ISO/IEC JTC 1/SC 29 N 2027
DATE: 1997-05-01
OUR REF.: 29CL383/29F055/29D139/


Coding of Audio, Picture, Multimedia and Hypermedia Information

Secretariat: Japan (JISC)

DOC. TYPE Meeting Report
TITLE Meeting Report, the Thirty-eighth ISO/IEC JTC 1/SC 29/WG 11 Meeting, 1997-02-17/21, Sevilla, Spain [JTC 1/SC 29/WG 11 N 1510]
SOURCE Convener, ISO/IEC JTC 1/SC 29/WG 11
STATUS In accordance with Resolution 1 taken at the thirty-ninth SC 29/WG 11 meeting, 1997-04-07/11, Bristol, UK, WG 11 has approved this report.
REFERENCES SC 29 N 1941 (WG 11 N 1508): Resolutions
REQUESTED ACT. For SC 29's information
DISTRIBUTION P-, O- and L-members, ISO/IEC JTC 1/SC 29
Officers, ISO/IEC JTC 1/SC 29
Secretariat, ISO/IEC JTC 1

Narumi Hirose, Secretariat, ISO/IEC JTC 1/SC 29
IPSJ/ITSCJ*, Room 308-3, Kikai-Shinko-Kaikan Bldg., 3-5-8, Shiba-Koen, Minato-Ku Tokyo 105 Japan
Telephone: +81-3-3431-2808; Facsimile: +81-3-3431-6493;
Telex: 2425340 IPSJ J; E-mail:
* Information Processing Society of Japan/Information Technology Standards Commission of Japan (A standards organization accredited by JISC)


ISO/IEC JTC1/SC29/WG11 N1510
MPEG 97/
March 1997

Source: Leonardo Chiariglione
Title: Report of Sevilla meeting
Status: Draft

1. Opening

The 38th WG11 meeting was held in Sevilla, Spain on 17-21 February 1997 at the kind invitation of AENOR, the Spanish National Body.

2. Roll call of participants

The attendance list is given in Annex 1.

3. Approval of agenda

The agenda is given in Annex 2.

4. Allocation of contributions

The list of contributions is given in Annex 3.

5. Communications from Convenor

The Convenor stressed the importance of the goal-driven attitude of MPEG as follows:

  1. MPEG is a standards committee
  2. MPEG is a result-driven committee, hence
    1. the workplan plays a primary role
    2. at MPEG meetings subgroups are given tasks
    3. between MPEG meetings ad-hoc groups are established with precise mandates
  3. Souls searching exists in MPEG but only at preliminary phases of work

6. Report of previous meeting

The Maceio meeting report could not be approved. The Convenor stressed his determination to get the necessary support from subgroup Chairs.

7. Processing of NB Position Papers

Documents from Japan and US were considered and a reply provided.

8. MPEG Phase 2

8.1 Systems

Corrigendum 1 to 13818-1 was approved.

8.2 Audio

Version 2 of 13818-3 was promoted to IS.

8.3 Verification of MPEG-2

No activity took place at the meeting.

8.4 Amendments

Amendment 4 to 13818-1 was promoted to DAM.

A new Amendment 5 to 13818-1 was initiated.

An amendment to 13818-4 to accommodate AAC was initiated.

An amendment to 13818-5 to accommodate AAC was initiated.

8.5 Part 7 (AAC)

The AAC being under ballot the activity was confined to discussing clarification to the text. Patent statements were received from several IPR owners.

8.6 Part 10 (DSM-CC Conformance)

This being the first time DSM-CC conformance was discussed no substantive output could be produced.

8.7 Workplan

This was approved.

9. MPEG Phase 4

9.1 Requirements

A new version of the requirements document was produced. This included a new section on IPR management.

9.2 Tools

A large number of contributions addressing all parts of the MPEG-4 standard were submitted and processed.

9.3 Verification Models

New versions of the Systems, Video Audio and SNHC VMs were produced.

9.4 Call for proposals

Several submissions were received. Parts of them were channelled through the Core Experiment mechanism. Another set was retained for formal tests in July.

9.6 Simulation software

MPEG will produce simulation software for Systems, Video, Audio and SNHC both for encoder and decoder. A decision about DMIF will be made in Bristol.

The need to achieve a timely integration of the software made available by members promted the kick-off of an "MPEG-4 player" project. Mike Coleman kindly undertook the task of leading this activity that will provide a first level of integration in July.

9.7 Working Draft

New versions of the Systems, Video and Audio WDs were produced.

9.8 Workplan

This was approved.

10. MPEG Phase 7

A new version of the MPEG-7 Context and Objectives was produced.

It was decided to organise a seminar in Bristol with invited speakers.

11. Overall WG11 workplan

This was approved.

12. Liaison matters

Liaison documents were processed and liaison letters produced.

13. Administrative matters

13.1 Schedule of future MPEG meetings

This was approved.

13.2 Promotion of MPEG

An "MPEG-4 overview" document was produced. As the document was still in a draft form publication was withheld until Bristol.

A press release was approved.

14. Organisation of this meeting

14.1 Tasks for subgroups

The list of tasks assigned to subgroups is given below. Reports of subgroups appear as annexes 4 to 12.


MPEG-4 Requirements document including
Profiles document
MPEG-7 Context and objectives
MPEG-4 technical overview


DSM-CC Conformance Testing PWD
DMIF Reference Software
Joint Seminar with ITU-T SG 16
"DMIF context and goals" document
DMIF workplan


1/COR 1 DoC & Final Text
1/AMD 4 DoC & Final Text
Identify WD editors
Systems VM 3.0
Systems WD 3.0
Systems FAQ
Document on MPEG-4 technical description



Plan for SNHC reference software
Identify WD editors
Document on MPEG-4 technical description


Selection of test material
Competition test plan and procedures
Verification plan (July '97)
Verification plan (beyond)
Advise Video


Study liaisons received
Receive proposed external liaisons from other subgroups
Issue liaison statements to other bodies


MPEG-4 Technical Report
Plans for MPEG-4 viewer
Liaisons (all through Liaison sub-group)
Resolution drafting
Collection of subgroup resolutions
Collection of ahg forms
Responses to National Bodies
Promotion of MPEG
Press Release
MPEG Home Page

14.2 Finalisation of meeting allocation
Req-Sys-DMIFTu 16-18
Req-Aud-TstTu 15-16
Tst-Vid (Err. Rob.)Tu 09-11
Tst-Aud-Vid (AV Tests)Mo 16-17
Vid-Sys-TstWed 15:00-16:00
SNHC-SysWed 10:30-11:30
Content mgmtWed14:00-
Patent issuesWed 19:00-20:00
DMIF SeminarFri09:00-12:00

15. Planning of future activities

The following subgroups were established
1596Ad-hoc group for frequently asked questions on MPEG-4 video
1533Adhoc group on audio FAQ for MPEG homepage
1603Adhoc group on composition
1605Adhoc group on Computational Graceful Degradation Techniques
1574Adhoc group on content related IPR issues
1592Ad-hoc group on core experiments on coding efficiency in MPEG-4 video
1589Ad-hoc group on core experiments on coding of arbitrarily shaped objects in MPEG-4 video
1590Ad-hoc group on core experiments on error resilience in MPEG-4 video
1591Ad-hoc group on core experiments on multifunctional coding in MPEG-4 video
1575Adhoc group on developing "MPEG-4 Technical Overview"
1576Adhoc group on DMIF activities for MPEG-4
1593Ad-hoc group on editing the documents of the MPEG-4 video verification model and the MPEG-4 visual working draft
1556Adhoc group on evaluation methods and procedures for MPEG-4 tests
1604Adhoc group on evaluation of MPEG-4 video VM complexity
1548Adhoc group on face/body animation
1550Adhoc group on generic SNHC object coding
1577Adhoc group on low delay issues
1549Adhoc group on media integration of text and graphics
1573Adhoc group on MPEG requirements
1530Adhoc group on MPEG-2 AAC conformance and technical report
1528Adhoc group on MPEG-2 AAC IS preparation
1529Adhoc group on MPEG-2 AAC prediction complexity reduction
1532Adhoc group on MPEG-4 audio core experiments
1531Adhoc group on MPEG-4 audio WD and VM
1602Adhoc group on multiplex specification and signaling
1606Adhoc group on Padding Complexity Evaluations
1597Adhoc group on SNHC VM/WD editing
1551Adhoc group on SNHC VM/WD integration into MPEG-4 WD
1601Adhoc group on systems VM
1600Adhoc group on systems WD editing

16. Resolutions of this meeting

These were approved.

17. A.O.B

Constant progress of MPEG-4 towards CD made MPEG aware of the urgency for industrial users of the standard to identify a mechanism to solve the problem of IPR that is necessary to implement the MPEG-4 standard. Members were encouraged to bring to the next meeting information on similar experiences and suggestions about their possible applicability to the MPEG-4 case.

18. Closing

The meeting closed on 97/02/21 21:30.

