me puiod.
`
`2.1.16 flock [video]: All 8-row by 8-cnlum orthogonal blod: of pds.
`
`2.1.11 bound [audio]: The lowest sullband in which intensity stereo main; is used.
`
`Page 13 of 124
`
`
`
`ISO/IEO 11172-2 1993 (E)
`
`O I%flB3
`
`2.1.1! layh nligt-od:Abiin1codedbiIsu-can Isbylc-alignodifltspcnitionis amulIip1ea£8-hits
`fiomthctixstbitlnthennun.
`
`2.1.19 byte: Sequ-me ol 8-bits.
`
`2.1.20 «banal: A digital medium that stares or umspons an ISOFIE 11 W2 -mean-
`
`2.1.21 dunno! lsudlol: ‘Helen andrighl channels of a stereo dgnal
`
`2.1.23 chronalnncc (component) [video]: A matrix. block at sixglc pel representing one at the
`two colourdiffctencesignalstelated to Inepdmuy colaursin the umnnnlefinad in CUR Rec601 . The
`symholsusslforneoobwtiflamcasigxnlsaaccrnflcb.
`
`2.1.2: coded mule bltstrum hallo]: A coded teplesamlbn ofan undo signal as spedfied ‘u
`rsolusc 111713.
`
`2.l..24coded vldnlmsu-cam Mao}: A codedrepxescnationolasaricscfoneormotcpicttnesas
`spacifiedinthispmoIlSOIll'JC 11112.
`
`Llzsodadudcr Ivileolzneorlctin which mepictum restored mddecodai. Thisordetis wt
`uenplsnlilynnameasnhedkplayoader.
`
`2.1.26 codod rcpruanhtlm: A data element as repnsentcd in Its encoded fonn.
`
`2.1.11 todlng parameters [video]: The setof user-definable pumnccts that charnccfzc a coded video
`bitsu-cam. Bismcamsue dnmaaisedbycoding pmnelas Deoodusm ctamcetiaedby Ixebitsueans
`that the} an capable of deaoding.
`
`2.1.28 w|'1poueIl.lvitleo]:Annl.Iix.bbckor single plfrunone ofthc flue: mazficesmmimnee
`andtvIod:tomiamc:)maInnl:npnpt:mc
`
`2.1.20 coup-usbu lleductiminlhcnmtoertfbitsusedto rcpesemar;iIam ofrhn.
`
`2.130 constant bltrnk coda! video [Video]: A compnssed video biuuean with a whsmn:
`avumebinn.
`
`2.1.3! mutant hitnte: qaeradon whaemc hilrau is constant frun sun to finish of in: compress-A
`bisuean.
`
`2.1.32 eonslralmd plnmeta-s [video]:
`2.4.3.2.
`
`111-, values 1!‘ mean a‘ coding panama: defined In
`
`2.1.33 coma-ulna! nystun pnnnncler strum (CSP8) [system]: An ISOIIEC 11172
`mllfplcxed strlam to: which the oonanin:sauinedinza.ooI|so!IB(‘.11172,Iapp1y.
`
`2.1.33 CRO Cyclic ndwdmcy code.
`
`2.13 crllul band nu IIIIIIOII Psycbmoousic functim at frequency. A! a given audible
`freqnxemyitis prq:vo1ionIl In thcnumbctofcriial bmdsbelow that frequency. The units ofthecliticai
`Imdmtescalesemdn.
`
`2.1.3 ex-Dial baud [audio]: Psychoawustic mwsune in the spectral domain wlicl coaesponds to the
`mquemyseleaivixyofltehunanear. This selectivityis exptessed it Rank.
`
`2.1.37 lam elunent An item of data as rcpuaaemed bcforcenooding and after dacodiug
`
`2.1.38 dc-coellldent [Vila]: The DU ooetficielt for which the fieqsency in mm in both
`dhnmsious.