JohnArnold University of New South WalesAustralia
AlbertDeKnuydt ESAT-VISICS KuLeuvenBelgium
StefDesmet ESAT-VISICS KuLeuvenBelgium
EricBadiqu European Comission DG XIII/B2Belgium
JanBoemans IMECBelgium
JanBormans IMECBelgium
ChristophDe Vleeschouwer UCL-Laboratory TeleBelgium
Mauri Vaananen NokiaFinland
MartaKarczewicz Nokia Research CenterFinland
PasiOjala Nokia Research CenterFinland
MarcinRzeszutko Nokia Research CenterFinland
BrigitteSchuster Alcatel Business SystemsFrance
Jean Pierre HenotCCETTFrance
BernardPeign CCETTFrance
Jean-Bernard RaultCCETTFrance
YonDominique CISACFrance
NasseDominique France TelecomFrance
OlivierAvaro France Telecom CnettFrance
SereeGourrier Lab d'Electronics PhilipsFrance
SylvieJeannin Lab d'Electronics PhilipsFrance
BolichardLionel Philips LEPFrance
EdouardFrancois Thomson MultimediaFrance
KlausDiepold Betatechnik GMBHGermany
SvenBauer BoschGermany
GunnarNitsche BoschGermany
PeterList Deutche Telecom AGGermany
CarstenHerpel Deutche Thomson-BrandtGermany
FrankFeige Deutsche Telekom AGGermany
JensSpille Deutsche Thomson-BrandtGermany
KarlheinzBrandenburg FHG-11SGermany
JrgenHerre FUG-11SGermany
JanDeLameillieur Heinrich-Hertz-InstitutGermany
ThomasSikora Heinrich-Hertz-InstitutGermany
BennoStabernack Heinrich-Hertz-Institut-BerlinGermany
JurgenRuf Media TransferGermany
Dr. GeraldKnabe Q-Team Dr. Knabe GmbhGermany
WolfgangHiehsen RWTH AachenGermany
FritzSeytter SiemensGermany
Dr. Andr KaupSiemens AGGermany
Dr. Christoph HeerSiemens AG ZT ME 3Germany
PeterKuhn Technical University of MunichGermany
AndreasHutter Tu MunichGermany
JochenHammerschmidt TV MunichGermany
MarkusZeller TV-MunchenGermany
ThomasWiegand University ErlangenGermany
BernhardGrill University of ErlangenGermany
UlrichBenzler University of HannoverGermany
BerndEdler University of HannoverGermany
PeterGerken University of HannoverGermany
BrunoNagel University of HannoverGermany
NoelBrady Teltec IrelandIreland
LiamWard Teltec IrelandIreland
YairBe`ery Tel Aviv University & VDOnetIsrael
EugeneTaits VDOnet CorporationIsrael
Dr. IlanSadeh Visnet Ltd.Israel
AharonGill Zoran Microelectronics Ltd.Israel
GiorgioParladori Alcatel TelecomItaly
RicardoPannocchia Consorzio Pisa RicercheItaly
StefanoBattista CSELTItaly
LucaCellario CSELTItaly
LeonardoChiariglione CSELTItaly
LauraCont!n CSELTItaly
MarcoGandini CSELTItaly
ClaudioLande CSELT s.p.a.Italy
RobertoPockaj Dist-University of GenovaItaly
GiuseppeRusso Fondazione Ugo BordoniItaly
VittorioBaroncini FUBItaly
GiorgioDimino RAI Radiotelevisione ItalianaItaly
EishiMorimatsu Fujitsu Laboratories Ltd.Japan
YasufumiNakamura FUJITSU LTDJapan
ItaruKaneko Graphics Communications LabJapan
JunichiKimura HitachiJapan
YuichiroNakaya HitachiJapan
NarumiHirose IPSJ/ITSCJJapan
AndoIchiro JVCJapan
KojiImura Matsushita Communication Ind.Japan
YutakaMachida Matsushita Communication Ind.Japan
NaoyaTanaka Matsushita Communication Ind.Japan
KojiYoshida Matsushita Communication Ind.Japan
TakeshiNorimatsu Matsushita Electric Ind. Co.Japan
Choong Sang BoonMatsushita Electric Industrial Co. Japan
YoshinoriMatsui Matsushita Electric Industrial Co.Japan
TaisukeMatsumoto Matsushita Research Inst. TokyoJapan
KotaroAsai Mitsubishi Electric CorporationJapan
Shun-ichiSekiguchi Mitsubishi Electric CorporationJapan
MasahiroIwadare NECJapan
ToshiyukiNomura NECJapan
HitoshiKoyama NEC CorporationJapan
YoshihiroMiyamoto NEC CorporationJapan
MusumiOhta NEC CorporationJapan
EisukeNakasu NHKJapan
ShinichiSakaida NHK (Japan Broadcasting Corp.)Japan
KinyaOosa Nippon Steel CorporationJapan
NaokiIwakami NTTJapan
HirohisaJozawa NTTJapan
MasahisaKawashima NTTJapan
TakehiroMoriya NTTJapan
YasuhiroTomita NTTJapan
SanaeHotani NTT DoCoMoJapan
ToshioMiki NTT DoCoMoJapan
ShinyaNakajima NTT Human Interface LabsJapan
ToshihisaNakai OKI Electric Ind. Co. LtdJapan
ShigeruFukunaga OKI Electric Ind. Co. Ltd.Japan
YasukoMatsumura OKI Electric Ind. Co. Ltd.Japan
ZhixiongWu Oki Electric Industry Co. Ltd.Japan
HiroyukiKatata Sharp Co.Japan
ShuichiWatanabe Sharp CompanyJapan
KeiichiHibi Sharp CorporationJapan
KazuyukiIijima Sony Corp. Architecture LabsJapan
KenzoAkagiri Sony CorporationJapan
KazuhisaHosaka Sony CorporationJapan
TakashiKoike Sony CorporationJapan
MasayukiNishiguchi Sony CorporationJapan
MasamiOgata Sony CorporationJapan
YoshiakiOikawa Sony CorporationJapan
TeruhikoSuzuki Sony CorporationJapan
YoichiYagasaki Sony CorporationJapan
YujiItoh Texas Instruments Tsukuba R&D Ctr. Japan
YoshihiroKikuchi ToshibaJapan
ToshiakiWatanabe ToshibaJapan
AtsutoMaki Toshiba Co.Japan
HirokazuTanaka Toshiba CorporationJapan
EijiKasutani Waseda UniversityJapan
HisashiMiyamori Waseda UniversityJapan
ShigekiFujii Yamaha CorporationJapan
ByungCheol Sung
Jong Il-lain
Jin HunKim Daewood Electronics Co. Ltd.Korea
ChieteukAhn Electronics/Telecom Res. Inst.Korea
Young-KwonLim Electronics/Telecom Res. Inst.Korea
Joo-heeMoon Hyundai ElectronicsKorea
Vong JungKang KAISTKorea
Yung-LyulLee KAISTKorea
SangheeLee KaistKorea
Sung DeukKim Korea Adv. Inst. Science & Tech.Korea
YoungSu Lee LG SemiconKorea
Hyun MunKim LG Semicon Co.Korea
Euee S.Jang Samsung AITKorea
SangwookKim Samsung AITKorea
Jae-SeobShin Samsung AITKorea
Sung-hooHan Samsung ElectronicsKorea
Hoon-JaeKim Samsung ElectronicsKorea
RobKoenen KPN ResearchNetherlands
CorQuist KPN ResearchNetherlands
A.W.G.Oomen Philips ResearchNetherlands
R.Taori Philips ResearchNetherlands
Leonvan de Kerkhof Philips S & VNetherlands
RobertDanielsen TelenorNorway
Jan Eikeset KnudsenTelenor R & D Norway
Anibal J.F. De Sousa FerreiraINESCPortugal
Maria Isabel MartinsINESCPortugal
FernandoPereira Instituto Superior TecnicoPortugal
Sheng-MeiShen Panasonic Singapore LaboratoriesSingapore
Ah-PengTan Panasonic Singapore LaboratoriesSingapore
Thiow KengTan Panasonic Singapore LaboratoriesSingapore
LucasHui SGS-Thomson MicroelectronicsSingapore
Jeong SuKim Samsung AITKorea
JeongSu Kim Samsung AITKorea
LuisTorres Polytechnic University CataloniaSpain
EnriqueLoras SGAESpain
EduardoBallesteros TelefonicaSpain
FerranMarques UPCSpain
PerThorell EricssonSweden
GranRoth Ericsson RadioSweden
MorganLindqvist Ericsson Radio SystemsSweden
BoBurman Ericsson Radio Systems ABSweden
Torbjrn EinarssonEricsson Telecom Sweden
RickardSjöberg Ericsson TelecomSweden
RobertForcheimer Linkoping UniversitySweden
JacobStröm Linkoping UniversitySweden
Jörgen AhlbergLinkping University Sweden
HaiboLi Linkping UniversitySweden
MichaelRanden Telia ResearchSweden
Matsgren Telia Research ABSweden
OlleFranceschi TeracomSweden
SofieOlsson TeracomSweden
NilsSundström Teracom ABSweden
MehrdadAhmadi TerecomSweden
MarcoMattavelli Ecole Polytechnique FederaleSwitzerland
TolgaCapin EPFLSwitzerland
TouradjEbrahimi EPFLSwitzerland
PandzicIgor University of GenevaSwitzerland
IgorPandzic University of GenevaSwitzerland
David JMeares British Broadcasting Corp.UK
Dr. Theodore VlachosBritish Broadcasting Corporation UK
BillWelsh BTUK
StephenMinnis BT LabsUK
Dr. ColinDavies DMV Ltd.UK
AlbertoDuenas DMV LTD.UK
Dr. CliveStubbings DMV Ltd.UK
PaulSmith LSJ Logic Europe Ltd.UK
JohnMagill Lucent Technologies UK Ltd.UK
Dr. Thomas J. Caseytcasey Ltd.UK
AtulPuri AT&TUSA
SchuylerQuackenbush AT&TUSA
BarryHaskell AT&T LabsUSA
JoernOsterman AT&T Labs-ResearchUSA
EricPetajan Bell Labs Lucent TechnologiesUSA
AlexandrosEleftheriadis Columbia UniversityUSA
YihanFang Columbia UniversityUSA
TihaoChiang David Sarnoff Research CenterUSA
HeidiPeterson David Sarnoff Research CenterUSA
IrajSodagur David Sarnoff Research CenterUSA
Ya-QinZhang David Sarnoff Research CenterUSA
RichardSchaphorst Delta Information SystemsUSA
FredericDufaux Digital Equipment CorporationUSA
DonHooper Digital Equipment CorporationUSA
MatthewGoldman Divi ComUSA
MarinaBosi Dolby LaboratoriesUSA
AlexanderLoui Eastman Kodak CompanyUSA
BobSenn Eastman Kodak CompanyUSA
Peter K.Doenges Evans & SutherlandUSA
MikeColeman Five Bats ResearchUSA
BobEifrig General InstrumentUSA
SamNarasimhan General InstrumentUSA
XueminChen General Instrument CorporationUSA
GaneshRajan General Instrument CorporationUSA
JimWilliams Hughes (Directv)USA
PeterAu Hughes ElectronicsUSA
Ron Burns Hughes ElectronicsUSA
Donald Mead Hughes ElectronicsUSA
RamNagarajan Hughes ElectronicsUSA
Chao-KungYang Hughes Electronics-DirectvUSA
Cesar A.Gonzales IBMUSA
PeterWesterink IBMUSA
PeteSchirling IBM Video Enabled SolutionsUSA
RohitAgarwal Intel CorporationUSA
Thomas R.Gardos Intel CorporationUSA
ChrisHansen Intel CorporationUSA
MichaelKeith Intel CorporationUSA
Juande la Cierva Interaxx Television NetworkUSA
John H.Muller Iterated SystemsUSA
Michael R.Zeug Iterated Systems Inc.USA
WeipingLi Lehigh UniversityUSA
AndrewTescher Lockheed MartinUSA
CasparHorne Mediamatics Inc.USA
Wei-geChew MicrosoftUSA
Ming-ChiehLee MicrosoftUSA
EricSchreier MITUSA
BennyVenco MITUSA
HuifangSun Mitsubishi ElectricUSA
DavidThom Mitsubishi Electronics AmericaUSA
CheungAuyeung MotorolaUSA
JamesBrailean MotorolaUSA
JiroKatto NEC Corporation/Princeton Univ.USA
KwokChau NJR CorporationUSA
VaheBalabanian NortelUSA
MichaelSzwarc NYNEXUSA
JeffNeedham OracleUSA
HayderRadha Philips ResearchUSA
Homer H.Chen RockwellUSA
Peter G.Schreiner III Scientific AtlantaUSA
TimAddington Scientific-AtlantaUSA
Si JunHuang Scientific-AtlantaUSA
JordiRibas Sharp LaboratoriesUSA
Dr. Regis J. CrinonSharp Laboratories of America USA
ShawminLei Sharp Labs of AmericaUSA
IbrahimSezan Sharp Labs of AmericaUSA
Chien-MinHuang Sorenson Vision Inc.USA
KyminhLiang Sorenson Vision Inc.USA
PallaviShah Sun MicrosystemsUSA
JamesVanLoo Sun MicrosystemsUSA
Dr. SinanOthman Team LogicUSA
AliTabatabai TektronixUSA
ChuckLueck Texas InstrumentsUSA
RajTalluri Texas InstrumentsUSA
FritzWhittington Texas InstrumentsUSA
IoleMoccagatta Texas Instruments Inc.USA
RalfNeff U.C. BerkeleyUSA
OsamaAlshaykh University of California-BerkeleyUSA
A. MuratTekalp University of RochesterUSA
AlbertSimmonds US ISBN AgencyUSA
AlbertSimonds US ISBN AgencyUSA
Yousef W.Nijim Vela ResearchUSA
IrvingRobowsky Worldwide Satellite BroadcastingUSA

1. Opening
2. Roll call of participants
3. Approval of agenda
4. Allocation of contributions
5. Communications from Convenor
6. Report of previous meeting
7. Processing of NB Position Papers
8. MPEG Phase 2
8.2 Audio
8.3Verification of MPEG-2
8.4.1Private Data (System #3)
8.4.2Semantic & table def. f. 4:2:2 and MV profiles
8.4.3ITU-T extension
8.5Part 7 (NBC Audio)
8.6Part 10 (DSM-CC Conformance)
9.MPEG Phase 4
9.1 Requirements
9.2 Tools
9.2.3Natural Video
9.2.4Synthetic Video
9.2.5Natural Audio
9.2.6Synthetic Audio
9.MPEG Phase 4 (cont. 1)
9.3 Verification Models
9.4 Call for proposals
9.5 Tests
9.6 Simulation software
9.MPEG Phase 4 (cont. 2)
9.7 Working Draft
9.8 Workplan
10.MPEG Phase 7
11. Overall WG11 workplan
12. Liaison matters
13. Administrative matters
13.1 Schedule of future MPEG meetings
13.2Promotion of MPEG
14. Organisation of this meeting
14.1 Tasks for subgroups
14.2 Finalisation of meeting allocation
15. Planning of future activities
16. Resolutions of this meeting
17. A.O.B
18. Closing

Annex 3

List of input documents
1596Pete SchirlingContribution Register for 38th meeting in Sevilla, Spain
1597Patrick Mulroy, Bill Welsh, Andy Breen Face Animation and TTS files for SNHC Demonstrator
1598TK TanTables of Bitrate and PSNR for I-VOP with DC/AC Prediction
1599Minhua ZhouVariable Length Coding of DQUANT
1600Minhua ZhouAn improved adaptive selection of scanning path for DC/AC prediction
1601T.K. Tan, S.M. Shen Results of Core Experiment T16
1602Franco Casalino, Matthew Leditschke, Mauro Quaglia MPEG-4 PC System Architecture Overview
1603Marina BosiReport of the Ad Hoc Group on MPEG-2 AAC DIS 13818-7 Progression
1604T. Alpert, L. Contin DSCQE experiment for the evaluation of the MPEG4 VM on error robstness functionality
1605Richard J. Qian, M. Ibrahim Sezan Generalized Content-Scalable Shape Coding
1606Pasi OjalaResults of the core experiment of NEC compression algorithm proposal for MPEG-4/Audio
1607C. Le Buhan Jordan, T. Ebrahimi Inherently progressive polygonal contour coding
1608SC 29 Secretariat Summary of Voting on ISO/IEC 13818-1/PDAM 4
1609SC 29 Secretariat Summary of Voting on ISO/IEC 13818-1/DCOR 1
1610SC 29 Secretariat Summary of Voting on ISO/IEC DIS 13818-3 (Second Edition)
1611T.K. Tan, S.M. Shen Truncated Quantization Matrix and Adaptive DC Quantization Scaling
1612T.K. Tan, S.M. Shen Deblocking Filter
1613S.M. Shen, T.K. Tan New Results of Merged Mode for Spatial Scalability
1614S.M. Shen, T.K. Tan Merged Mode Applied to ROI for Spatial Scalability
1615David Thom - MELA, Mike Coleman - Five Bats Ad Hoc Group Report on MPEG-2 AAC Conformance Testing & Technical Report
1616Caspar Horne, Ganesh Rajan Report of the AdHoc Group on SNHC VM editing
1617Caspar Horne, Ganesh Rajan SNHC Verification Model 2.1
1618Wei Wu, Homer Chen, Hai Tao Rockwell C Implementation of SNHC FAP coding
1619A. Eleftheriadis, H. Kalva A Proposed Architecture for an Object-Based Audio-Visual Bitstream and Terminal
1620Kevin BarronVRML node inclusion in MSDL model
1621Noel BradyResults of Shape Coding CE S4F
1622Raj Talluri, Iole Moccagatta, Yashoda Nag E8 - Core Experiment on Error Concealment by Data Partitioning
1623Ram NagarajanReport of Ad Hoc Group on the Media Integration of Text and Graphics
1624Andre Kaup, Anke Lorenz A Modified SA-DCT with Improved Coding Efficiency
1625S. R. Quackenbush Report of the Ad-hoc Group on Reducing AAC Prediction Complexity
1626John MullerReport of the Adhoc group on coding efficiency in MPEG-4 video
1627A. M. Tekalp, P. J. L. van Beek Description of CE M2 (Triangular mesh-based coding with object-based functionalities)
1628A. M. Tekalp, P. J. L. van Beek Results of CE M2 (Triangular mesh-based coding with object-based functionalities)
1629A. M. Tekalp, P. J. L. van Beek, C. Toklu Description of CE X4 (2D Mesh-based texture mapping for synthetic and natural content manipulation)
1630A. M. Tekalp, P. J. L. van Beek Results of CE X4 (2D Mesh-based texture mapping for synthetic and natural content manipulation)
1631Anthony Vetro, Huifang Sun Joint Rate Control for Coding Multiple Video Objects Based on Quadratic Rate-Distortion Model
1632V. BaronciniProposal to host the July '97 MPEG-4 Tests
1633G. Russo, S. Colonnese, U. Mascia, C. Tabacco New FUB Results on Core Experiment N2 on Automatic Segmentation Techniques
1634Marco Mattavelli Report of the ad-hoc group on Computational Graceful Degradation
1635Koji Imura, Yutaka Machida The Results of the Core Experiment E3b(HEC)
1636Marco Mattavelli, Sylvain Brunetton Draft proposal for the specification of the ISG Computational Graceful Degradation core experiment
1637Pierre COMBELLES, Dominique CURET protection policy within the MPEG4 multiplex1
1638Pierre COMBELLES, Dominique CURET one question about channel coding efficiency
1639Jean-Claude Dufourd Application Study: Interaction with a Multimedia Database
1640Shi-Hwa Lee,Dae-Sung Cho,Se-Hoon Son,Euee-Seon Jang,Yu-Shin Cho,Seong-Jin Kim,Jae-Seob Shin Results of Core Experiment S4
1641Sung-Gul Ryoo,Jae-Seob Shin,Euee-Seon Jang Results of Core Experiment Q2
1642Seok-Yoon Jung,Euee-Seon Jang,Jae-Seob Shin A New Gray-scale shape coding technique
1643Y.B. Thoman Kim, S.H. Park A Proposal for the MPEG-4 Audio T/F based scalable codec test
1644Toshiaki Watanabe, Yoshihiro Kikuchi Comparison of Binary Shape Coding (Core Experiment S4d)
1645Toshiaki Watanabe, Yoshihiro Kikuchi Scalability for Modified M4R
1646Toshiaki Watanabe, Yoshihiro Kikuchi Error Resilient Performance of Header Extension Code (Core Experiment E3b)
1647Toshiaki Watanabe, Yoshihiro Kikuchi Error Resilient Performance of EREC with RVLC (Core Experiment E7)
1648Hirokazu TanakaError control scheme for a multiplexed frame header
1649Bill Welsh, Andy Breen, Steve Minnis Description of Synthetic Talking Head Demonstrator for SNHC
1650Taisuke MATSUMOTO, Sinya KADONO, C.S.Boon Results on Shape Core Experiment S4d (Improved M4R Method)
1651Minhua Zhou,, Jan De Lameillieure HHI report on Core Experiment P9
1652Minhua Zhou,, Jan De Lameillieure HHI report on Core Experiment T3

1654Peter H. AuCore Experiments Report on SNHC-MITG
1655Peter H. AuCore Experiments Report on N3 - Low Latency Coding of Static Sprites

1657Xuemin Chen, Bob Eifrig, Ajay Luthra Rate Control for Multiple Higher Resolution VOs : A Report on CE Q2


1660Bernhard Grill, Heiko Purnhagen Report of the AhG on MPEG-4 Audio Working Draft Editing and VM Software Implementation
1661Y. Matsui, C. S. Boon Content Multiplex Syntax
1662Yoshinori Matsui, C. S. Boon Content Multiplex Syntax

1664Norio Ito, Hiroyuki Katata A report of core experiment T1 (Wavelet coding of I pictures)
1665Norio Ito, Hiroyuki Katata A report of core experiment T1 (Wavelet coding of I pictures)
1666Hiroyuki Katata, Ryuji Kitaura, Hisashi Saiga Proposal of the refined syntax for Object-based temporal scalablity
1667Tomoko Aono, Hisashi Saiga, Hiroyuki Katata Spatial scalability with fine granularity
1668Shuichi Watanabe Continuous scalability application in MPEG-4 system level
1669Keiichi Hibi,Nobuyuki Ema Report of core experiment on P6 "Multimode Warping Prediction"
1670Eishi Morimatsu, Akira Nakagawa Results of Core Experiment P13(Dynamic Resolution Conversion)
1671Eishi Morimatsu, Akira Nakagawa Results of Core Experiment Q2(Improved Rate Control)




1676Keiichi HIBI,Nobuyuki EMA Report of core experiment on P6 "Multimode Warping Prediction"
1677Naoya Tanaka, Koji Yoshida Results of the Core Experiment m1509
1678Keiichi HIBI, Tadashi UCHIUMI, Seiji SATO Proposal of Improvement on E8 Core Experiment (with Texture Data Partitioning)


1681Yoshinori Matsui, C. S. Boon Content Multiplex Syntax
1682Yoshinori Matsui, C. S. Boon Content Multiplex Syntax
1683Yoshinori Matsui, C. S. Boon Content Multiplex Syntax

1685Hirohisa Jozawa, Kazuto Kamikura Results of Core Experiment P1
1686Hirohisa Jozawa, Kazuto Kamikura Results of Core Experiment P1
1687Hirohisa Jozawa, Kazuto Kamikura Modified OBMC for Sprite/GMC prediction
1688Hirohisa Jozawa, Kazuto Kamikura Modified Bitstream Syntax for Sprite Coding
1689Eishi Morimatsu, Akira Nakagawa Proposal of the Simplified Dynamic Resolution Conversion of Prediction Error(Simplified DRC-PE) scheme
1690B. DeKnuydt, S. Desmet, L. Van Gool, L. Van Eycken Coding of dynamic Texture: problem and example


1693G.FranceschiniInternet and User-Network signaling integration
1694G.FranceschiniDMIF directions for User-User signaling
1695G.FranceschiniElementary Stream to Adaptation Layer interface and Adaptation Layer Headers

1697Masayuki Nishiguchi, Kazuyuki Iijima, Jun Matsumoto Bit rate controllable/scalable tools for the parametric speech coder core of the VM
1698Angelika KnollDescription of the multiplex solftware of DTAG
1699Dominique Yon, CISAC Report of ad hoc group on content-related IPR
1700Michael ZeugReport of Ad Hoc Group on MPEG4 Low Delay Evaluations

1702Atul PuriReport of Ad hoc Group on Multifunctional Coding in MPEG-4 Video
1703A. Puri, R. L. Schmidt, B. G. Haskell Results and description of coding efficiency modified experiment T16 in MPEG-4 Video
1704A. Puri, R. L. Schmidt, B. G. Haskell KeyRegions: Improved slices for MPEG-4 Video