`
`Page"14 of 124
`
`
`
`(DISC/IEC
`
`ISO/IEO 11172-2: 1993 (E)
`
`2.1.39 dc-coded picture; 1)-plcture[video): A picture that is coded using only information from
`itself. 0ftheDCl' eoetficiems in theoodeil representation, only the dc-ooetficientsase pxesenl.
`
`LL40 DCTeoell‘1cient: The amplitude of a specific cosine basis function.
`
`2.1.41 decoded stream: The decoded reconstruction of aeompressed bitstream.
`
`2.1.42 decoder input buffa [video]: The fitst—in first-out
`
`(FIFO) hufier specified in the video
`
`2.1.6 detoder Input rate lvileolz The clan ntespedfied in the video buffexing verifier mdeocoded
`inthe coded videobiisueam.
`
`2.1.44 decoder: An embodiment of a decoding process.
`
`2.1.£ decoding (process): The process defined in ISO/11-IC 11172 tint reads an input coded bitstrean
`anti produoesdeaoded pictures or audio samples.
`
`2.1.46 decoding tine-stump; DTS Isysteml: A field that may be plaeut in a packa headet that
`indicates the time-ha an access tnit is decoded in hesystetn targetdoeodet.
`
`2.1.47 dc-emphasis laudiol: Filming applied to an aucio signal after storage or transmision to undo
`a linear distortion due to enlflaasis.
`.
`
`2.1.48 degnmtlntion [video]: The process of rescaling the quantized DCI‘ cocifieients aha their
`representation in the bitstream has been deooaadalxlbefmethcyueptesemedto the invetseDC'I‘.
`
`2.1.49 digital storage media; osn: A digitnlstomge on transmission cbvice a} system.
`
`2.1.5) disarm cosine ttaltsforln; new lvideolz Either the forward discrete cosine tmsrom or the
`invetse discrete ctsire transform. The DC!‘ is at invertible. discrete orthogonal tmnsfomatim. The
`inverse DCl‘is defined in annex A.
`
`2.1.51 display order [video]: The ordu ln which the decoded phone: should be dsplayed. Normally
`this iultesameonlerinwhida they were pesentedatthe input tithe enoodet.
`
`2.1.52 and l:IIlI'1ncl_lI0d¢ [audio]: A mode, when two audio channels with indqaendeut pmxralnnte
`curtail: (e.g. bflingualhreeocoded within onebisneam. Thecoding puocess isthe sunem lbnhe stereo
`mode-
`
`2.1.53 «ding: The process uy which one ormore compressedbilsnearns aremanipulnted to pounce a
`new comptessed bilstneam. . Conforming edited biuueans rust meat the zequitenattsdtflned in this part of
`ISO/IEC 11172.
`
`2.1.54 elementary stream [system]: A generic term for one of the coded video. oodedaudio or other
`coded bltsueants.
`
`2.1.56 emphasis [audio]: filtering appli.-d to an audits signal before storage or transmission to
`improve the signal-to-noise ratio at high fiequencies.
`
`2.1.56 encoder An embodiment ofan encoding process.
`
`2.1.57 encoding (process): A process, mt specific! in 180/[EC 11172. tlnt reads nsuean of input
`pieturesoraudio samples andpmduces a valid coded bitsumm as defined inISOI[EC I1 172.
`
`2.1.58 entnpy ending: Vatiable length Jossless coding of the digital representaion of A signal to
`tehmmdnndanqy.
`
`2.1.59 fast Ion-ward playback lvideolz flu: graces of displaying a sequence, or pans of a sequence,
`ofpicmres indisphy-ordet faster than xtaltime.
`
`Page 15 of 124
`
`5
`
`
`
`ISOIIEC 11172-2:1993 (E)
`
`O ISOIIEC
`
`2.1.50 FIT: Fat Fourier 'I‘mnsfa:matbn. A fast algoiitbm for pertaining a discrete I-‘oixier tnmsfoan
`(an omoxonal transform).
`
`2.1.61 fllatlunk [audio]: A set of band-pass fillfls oovczing the entire audio frequency range.
`
`2.1.62 fixed segmonulbn [audio]: A subdivision of the digital xepresntalion of an audio signal
`into fixed segnansuftine.