1706Luis Torres, David Gimeno, David Garcia Results of core experiment X1

1708Dean Messing, M. Ibrahim Sezan Result of X4 for Augmented Reality
1709Peter H. AuCore Experiments Report on N3 - Low Latency Coding of Static Sprites
1710Peter H. AuCore Experiments Report on N3 - Low Latency Coding of Static Sprites

1712Bob Eifrig, Xeumin Chen Interlaced Tools for MPEG4 Video
1713Dong-Jung Kang, In-So Kweon Results of Core Experiment Q2
1714C.S. Boon, J. Takahashi, K. Horiike Experiment Results of Core Experiment O2
1715C.S. Boon, T. Nishi, S. Kadono Result of Core Experiment S14 (Quantization of Gray-scale Alpha Plane)
1716C.S. Boon, T. Nishi, S. Kadono Selection of Motion Vectors for Motion Compensation of Grayscale Alpha Plane : A New Proposal
1717C.S. Boon, J. Takahashi Comments on Video Verification Model
1718Toshihisa Nakai,, Shigeru Fukunaga,, Yasuhiro Tomita Error Resilient Video Coding using Backward Channel Signaling
1719Toshihisa Nakai, Yasuko Matsumura Proposal of Reversed Ordering of RVLC for Error Resilient Video
1720Touradj Ebrahimi (editor) Video Verification Model 5.1
1721Touradj Ebrahimi, Atul Puri Report of Ad hoc group on editing video VM and WD
1722Touradj Ebrahimi Report of Ad hoc group on SNHC texture coding
1723Touradj Ebrahimi Description of core experiments in SNHC texture coding
1724Jin-Hun Kim, Jong-il Kim, Sang Hoon Lee Results of Core Experiment S4 - Vertex Based
1725Karlheinz Brandenburg Report of the Ad-hoc Group on core experiments for MPEG-4 audio
1726Laura ContinReport of the ad hoc group on MPEG-4 July 97 audio, video and audio-visual Tests.
1727Masami Ogata, Andrew Mackinnon Results of Core Experiment T1:wavelet coding of I picture
1728Jun Matsumoto, Masayuki Nishiguchi, Kazuyuki Iijima Core experiment results in Sony for NEC proposal
1729Itaru Kaneko/GCL, Shinya Nakajima/NTT Chairman's report of Ad-hoc group on the SNHC Audio
1730Itaru Kaneko/GCL A proposal for content related IPR's
1731Toshio Miki, Satoru Adachi, Tomoyuki Ohya Results of Core Experiments on Error Resilience (E7)
1732Toshio Miki, Satoru Adachi, Tomoyuki Ohya Comparison of Coding Efficiency Between VM VLC and ERVLC (CE7)
1733The National Body of Japan Requests on MPEG document handling
1734John Arnold, Michael Frater, Martin Kuchlmayr Results of Core Experiment T15
1735Teruhiko Suzuki, Yoichi Yagasaki The results of CE B1.1
1736K. Hosaka, Y. Yagasaki Results of CE S4f(CAE)
1737Fred Jordan, Stefan Horbelt, Touradj Ebrahimi Results of Core Experiment X3 for Texture Coding
1738Julien SignesReport of the adhoc group on MPEG-4 Systems evolution
1739Minsoo Hahn, HangSeop Lee, Jae-Woo Yang Revision of MPEG-4 TTS Interface
1740Sofie OlssonMPEG-4 Broadcast Profile : Revision and Comments
1741Julien Signes, Dominique Curet, Bernard Peigne An example of audio and 2D/3D content manipulation application: Multiplex, 3D composition and distant interactivity issues
1742Christian Stoffers Results of Core experiment P6
1743Jae Gark Choi, Munchurl Kim, Myoung Ho Lee, Jisang Yoo Automatic segmentation based on spatio-temporal information
1744Jin-Soo Choi, Myoung Ho Lee, Chieteuk Ahn Geometry compression using predictive vector quantization
1745Young-Kwon Lim, Sanggyu Park, Munchurl Kim, Chieteuk Ahn, Yong Han Kim (Seoul City University) Proposal for the ISG core experiments on Computational Graceful Degradation in multiple VOs case
1746Frank BossenResults of CE S4f on binary shape coding (CAE)
1747Sung Deuk Kim, Jaeyoun Yi, Jong Beom Ra (KAIST),, Hyun Mun Kim, Young Su Lee (LGS) A New Blocking Artifact Reduction Scheme for MPEG-4 Video Coding(N1)
1748Sung Deuk Kim, Jong Beom Ra(KAIST),, Hyun Mun Kim, Young Su Lee(LGS) Improved Motion Vector Coding Scheme
1749Thomas wiegand, Hans-Juergen Leicht, Klaus Stuhlmueller Results of Core Experiment T3 (Matching Pursuit Coding of Prediction Errors)
1750Byung Cheol Song, Jong Beom Ra(KAIST),, Hyun Mun Kim, Young Su Lee(LGS) A Variable Block Size Motion Estimation by Adaptively Using Bilinear Interpolation
1751Klaus Stuhlmueller, Thomas wiegand Results of Core Experiment P9 (Motion-Compensated Prediction Using Quadtree Segmentation and Polynomial Motion Fields)
1752Carsten HERPELReport of adhoc group on Systems multiplex specification and signaling
1753Kenzo Akagiri, Takashi Koike A description of the MPEG-2 AAC compliant pre-/post- processing tools for the MPEG-4 VM framework
1754Carsten HERPEL, James VAN LOO Report of adhoc group on MPEG-4 signaling
1755Rakesh TaoriProposal for new features in the MPEG4 wideband coder: bit-rate scalability and complexity scalability
1756Rob Koenen, Laura Contin Registrations for July 97 Tests
1757Rob KoenenReport of ad hoc group on MPEG-4 Requirements
1758Rob KoenenDraft Revision of MPEG-4 Requirements Document
1759Rob KoenenDraft Revision of MPEG-4 Profiles Document
1760Akio Jin, Naoki Iwakami, Takehiro Moriya, Satoshi Miki, Kazunaga Ikeda Proposal of bit-rate scalability tools for MPEG-4/audio VM
1761Yoshihiro Miyamoto Results of Core Experiment P6 (Multimode Warping Prediction)
1762Ryoma Oami, Mutsumi Ohta Comparison of Wavelet Coding for Core Experiment T1
1763Akio Yamada, Mutsumi Ohta, Takahiro Kimoto, Ryoma Oami Comparison of Block-based Encoding Methods for Wavelet Coefficients
1764Akio Yamada, Mutsumi Ohta, Ryoma Oami, Yoshihiro Miyamoto An Adaptive Scanning Methods for Runlength Wavelet Coding
1765Gerhard StollReport of Ad-Hoc Group on Audio FAQ for the MPEG Home Page
1766Yuji ItohResults of Core Experiment N1
1767Yuji ItohUniversal VLC for MV and DCT coefficient
1768Yuji ItohPreliminary results of Universal VLC
1769Yuji ItohSyntax for Data Partitioning
1770I. CorsetResults on T3 : matching pursuit inter mode coding
1771James Irwin, Ralf Schafer MPEG-4 Terminal Architecture, Introduction
1772Y. Suzuki, Y. Nakaya, S. Misaka Results of Core Experiment P1 (Global Motion Compensation)
1773Y. Nakaya, Y. Suzuki, S. Misaka Results of core experiment P6 (Multimode warping prediction)
1774Christian Bouville, Julien Signes 2D/3D Scene Modelling and Presentation
1775Phillipe Schmid, Gloria Menegaz Results of core experiment X1 for efficient coding of homogeneous textures
1776Eddie Cooke, Christian Bouville Proposal for 2D & 3D Composition
1777Eddie Cooke, Liam Ward Proposal for 3D Scene Description and Composition
1778Systems VM AHGSystems VM 2.0 (Editor's revision)
1779Fabio Bellifemine, Stefano Battista Systems VM software development: Netsccs tutorial
1780Paul HaskellMPEG-2 Systems Test Bitstream Status
1781Cheung AuyeungResults of Core Experiment T14
1782Haibo Li, Robert Forchheimer, Jorgen Ahlberg, Jacob Strom Efficient Coding of Facial Texture Result of Core Experiment X2
1783Frank BossenReducing the cost of initialization and flushing of the arithmetic coder for CAE
1784Frank BossenEfficient representation of triangulations
1785Ming-Chieh Lee, Wei-ge Chen, Chuang Gu Microsoft Software C++ Implementation for MPEG-4 Video VM 5.x
1786Chuang Gu, Simon Winder, Ming-Chieh Lee Results Report on O2 -- Padding for Motion Compensation
1787Ming-Chieh Lee, Chuang Gu Results Report on N3 -- Low Latency Video Coding with Sprites
1788Eric D. Scheirer, Barry Vercoe Synthetic Audio Specification Language Proposal for SNHC Audio
1789Doris Bittruf, Thomas Stingl, Yankin Tanurhan, Alexandra Winter Events in MPEG-4 Systemarchitecture
1790Christian Koechling, Gunnar Nitsche MSDL VM and Audio
1791Andreas HutterConsiderations on the video input library for the July `97 tests
1792Ganesh Rajan, Sati Banerjee Results of CE X4 (2D Mesh Based Texture Mapping for Synthetic and Natural Content Manipulation
1793Ganesh RajanResults of CE X3 (View Dependent Texture Coding for Transmission of Virtual Environments)
1794Ganesh RajanResults of Core Experiments on Media Integration of Text and Graphics
1795Wei-ge Chen, Ming-Chieh Lee Results Report on S14 -- Quantization of Gray-scale Alpha Plane
1796Jie LiangResults Report on Core Experiment T1: Wavelet COding of I Pictures
1797Anthony Vetro, Huifang Sun Results of Core Experiment Q2: comparsion of VM and TMN6 rate control
1798Jie Liang, Iole Moccagatta Results Report on Core Experiment X5: Efficient Coding of Static SNHC Texture to Be Mapped in 2D/3D Surface Models
1799Vahe BalabanianPreliminary answers to the DSM-CC FAQ
1800Vahe BalabanianPotential Joint MPEG-4 Systems and DMIF Activties
1801Pete SchirlingIEEE 1180 Standard Reinstatement Process
1802Vahe BalabanianLiaison reply to DMIF from Q2 ATM & Other networks experts Group
1803Vahe BalabanianLiaison reply from ITU-T SG16, Rapporteur Group for Q3/16 (T.130)
1804Vahe BalabanianFirst cut at networking SRMs in DAVIC
1805Vahe BalabanianProposal to DAVIC for access to DAVIC Services from within Internet
1806Vahe BalabanianProposal to DAVIC for use of DAVIC Delivery to access Internet Multimedia Tools
1807Vahe BalabanianProposal to establish a DMIF reflector
1808Eric Petajan, Tolga Capin Ad Hoc Group Report on Face and Body Animation
1809Chris HansenProposed MPEG4 Transport Mux Interface to ITU-T H.series Terminals
1810Paul FellowsN1343 Ad-Hoc Group report
1811Frederic Dufaux, Regis J. Crinon Update Results Regarding N3 Core Experiment
1812Jordi Ribas-Corbera, Shawmin Lei Contribution to rate control Q2 experiment: A quantizer control tool for achieving target bit rates accurately
1813JavasoftJava Media Framework
1814Simon WinderReport of the Ad-Hoc Group on Investigating Reduced Complexity Padding Techniques
1815Bernhard Grill, Bodo Teichmann Additional Block Sizes for AAC-based Coding in MPEG-4
1816Bernhard GrillImplementation of a Scalable Audio Coder based on AAC and a Core Coder
1817Ralph Neff, Eugene Miloslavsky, Osama Alshaykh,, Shang-Pin Chang, Emin Martinian, Avideh Zakhor Experiment T3: Matching Pursuit Coding of Prediction Errors
1818Ram NagarajanProposal for Core Experiments in SNHC
1819Yihan Fang, Alexandros Eleftheriadis MPEG-4 Video WD 1.0 Syntax using Flavor (MSDL-S)
1820Y. Fang, A. Eleftheriadis Flavor (MSDL-S) Translator to C++ Version 1.2a
1821Jvrn OstermannReport of the Ad-Hoc group on core experiments on coding of arbitrarily shaped objects in MPEG-4 video
1822Jörn Ostermann, Atul Puri Shape coding for interlaced video coding
1823James BraileanError Resilience AHG Report
1824Yung-Lyul Lee, Yoon-Soo Kim, Hyun-Wook Park Results of Core Experiment N1(Coding Noise Removal Techniques)
1825Jiro KattoResults of Core Experiment M1
1826Bruno Nagel, Lutz Falkenhagen, Andreas Graffunder A new 3D graphics interface using general scene initialisation and update parameters
1827Ulrich BenzlerProposal for a new Core Experiment on prediction enhancement at higher bitrates
1828Jens Spille, Carsten Herpel Definition of Audio Objects and proposed changes to the MPEG-4 Requirements
1829Sang-Woon Seo, Yung-Lyul Lee, Hyun-Wook Park Results of Core Experiments S4e
1830Lin Yin, Mauri Vaananen Proposal for a Core Experiment on Backward Prediction in MPEG-4 Audio
1831Peter Gerken, Roland Mech Automatic Segmentation of Moving Objects (Partial Results of Core Experiment N2)
1832Marta Karczewicz Proposal for efficient coding
1833Marta Karczewicz Report on Core Experiment P9
1834Marta Karczewicz Description of Core Experiment P9
1835Nadia Magnenat Thalmann, Daniel Thalmann, Igor Pandzic,, Tolga Capin, Marc Escher, Ronan Boulic, Christial Miccio,, Tom Molet MIRALab/LIG contributions to SNHC in the project VIDAS
1836Klaus Stuhlmueller, Thomas Wiegand Rate-Distortion Curves for Core Experiment P9
1837Peter KuhnResults of Core Experiment P9 (Motion Compensated Prediction using Quadtree Segmentation and Polynomial Motion Fields)
1838Olivier AVAROReport of the AdHoc Group on Systems Working Draft Editing
1839Tolga K. Capin, Ronan Boulic, Joaquim Esmerado, Tom Molet Proposal for Update of Body Animation Section in the SNHC-VM
1840Olivier AvaroReport of the AdHoc Group on Systems Working Draft Editing
1841Joo-hee Moon, Ji-heon Kweon Results of Core Experiment O10 - Boundary Block Merging Technique
1842Carsten HERPELElementary Stream Naming Convention for 2 Layer Mux
1843Carsten HERPELMux2 (FlexMux) Layer Design Considerations
1844AVARO OlivierSystems WD 2.0 (Editor's revision)
1845Carsten HERPELError Resilience of the Mux2 (FlexMux) Layer
1846Carsten HERPELFraming ('Synchronization') Mechanisms in MPEG-4
1847Carsten HERPELTowards Mux1 (TransMux) Interfaces to various Transport Protocols
1848Joern Ostermann, Marc Beutnagel Comments on the current definition of TTS in SNHC
1849MoMuSys partners MoMuSys C implementation of MPEG-4 video VM 5.0
1850Robert Danielsen Results from CE T14
1851USNBUSNB Comments on 13818-1 DAM 3
1852USNBUSNB Comments on 13818-2 DAM 4
1853USNBUSNB Comments on DIS 13818-7
1854USNBUSNB Contribution re 13818-4 and 13818-5
1855USNBUSNB Contribution re MPEG-2 AAC SRS profile
1856Jae-won Chung, Jae-kyoon Kim, Joo-hee Moon Results of Core Experiment S4h : Geometric Representation Method
1857Wei Wu, Homer Chen Results of T15: N-Bit Video Coding
1858Wei Wu, Homer Chen Results of X3: View-Dependent Textture Coding
1859Jim Scholl, Wei Wu, Homer Chen Results of T1: Intra Wavelet Coding
1860Jim Scholl, Homer Chen Results of Core Experiment X5
1861Homer Chen, Hai Tao, Wei Wu Coding of Face Animation Parameters: Further Results
1862Thomas R. Gardos Results of Core Experiment T16
1863Yuval Fisher, Homer Chen Abstract Bit Stream Syntax of MITG
1864John Muller, Ghassan Farah, Mike Zeug Report on coding sequences with the Microsoft VM implmentation, version 5.01
1865John Muller, Ghassan Farah, Mike Zeug Report on results from core experiment T3
1866John MullerDescriptions of core experiments on coding efficiency in MPEG-4 video: prediction and sprite experiments
1867John MullerDescriptions of core experiments on coding efficiency in MPEG-4 video: texture coding and rate control experiments
1868Gary Sullivan, Tom Gardos, John Muller, Sen-ching Cheung MPEG-4 Video Compatibility with H.263/H.263+
1869Iraj Sodagar, Hung-Ju Lee, Ya-Qin Zhang Report of Core Experiment T1; Wavelet Coding of Intra Frames
1870Iraj Sodagar, Hung-Ju Lee, Ya-Qin Zhang Report of Core Experiment X5; Efficient Coding of Static SNHC Texture
1871Tihao Chiang, Hung-Ju Lee, Ya-Qin Zhang Report of Core Experiment Q2; Improved Rate Control
1872Tihao Chiang, Hung-Ju Lee, Ya-Qin Zhang Report of Core Experiment Q2; Comparison of VM and TMN6 Rate Control
1873Iraj Sodagar/DSRC, Ya-Qin Zhang, Raj Talluri, Ibrahim Sezan, Shawmin Lei, Weiping Li, Homer Chen, Ganesh Rajan, Masami Ogata Common Requirements For Still Image/SNHC Texture Coding in MPEG-4
1874John MullerPartial results for core experiment T14
1875Peter ListResults from Core Experiment T16
1876Sang Hoon Lee, Sungryul Cho and Seok-won Han Results of Core Experiment - S4e

Annex 4

Report Requirements Meeting, 17-21 February 1997, Sevilla

Source: Rob Koenen, chairman Requirements


The Requirements Group met during the entire week of the MPEG Sevilla meeting. The most important topics addressed were:

These issues will be discussed in detail below.

(NB: Although it was the intention to have a list of people that attended the meeting, the list that was drawn up is too incomplete to publish in this report.)

General Requirements

Version 2 of the General Requirements Document (N1595) was issued. Apart from some editorial issues, the most important changes were:

In a joint meeting with the Video group, a proposal for Still Picture Requirements was discussed (M1873). The conclusion was thaleonat the effort of joining requirements coming from the synthetic and natural world is a very useful one. More thought, however, needs to be put into understanding what is really necessary from an application point of view. The question will be addressed in the AHG on requirements.

Also in this joint meeting, the upper limit to the bitrates was discussed. The conclusion was that there is some indication that MPEG4 Video also performs well at higher bitrates. Furthermore, core experiments addressing interlaced video are being carried out. A definitive conclusion was left to the next (Bristol) meeting.

In a joint meeting with Audio, Audio requirements were discussed. Following proposal M1828, the definition of audio objects was adjusted. The suggestions made in the requirements AHG were approved. The need for adjusting Video synchronisation to Audio was discussed; it was agreed that information about this (a composition issue) should be in an informative part of the standard.

In a joint meeting with Systems, many issues were addressed. Here follows a brief overview:

The joint meeting with SNHC was dedicated to SNHC requirements. The need was expressed to get them in a much better shape, as there was too much overlap, resulting in contradicting requirements. Before the end of the MPEG meeting, revised input was received on SNHC Visual Requirements, and this input was included in he Requirements Document.



'How to deal with Profiles' was a much debated issue during the meeting, as was the reorganization of the Profiles document. A new Profiles document was issued (M1594). The conclusions on Profiles are the following:

The name 'Main' was discussed to some extent. The Requirements Group felt that it was the most appropriate name, as these profiles contain the most common tools and have no special features.

IPR Issues

Again experts from the content (management) world were present to work on the requirements for content-related IPR issues. In well-visited sessions of the Requirements Group, they discussed, together with the people previously present in MPEG, the requirements for Identification and Protection of Content. Dominique Yon chaired several break-out sessions with experts and interested people. The result of this work is laid down in document N1580: Requirements for identification and protection of content in MPEG-4. Also, a corresponding section was included in the MPEG-4 Requirements Document (N1595, Section 4.1.10). A proposal from GCL (M1730) with a potential solution was valuable guidance for the discussion.

A crucial element in the conclusions is that MPEG-4 should allow measures for protection to be implemented, by providing the appropriate hooks, but there is no requirement stating that MPEG-4 shall solve the issues itself.

Also at the Bristol meeting experts from the content (management) world will be present to further discuss the requirements, and it is expected that also experts will be present to start working on candidate solutions matching the requirements. Because of limited availability of these content experts, three days have been designated for content-related IPR discussions: Tuesday 8 through Thursday 10 April. Discussions will again partly take place in the requirements group, and partly in a break out group.

MPEG-4 Technical Overview

Using input from the different subgroups, a first draft was produced of the MPEG-4 Technical Overview Document (N1610). This is a much needed document, that shall inform MPEG members about the parts of the standard they are not directly involved in. In a later stage, it shall also be made available for public use, but a considerable amount of work is still needed before this is possible.

Already during the meeting many MPEG members took the trouble to give comments and suggestions for improvement - a sign that the document is indeed read (and necessary).

A special AHG was set up to further edit the document, with 'compulsory membership' of all the chairmen.


Another very fruitful session was held to discuss the future MPEG-7 standard. The major new conclusions, that were included in the 'MPEG-7 Context and Objectives (N1578):

An attempt was made to see how the work on MPEG-4 could be used to build on. The following potentially interesting starting points were mentioned:

It was also concluded that while the work on MPEG-4 may be useful, different optimizations are needed for MPEG-4 and MPEG-7. In MPEG-7 an efficient search is important, whereas in MPEG­4 it is the efficient transmission and storage that matters.

Some analogies with MPEG-4 exist:

MPEG members were invited to inform MPEG of work they have done that could be of interest to MPEG-7. This could be done in the form of a short contribution, describing the work, to one of the next two meetings.