`
`2.1.63 forbidden Themnn "forbidden" when used in Iheclauses defining the mdedbilsuan ildicaes
`thauhe value small uesetbeimd. ‘Ibis isnsuallyloavoidexnuhlinnofsmn macs
`
`11.64 {crud updating [video]: The process by whid: mmmblocls are inm-coded tom tine-to-Lime
`noeasueflaatnisnmchenushetwean IhcinveIseDCI‘ptooesscsinencodasandnbeodascannotbui11up
`excessively.
`
`2.1.55 fa-ward motion vector [video]: A motion vector that '3 used for motion eunpelsaion {turn
`ateference plcutcal an earlier time in dsplay cider.
`
`2.1.6611-am: [audio]: A part of the audio signal that corresponds to nude I’-‘CM sample: Ii-om an
`Audio Access Unit.
`
`2.1.61 the format [aud|o): Any hitme othe: than the defined bitmms mat is less than the maximum
`valid hiuatefot acblayet.
`
`11.68 (lure refeteuee picture {video}: The futuze referent: picture is tie neferem: pimne mm
`mcmsatahlathnennnmecumnlpiuimmdisplay order.
`
`2.1.69 ganulos [Layer II} [audio]: The set of 3 consecutive snbband samples from :11 32 subbmds
`tlataeeonsidctedmgelherbefonqumlimion. Ihey eonaspoui ln96l’CMsampbs.
`
`2.1.70 granules [Layer III] [audio]: 576 frequency Ii 11131 catty their own side ‘nfonnalioe.
`
`2.1.71 gnup of pictures [video]: A series of one or momecoded pictures intended to assist random
`access.
`'l1:egmup<f1x‘cmn:s's oneofthc Iayersin Ihecoding sytmx defined in this panofISOIIEC
`11 172.
`
`2.1.12 Ham window [audio]: A time function applied sample-by-sample to a block 01 aud‘n samples
`bet'aeFouri:r uanslomaiion.
`
`7.1.73 llnffman coding: A specificlnethod for entropy noting.
`
`2.1.74 hybtld fllterbank lnudiol: Aserial cnmbinalion of subband filenbank and MDCT.
`
`2.1.75 INDCT [audio]: Invexse Modfied Dictctc Cosine Trarsform.
`
`21.76 lnlaeslty sures laudiolz A method of exploiting stereo inelevancc Q’ redundancy in
`sxereophonic audio programmes based an netainingathigh frequencies onlythc enetgyenvelopecf me right
`andlefi c|aImeLe.
`
`2.1.711nterla¢e [video]: The property of conventional television pictures where alternating linasof
`the piaun: reptesemditfcrem instances in time.
`
`21.78 inlra ending [video]: Coding of a macroblock or picture um uses intornaion only {mm that
`nanmblockor picmxe.
`
`11.191mra-ended picture; I-pictuxe [video]: A picture coded using information only from itself.
`
`Page-616 of 124
`
`
`
`O ISO/IEC
`
`ISOVIEC 111722: 1 993 (E)
`
`2.1.80 ISOIIEC 11112 (mnltigdsud) stream [system]: A bilstraan composed of am) or mom
`elenentaysttennsounbinadinthcnanurd:fi1edinlS(YlI-1C1ll72-1.
`
`stotoo oodtrg [audio]: Any named that expldts sueophoni: imelevnc: ct
`2.1.81
`suuqhuncrodxndancy.
`
`2.1.82 jam stereo node lnndiol: A mode of the audio coding algorithm usizg juinl sumo coding.
`
`2.1.3 layer [omllo]: Oncof tin lcvcls in ct: coding liemruby of lhcaudio system deflnad in ISOIIEC
`111
`-3.
`
`2.134|ayer lvtaleoand systeIs]:0tnoftbelcvcls1ntha¢hnIic:archyotbcviieomdsysnu
`spedfii-aicnsdetinodi:IlSDIlEC1l172-18dthispnuof[SOIIEC11171
`