The schedule for MPEG-7 was discussed, and kept the same: CfP in November 1998, IS in November 2000. Like in MPEG, the work will of course start much earlier, perhaps with preliminary versions of the CfP before November 1998

Public Documents and Ad Hocs

The following documents were released for public usage:

The Requirements Group spawned 4 ad hoc groups:

Annex R1

Agenda Requirements meeting

14.00 - 15.00lunch
15.00 - 16.00agenda

goals of meeting

report of chairmen's meeting

16.00 - 18.00 Profile development policy, profiles document
9.00 - 11.00Intellectual Property Rights Issues

1699 - AHG on content-related IPR - AHG Report

1730 - Itaru Kaneko/GCL - A pro/posal for content related IPR's

11.00 - 13.00Requirements coming from the broadcast application

1740 Sofie Olsson - MPEG-4 Broadcast Profile: Revision and Comments)

Mxxxx - David Meares - Comments on MPEG-4 Audio Requirements

13.00 - 14.00lunch
14.00 - 15.00M1639 - Jean-Claude Dufourd - Application Study: Interaction with a Multimedia Database
15.00 - 16.00Joint with Test and Audio
16.00 - 18.00Joint meeting with DMIF/Systems
10.30 - ??.??MPEG-7
??.?? - 13.00Profiles continued (if time after MPEG-7 allows)
13.00 -14.00lunch
14.00 - 14.30Joint with Audio (audio requirements)
14.30 - 15.00Joint with Systems (Systems Requirements)
15.00 - 15.30Joint with ISG (complexity in broadcast)
15.30 - 17.30Revision of Requirements Document (M1758);

profile discussion continued

17.30 - 18.00 Joint with Video
9.00 - 10.00Broadcast profile
10.00 - 11.00Joint with SHNC
11.00 - 13.00Merging of profiles
13.00 - 14.00lunch
14.00 - 15.00General Requirements document, including AHG report
15.00 - 16.30Profiles continued
16.30 - 18.00IPR Issues: report back from break-out group
8.00 - 9.00Joint meeting with Systems
9.00 - 10.00Revision of 'MPEG-7 Context and Objectives'
10.00 - 12.00approving documents / resolutions / AHGs
12.00 - 13.00final preparation of plenary
13.00 - 14.00lunch
14.00 - 21.30plenary meeting

Annex 5

Report of MPEG DSM-CC

Source: Vahe Balabanian, Chair, DSM-CC

1. Contributions
Doc #TitleContact Relevance
M1693Internet and User-Network signaling integration Guido FranceschiniDMIF
M1694DMIF directions for User-User signaling Guido FranceschiniDMIF
M1799Preliminary answers to the DSM-CC FAQ Vahe BalabanianDSM-CC
M1800Potential Joint MPEG-4 Systems and DMIF Activities Vahe BalabanianDMIF
M1802Liaison reply to DMIF from Q2 ATM & Other networks experts Group Vahe BalabanianDMIF
M1803Liaison reply from ITU-T SG16, Rapporteur Group for Q3/16 (T.130) Vahe BalabanianDMIF
M1804First cut at networking SRMs in DAVIC Vahe BalabanianDMIF
M1805Proposal to DAVIC for access to DAVIC Services from within Internet Vahe BalabanianDMIF
M1806Proposal to DAVIC for use of DAVIC Delivery to access Internet Multimedia Tools, Vahe BalabanianDMIF
M1807Proposal to establish a DMIF reflector Vahe BalabanianDMIF

Because of priorities and the shortage of time M1804 was not dealt with at this meeting.

2. Attendance

Attendance at this meeting was 16 a considerable increase over 5 in Maceio.
Bill WelshBTUK
Bernard PeigneCCETT France
Matthew GoldmanDiviCom USA
Alex LouiEastman Kodak USA
Chris HansenIntelUSA
Jurgen RufMedia Transfer Germany
Vahe BalabanianNortel Chairman
Masa KawashimaNTTUSA
Michael SzwarcNYNEX USA
Jeff NeedhamOracleUSA
Hoon-Jae KimSamsung Korea
Tim AddingtonScientific Atlanta Editor
Jim VanlooSun Micro USA
Yair Be'eryTAV/VDO net Israel
Mehrdad AhmediTeracom Sweden

The full attendance list is uploaded as file on the dsmcc

3. Delivery of the DSM-CC Tutorial

Presentation took 2:30 hours with a break of 20 mins.

The attendance was 36-40 and it remained constant throughout the tutorial. Two from ITU-T. Feedback it is mostly clear but it is still too abstract to some, need to bring in concrete examples.

14 evaluation sheets were received:

Coverage= just right 100%

Level of Interest= just right 100%

Level of Clarity= clear 91%, not clear 9% (one response in 14)

An e-mail sent out on the reflector provides more detail on the improvement of the tutorial.

4. Preparation of DSM-CC FAQ answers (version 1.0) - N1559

All the questions except 2 are answered. Will revise the FAQ based on the feedback received

Both the tutorial and FAQ will appear on the MPEG Web

5. Release of DSM-CC Conformance 13818-10 Preliminary Working Draft - N1560

Lots more work still needs to be done in order to identify additions to the conformance specification. Hope to receive strong input at the next meeting.

Many companies are conducting tests with others on an inter-company agreement basis.

We hope that the N1560 will give the base for them to open up their test with other companies.

6. Request for Participation in the DSM-CC Conformance Tests - N1562

and we are seeking to widen the membership by extending the invitation to Bellcore, Columbia University, Jupiter and NIST to feedback and participate.

It is requested that all organizations give serious consideration to this subject and report at the Bristol meeting.

7. Joint Meeting with Systems and Requirements

Improved the understanding between the groups. As a result a DSM-CC Sub Group has put a subset of DMIF (phase 1) on a fast track in 14496-6 to correspond to MPEG-4's schedule. The remaining part will move into 14496-7 as DMIF (phase 2).

8. Established a Joint Ad Hoc Group with Systems and Requirements


Walk through DMIF through a scenario provided by Requirements: MPEG-4 Broadcast scenario

Access MPEG-4 from/on Internet

Use of MPEG-4 resource descriptors

Binding of Flexmux to Transmux network resources

Use of download for MPEG-4

9. Priority of DMIF Functions in Support of MPEG-4 Systems - N1563

Identifies the priorities as "**" , phase 1, "*" phase 1 or phase 2, "unmarked" phase 2

This document will be kept up to date with ongoing interactions with Systems and other groups.

10. Policy Statement regarding DMIF VM - N1561

It defines DMIF VM and identifies the reasons and value drawn from the participation of organizations

It is requested that all organizations give serious consideration to this subject and report at the Bristol meeting.

11. DMIF Internet Integration - N1564 (DMIF Phase 1)

Identifies the scope of the work and outlines the work items and the inputs required.

This document will be kept up to date with the inputs received.

The delegate who will maintain this document is Jim VanLoo from Sun Microsystems, USA

12. MPEG-4 Resource Descriptors - N1565 (DMIF Phase 1)

Identifies the DMIF resource descriptors for MPEG-4.

This document will be kept up to date with the inputs received.

The delegate who will maintain this document is Tim Addington from Scientific Atlanta, USA

13. DMIF Binding between MPEG-4 Systems Network Resources - N1566 (DMIF Phase 1)

Defines how the network resources will be bound based on QoS parameters from MPEG-4 using DSM-CC definitions.

This document will be kept up to date with the inputs received.

The delegate who will maintain this document is Masa Kawashima from NTT, Japan

14. Use of DMIF Applications in Conjunction with MPEG-4 Systems - N1567 (DMIF Phase 1)

Defines how the applications such as download operate with MPEG-4 systems.

This document will be kept up to date with the inputs received.

The delegate who will maintain this document is Bernard Peigne from CCETT, France

15. DSM-CC Resolutions in Sevilla:

1- Accept N1559 as version 1.0 of answers to the DSM-CC FAQ

2- Accept N1560 as the Preliminary Working Draft of ISO/IEC 13818-10 DSM-CC Conformance

3- Resolved that DMIF will be divided into two phases, the first will be aligned with MPEG-4 time scale as part 14496-6 and the other will be defined as part 14496-7

4- Resolved that in Bristol, April/97, a project description will be provided for SC29 on DMIF phase 2 14496-7

5- Accept N1561 as the policy statement for DMIF VM

6- Accept N1562 as the request for participation in DSM-CC in DSM-CC conformance tests

7- Accept N1563 as a statement of the present priority of DMIF functions in support of MPEG-4 Systems

8- Accept N1564 as the scope, outline and some content of work on DMIF integration with Internet

9- Accept N1565 as the scope, outline and some content of work on MPEG-4 resource descriptors

10- Accept N1566 as the scope, outline and some content of work on DMIF binding between flex stream and the transmax network resources

11- Accept N1567 as the scope, outline and some content of work on DMIF applications in conjunction with MPEG-4 systems

Upcoming meetings:
Bristol UK, April 7-11/97
Stockholm SE, July 21-25/97
Fribourg CH, October 27-31/97

Annex 6

Report of MPEG Systems

1- Architecture

The architecture of the Systems is now stable. The main issue that was raised last meeting in Maceio was to see if the current specifications can allow for concurrency. This has been tested and documented since.

The concurrency feature has been recognised as highly desirable e.g. : an MPEG-4 player will certainly be more efficient if it can allow many decoders to run together on several processors. Up to the contributors, the architecture is acceptable, however, minor changes may be needed. More contributions on the topic that propose precise changes to the current spcifications.

2- Composition

The fact that an explicit scene decription was needed, as decided last meeting, was confirmed in Seville : an approach to reach this goal was defined and the first specification was produced. The starting point for scene description is the VRML 2.0 specifications.

A preliminary issue was to determine usefull sub-sets of the language and to identify missing functionnalities. Basically the following features will be added to VRML :

-1 Extension to 2D scene description,

-2 Extension of new nodes to deal with streamed media,

-3 Extension of new nodes for new MPEG-4 media type,

-4 Mechanisms for synchronisation of media types and

-5 Definition of an optimised binary representation.

Most of this points are now drafted in the systems verification model and working draft. Concerning point -5, no special feature is now defined, but the MPEG community will soon track the last redundant bits of the scene representation with its well-known efficiency in the domain. Feature like geometry compression have already identified by the SNHC sub-group and contributions have been presented. A name for this format : .BIFS has been proposed.

The next issue for the composition activity is to enter a phase of testing and validation of the specification. For the 2D case, it is forseen that the current APIs can be sufficient to implement 2D non-flexible composition in the MPEG-4 player. However, this has to be precise very soon by the verification model activity. Concerning the 3D case, the current APIs may be not sufficient. The tests will be therefore carried out using current extensible VRLM browsers. New APIs to build the MPEG-4 player may be defined if requested.

The Systems group strongly wishes to initialize a real collaboration (joint technical meeting, not bureaucracy ...) with the VRML community. In order to achieve this, a document was produce to evangelize the workplan above. This document should be broadly distributed, in particular to thye VRML experts.

3- Multiplex

The Systems experts have confirmed in the Seville meeting the need to separate the multiplex layer in two different layers : the FlexMux, that maps elementary streams (e.g : compressed data streams) of similar QoS in one single FlexMux stream, and the TransMux, that realises the interface between the FlexMux streams and the digital storage or compression medium. The specification of these two layers have been precised.

Links with the MPEG-4 DMIF sub-group are now taking shape. The relevant services provided by DSM-CC should be mapped to the specific needs of MPEG-4. The binding of Systems streams to network/storage media resources will be defined by DMIF. The binding of elementary streams to scene objects has been identified as a Systems activity. However, mechanisms similar to the ones defined by DSM-CC could be also usefull. Finaly, new services provided by DMIF experts needs further collaboration.

The mux activity is now entering a phase of validation. The current specifications allow for a software implementation that should be integrated as soon as possible in the MPEG-4 player.

The MPEG-4 multiplex has been designed to be the minimal layer that allow transport and storage of MPEG-4 content over various type of media. To achieve this, the MPEG-4 layer make some assumptions of the different properties of this underlying layers. In order to clarify these assumptions, the Systems sub-group is in the process of issuing a document that make these requirements explicit and that will be forwarded after the Bristol meeting to relevant bodies (ITU, ATM Forum, ...).

4- Syntactic Description

This activity was one of the first activity of the Systems group. The specification are stable and we have now entered a phase of evangelizing the language.

This is done in the following way :

Therefore, the main efforts on this activity mainly took place in other MPEG sub-group.

5- APIs

The APIs of MPEG-4 are one of the most promissing activity since they allow the standard to be extensible and flexible. However, the Systems group recognised early in the week that it could be very hard to fullfill all the objectives (mux, composition and APIs) for the November CD, and that we should therefore focuss our manpower in the more urgent areas of work : composition (non-flexible) and multiplex. The Systems group therefore decided to restrict the Nov. CD to the two previous areas.

However, some companies raised the point that the usability of MPEG-4 without APIs could be very limited, and that it was worth trying hard to get them in the Nov. CD. The Systems group accepted this very challenging task and the following was decided :

The motivated people of the verification model AHG will attempt to develop the APIs and a demonstrator based on them. If in july :

6- Future Work

The Systems activities will continue on the adequat reflectors until the Bristol meeting. Key issues are :


We are now entering a critical phase of our work where a little bit of method will be needed. Here follow some reminders that could help the process :

* editors of the document (working draft and verification model) propose and gather comments on the specification. These comments are then included in the document and proposed as inputs to the next meeting.

* if no agreement were reached on the reflectors, if some issues were treated incompletly, if new issues come up, individuals should make some writen contributions.

* in contributions, what part of Systems is adressed (architecture, multiplex, composition, ...), what part of the existing documents is adressed should be precised.

* if possible, propose a solution ;-) (it will save us quite a lot of time during the meetings)

List of Participants
Si Jun Huang Scientific Atlanta
Riccardo Pannochia Consorzio Pisa Ricerche
Yasufumi NakamuraFujitsu LTD
Shuichi WatanabeSharp Corporation
Stefano BattistaCSELT
Yoshinori MatsuiMatsushita EI
Fritz SeytterSiemens
Jean-Claude DufourdENST
Gunnar NitscheBosch
Lionel BouchardPhilips LEP
Junichi KimuraHitachi Ltd.
Ali TabatabaiTektronix Inc.
Hirokazu TanakaToshiba Corp.
Sofie OlssonTeracom
Rohit AgarwalIntel Corp.
Dominique CuretFrance Telecom
Bo BurmanEricsson Radio Systems
Chris HansenIntel Corp.
Giorgio ParladoriAlcatel
Julien SignèsFrance Telecom signè

Annex 7

Report of MPEG Video

Annex 8

Report of MPEG Audio

P. Schreiner, Chairman, Audio Subgroup
D. Meares, Secretary, Audio Subgroup

Opening of the meeting

The MPEG/Audio Subgroup meeting was held during the 38th meeting of WG11 in Seville, Spain on February 17 to 21, 1997. The list of participants is given in Annex A-I. The Chairman welcomed the delegates to the meeting and outlined the work for the five days. He explained the purpose and scheduling of the joint activities with other groups.

Approval of agenda

The agenda as presented in Annex A-II was discussed and approved.

Maceioó meeting report

The Audio Subgroup portion of the Maceio meeting report, November 1996, had been previously distributed and was approved.

Allocation of contributions

All contributions were listed (see Annex A-VI) and allocated to the agenda. All contributions directly related to the Subgroup were presented either in Audio plenary or in the task group discussions. Several relevant documents from Systems and Requirements were brought to the attention of the group.

Communications from the Chair

The Chairman summarised the detailed allocations and questions raised at the Chairman's meeting held on the evening before the main meeting started. The majority of these, by design, were already in the agenda.

The points raised by the Chairman related to

Joint meetings

Joint meetings were held with Test, Requirements, and SNHC as necessary to pursue the business of the week. The topics of the July 97 tests and verification testing were addressed with Test. The need for audio profile inputs to the requirements document and the technical overview document were discussed with Requirements. A discussion of the ad hoc activity in SNHC audio and demonstrations relating text to speech and structured audio were presented jointly for Audio and SNHC members. Further details of these activities appear later in this report.

Report of Ad-hoc group activities

All Ad-hoc Group reports (M1603, M1615, M1625, M1660, M1725, M1765, M1729) had been presented in the opening Plenary. Each was briefly presented again in the Audio Subgroup. Additional details were taken into account in the main part of the meeting.

Resolution of National Body comments

Comments from US NB on AAC DIS were addressed by the task group on AAC IS and will be conveyed through to the next meeting.

The US NB requested the rapid realignment of IS13818-4 & -5 to the changes to the standard given in 13818-3.2. Unfortunately the existing parts IS13818-4 & -5 have to be published as they are, but it was noted that we can ensure that the web site reflects this state of affairs. It was determined that no changes are required for IS13818-4 to align it with the revision. The foreword of IS13818-5 will have a statement indicating the minor technical change that is necessary for this part to agree with IS13818-3 (second edition). The output paper N1535 details the few technical changes that occurred in the revision to IS13818-3. A response to the US NB was prepared.

The US NB also requested early evaluation of the sampling rate scaleable profile of MPEG-2 AAC. The standard is about to go forwards without this profile ever having been tested. Sony has proposed to carry out such tests. A response was prepared accordingly.

Revision of schedules in line with new SC29 work plan rules

The new work plan rules were briefly reviewed, and it was concluded that there was no impact on the Subgroup's current schedules.

Temporary Task Group Formation

To accomplish the large number of tasks to be performed by the Audio Subgroup, task groups were formed as indicated in Annex A-V. The results of each of the task groups were presented to and discussed by the entire Audio Subgroup, including iterations as necessary. The conclusions of the task groups are presented elsewhere in this report and are included in the output documents.

MPEG-2 BC coding

IS 13818-3 (second edition) preparation

A task group was assigned to address the NB comments and prepare a DoC. Additionally, the earlier amendment, relating to copyright identifier, was taken into account. The DoC is given in document WG11/N1520 and the IS, including necessary changes, in document WG11/N1519. The Subgroup would like to express thanks to the contributors to the second edition and to Gerhard Stoll for his leadership of this thorough effort to improve the comprehensibility of the document and to address the few technical difficulties identified in the initial version of IS13818-3.

MPEG-2 audio quality update

Mr. v.d. Kerkhof reported that BC audio quality, relating to an improved algorithm, is planned to be evaluated by the ITU-R in about one year's time.

Bitstream tests

There were no further bitstream tests reported.

Technical Report and Conformance

Concern was expressed that IS13818 Parts 4 and 5 are now proceeding through publication when it is known that there is a feature in Part 3 which is in error and which has already been corrected in the Part 3 (second edition). Means by which corrections, errata etc. may be introduced prior to publication were explored. To deal with this matter formally, a contribution to a PDAM was prepared and approved. This is contained in document WG11/N1543. The introduction of this work will be delayed one meeting to align it with the other amendment to part 5 that is already in progress.


IS preparation

Ms. Bosi presented her ad-hoc report to the Subgroup and discussed with the Subgroup the changes that had been proposed within the ad-hoc group. The US National Body comments were also taken into account in the task group to ensure compatibility with the ad-hoc group's proposed changes. Contributions to the DoC which will be needed at the next meeting were prepared, see document WG11/N1518.

Mr. Meares raised the matter of matrix-downmix facilities for a stereo receiver of the multichannel bitstream, M1879. Two bits are to be reserved for the matrix-downmix options as specified in ITU-R Recommendation 775. This will need a National Body Comment in order to be formally adopted. This matter is covered in document WG11/N1521.

The issue of prediction complexity reduction was covered by the Mr. Quackenbush's ad-hoc group. Marked reductions in prediction algorithm implementation complexity are recorded for use of reduced word length variable representations in the algorithm and possibly a saving with a 24 bit fixed point implementation, but the issue of fixed versus floating point implementations is unresolved. The judgement of gain/loss is based on computational analysis rather than listening tests. Informal, but rigorous, subjective tests were included in the mandate of an ad-hoc group to report at the April 1997 meeting. Mr. Meares advised the Subgroup that the BBC would assist in the conduct of these tests.

The question of the name for the sampling rate scaleable profile was raised as the initials SRS are already registered by another organisation. SSR, scaleable sampling rate, was chosen. Additionally, it was decided that the SSR profile will be assessed in subjective tests to be run between now and the next meeting. Low anchors, as used in the previous AAC tests will be used again. The BBC or NHK will provide the appropriate files to Sony. Views were expressed over the reference codec in these tests, MPEG-2 BC or another AAC profile. Mr. v.d. Kerkhof objected strongly to the previous version of BC from the dTTb tests being used, but no newer version of BC has yet been assessed despite repeated requests. As Philips could not provide test stimuli for their new BC algorithm within the time scale of these tests, the AAC low complexity profile will be used. The details of these tests are given in document WG11/N1534.

Technical Report

Mr. Coleman presented the ad-hoc group report. The decoder from Mr. Quackenbush has been made available. Sony have also offered decoder software for the sampling rate scaleable profile. Additionally some encoder modules have been offered. The current status of the Technical Report preparation (WD) is given in document WG11/N1523


Copyright release has been received by Mr. Thom for many of the test items. Reference bitstreams are now required to judge conformance issues. The Working Draft of the Conformance document is given in document WG11/N1522

Subjective tests

AAC has not yet been tested in stereo or mono modes. Means of providing this information were discussed. It is reported by Mr. Brandenburg that the Canadian Research Corporation are planning to run stereo tests on AAC. We need to take note of these tests.

However, it was decided that MPEG Audio should be proactive and so a task group was set up to discuss the requirements further. A mono and stereo verification test plan was produced and is given in document WG11/N1538.

  1. Elementary Bitstream Type Identifier for AAC

Work was begun on a PDAM to IS13818-1 to have an elementary bitstream type identifier assigned in the systems layer for MPEG-2 AAC (13818-7 audio). The stream_id hex FA will be assigned to the AAC bitstreams. The necessary changes to two tables are presented in N1598.


VM activities

Mr. Brandenburg presented his ad-hoc group report on MPEG-4 audio core experiments. The matter of core experiments was addressed in a task group. This reviewed the tests to date, as well as the test methodology for the different functionalities, including the option to listen to some of the simulations. Additionally, the new core experiment proposals were reviewed. A test methodology for scaleable coding has been proposed, and after discussion and approval, this was merged with the previous test methodology document from Maceio. The final text is presented in document WG11/N1525.

Core experiments described in M1697 and M1755 were considered complete and proven. All of the other core experiments were thoroughly reviewed by the task group, the results from which are recorded in document WG11/N1524. Merging of new proven modules into the VM will continue to the next meeting. The current state of VM software for all the relevant modules is presented in document WG11/N1526.

A review was made of the availability of encoder software in source code. The majority of the CELP encoding modules are already available in source code, certainly sufficient for a simple embodiment of the MPEG-4 audio encoder. Many of the Parametric encoder and TF encoder modules are also in source code already. This code and much more is available on the MPEG-4 Audio ftp site described in document WG11/N1527.

MPEG-4 Audio Working Draft

Mr. Grill reported on his ad-hoc group's activities. Discussions with the Subgroup identified the additional features which had to be resolved during the meeting. Extensive work took place within a task group to pursue these points. By end of the meeting, all essential information from the core experiments that are approved were merged into the WD version 2.0, given in document WG11/N1536.

MPEG-4 Technical Report

The status of the Technical Report was presented by Mr. Coleman and work continued on it during the week. The Convenor additionally asked, in this context, for initial work to commence on an MPEG-4 Player. This was kindly picked up by Mr. Coleman who produced some initial proposals. The final status of the TR and that of the MPEG-4 Player is given in document WG11/N1542.


A joint meeting with Requirements Subgroup was held to discuss details relating to broadcast profile, real-time profile and synthetic audio. In particular the definition of audio objects needed to be refined to ensure that an audio object could be 1, 2 or up to channels, i.e. they are not restricted to mono sources, see documents M1828 and M1878. Resolutions to the various details were identified and are merged into document WG11/N1595. Audio Subgroup undertook to convert the remaining table at the back of the requirements document into a style of text common with the rest of the document.


Mr. Kaneko and colleagues from SNHC gave the Audio Subgroup an up-date on the progress of SNHC Audio and of his ad-hoc group since the Maceio meeting.

MPEG-4 TTS (Text to Speech) requirements (M1739) were presented to the Subgroup by members of SNHC. Problems of synthesis, lip-synch, pitch change, syntax, interface etc. are being addressed. A demonstration showing the current state of work with one of the proposals was presented. The question of how we proceed with the adoption of new features was raised. Some features such as the principles of TTS (functionality, bitstream syntax and semantics) have been adopted by SNHC. These now need to be incorporated into the VM. Core experiments are needed to assess specific embodiments.

The 'high level' programming language aspects of SNHC and system architecture were also discussed, M1788. 3-D spatialisation of the output audio, reverberation, post-mixing of natural audio and synthetic audio, synchronisation etc., are all important features that need to be controlled, as does the inclusion into the process of MPEG-4 Natural Audio bitstreams. The end product of this current approach is that Natural Audio and SNHC audio will be mixed and manipulated in the consumer unit according to the SNHC protocols.

In a joint meeting, SNHC members demonstrated various SNHC structured audio files, using both preloaded and downloaded sound tools in layered structures. This approach had been describe in a paper at the November 1996 meeting. Using this approach the file sizes are reduced significantly compared to an approach based on conventional encoding and lead to exciting new possibilities in the context of synthesised sounds. Source code VM modules will be available for this work at the next meeting.

Other developments

A document was prepared which identified the tools and functionalities expected to be included in MPEG-4 Audio. This helps to define the expectations between Subgroups and provides a useful contribution to MPEG publicity material. Much of what was drafted was included in the press statement.

Testing options for new proposals

July 1997 tests

The special tests required for new MPEG-4 Audio proposals were discussed in a joint meeting with Test and in a task group. Most of the proposals can be assessed by core experiments. Only 7 proposals were initially under discussion, of which 4 were immediately suitable to the core experiment approach. The advantage of handling by core experiments is that this procedure is much simpler. The proponents of these 'converted' proposals issued signed statements of the change of status.

If the core experiment approach is not possible for all proposals, then pre-screening (ref. document N1034) will have to take place before the July '97 tests. Formal cut off for this information and full technical details is formally 2 June. This poses severe problems in determining what goes forwards to July 97 tests.

The final situation was that only two proposals will be required to be evaluated in the July 97 tests assuming they pass the pre-screening tests. The pre-screening test requirements are that they should for preference be completed during April '97, such that the decisions needed for the July '97 tests can be made in time. The July '97 test requirements are given in document WG11/N1539. Additionally, it was resolved that those proponents pursuing the formal test approach should cover the entire cost of the formal test preparation and conduct. A resolution to this effect was approved in final Plenary.


The FAQs on MPEG Audio were reviewed and a revised full set were prepared as document WG11/N1537.

Preparation of a press statement

A contribution to the press statement was prepared and approved by the Subgroup.

Liaison matters

There were no liaison matters to be dealt with at this meeting.

Discussion of unallocated contributions

Mr. Meares reported the HoD's discussions relating to the Japanese NB paper on control of the MPEG file server, M1733.

The matter of IPR, M1699 and M1730, was dealt with in a general meeting on IPR matters for MPEG-4 .

Recommendations for final Plenary

A list of recommendations was prepared for approval at the final MPEG plenary meeting. It was recommended that document WG11/N1535 - "A guide to the second edition of ISO IEC 13818-3" and document WG11/N1537 - "MPEG Audio FAQs, version 2" be approved for public release.

Establishment of new Ad-hoc groups

The following ad-hoc groups were established:

Their mandates are given in the Error! Reference source not found..

Additionally, the following joint ad-hoc groups were established

Agenda for next meeting

The agenda for the MPEG Audio Subgroup meeting in April 1997 in Bristol, UK was approved (see Annex A-III).

Any other business

The Convenor requested that consideration should be given to what areas of work may be appropriate for MPEG at the end of 1998 and beyond. A small task group, led by Mr. Thom considered the matter and produced an ideas paper in document WG11/N1608.

Closing of the meeting

Mr. Schreiner thanked the participants for all their hard work in preparation for and during this meeting. He also thanked Mr. Meares for acting as Subgroup Secretary, and the hosts for the facilities provided. With that, he declared the Audio Subgroup meeting closed and wished members a safe return journey.

Annex A-I

38th MPEG/Audio Seville Meeting Participant List

(February 1997)

NameCountry Affiliatione-mail address
Akagiri, K.JSony
Bosi, M.USADolby Laboratories
Brandenburg, K.DEFhG - IIS
Cellario, L.ITCSELT
Coleman, M.USAFiveBats
Duenas, AUKDMV
Edler, B.DEUniversity Hannover
Feige, F.DEDeutsche Telekom AG
Ferreira, A.PTINESC
Grill, B.DEUniv. of Erlangen
Herre, J.DEFhG-IIS
Hotani, S.JNTTDoCoMo
Iwadare, M.JNEC
Iwakami, NJNTT
Kaneko, I.JASCII
Kerkhof, L. v. d.NLPhilips CE
Kim, S-W.KRSamsung
Knudsen, J. E.NTelenor
Koike, T.JSony
Lindqvist, M.SEricsson
Magill, J.UKLucent Technologies
Meares, D. J.UKBBC
Moriya, T.JNTT
Nishiguchi, M.JSony
Nomura, T.JNEC
Norimatsu, T.JMatsushita
Oikawa, Y.JSony
Ojala, P.FINNokia Research Centre
Oomen, W.NLPhilips
Parladori, G.IAlcatel Telecom
Purnhagen, H.DEUni Hannover
Quackenbush, S.USAAT&T
Scheirer, E.USAMIT
Schreiner, P. G.USAScientific Atlanta
Spille, J.DEThomson Multimedia
Tan, Ah-PengRSPanasonic Singapore Labs
Tanaka, N.JMatsushita
Taori, R.NLPhilips Research
Thom, D.USAMitsubishi
Väänänen, M.FIN Nokia Res.
Vercoe, B.USAMIT

Annex A-II

Agenda for the 38th MPEG/Audio Subgroup Meeting

in Sevilla, 17-21 February 1997

1 Opening of the meeting
2 Approval of agenda
3 Maceio meeting report
4 Allocation of contributions
5 Communications from the Chair
6 Joint meetings
7 Report of ad hoc group activities 1603, 1615, 1625, 1660, 1725, 1765, 1729,
8 Resolution of National Body Comments 1733, 1853, 1854, 1855
9 Timing of release of 13818-4 & -5
10 Revision of schedules in line with new SC29 work plan progress rules
11 Temporary task group formation
12 MPEG-2 BC
13 IS 13818-3.2 preparation 1610,
14 MPEG-2 audio quality update
15 Bitstream tests
16 Technical Report and Compliance 1854
17 MPEG-2 AAC 1603, 1615, 1625, 1853, 1855
18 IS preparation 1879
19 Technical report
20 Compliance
21 Assignment of bitstream type in MPEG-2 Systems layer
22 Other features: Identification of Language etc.
23 MPEG-4
24 VM activities 1606, 1643, 1677, 1697, 1728, 1753, 1755, 1760, 1815, 1816, 1830
25 Requirements 1757, 1758, 1759, 1828, 1878
26 WD
27 Systems 1619, 1741, 1752, 1778, 1790, 1838, 1840, 1846
28 SNHC 1616, 1617, 1649, 1788
29 Other developments 1634, 1636, 1661, 1695,
30 Complexity
31Testing options
July 1997 tests
Core experiments
1632, 1756, 1726,
32 MPEG 7
33 Participation in April seminar
34 MPEG Audio FAQ 1765,
35 Preparation of a press statement
36 Liaison matters
37 Discussion of unallocated Contributions 1699, 1730, 1733,
38 Recommendations for final plenary
39 Establishment of new Ad-hoc Groups
40 Agenda for next meeting
41 A.O.B.
42 Closing of the meeting

Annex A-III

Agenda for the 39th MPEG/Audio Subgroup Meeting
in Bristol, 7-11 April 1997

  1. Opening of the meeting
  2. Approval of agenda
  3. Seville meeting report
  4. Allocation of contributions
  5. Communications from the Chair
  6. Joint meetings
  7. Report of ad hoc group activities
  8. Resolution of National Body Comments
  9. Temporary task group formation
  10. MPEG-2 BC
  11. PDAM/1 IS 13818-5
  12. MPEG-2 audio quality update
  13. Bitstream tests
  14. MPEG-2 AAC
  15. IS preparation
  16. Technical report
  17. Conformance
  18. Verification test
  19. MPEG-4
  20. VM activities
  21. Requirements
  22. Profiles
  23. WD
  24. MSDL
  25. SNHC
  26. Other developments
  27. Complexity
  28. Testing options
  29. July 1997 tests
  30. Pre-screening results
  31. Core experiments
  32. MPEG Audio FAQ
  33. Preparation of a press statement
  34. Liaison matters
  35. Discussion of unallocated Contributions
  36. Recommendations for final plenary
  37. Establishment of new Ad-hoc Groups
  38. Agenda for next meeting
  39. A.O.B.
  40. Closing of the meeting

Annex A-V

Audio Task Groups

MPEG Audio FAQ - Kerkhof

13818-3.2 DIS to IS conversion and DoC - Kerkhof



AAC IS preparation, Conformance and Technical Report - Bosi
ColemanLueckHerre IwadareAkagiri

MPEG-4 WD editing, Functionalities and Tools - Grill
SpilleFerreiraIwakami DuenasNishigushi
KimHotaniIijima TanFeige
TanakaTaoriCellario RaultOomen
KnudsenLindquistKoike EdlerAkagiri

MPEG activities beyond 1998 - Thom

MPEG-4 Core Experiments, A/V test material - Brandenburg
FerreiraMoriyaIwakami NishiguchiKim
HotaniIijimaTan OomenTanaka
CellarioHerreLindquist KnudsenThom
KoikeEdlerFeige TaoriAkagiri

AAC stereo/mono verification tests - Brandenburg

Preparations for the Audio July 97 tests - Feige


Edler Taori

Annex A-VI

Input/Output Documentation

Contributed documents

The following documents were contributed to the Audio Subgroup and were considered during this meeting:
NumberGroup SectionSource Title
M1603MPEG-2Audio Marina BosiReport of the Ad Hoc Group on MPEG-2 AAC DIS 13818-7 Progression
M1606MPEG-4Audio Pasi OjalaResults of the core experiment of NEC compression algorithm proposal for MPEG-4/Audio
M1610MPEG-2Audio SC 29 SecretariatSummary of Voting on ISO/IEC DIS 13818-3 (Second Edition)
M1615MPEG-2Audio David Thom,
Mike Coleman
Ad Hoc Group Report on MPEG-2 AAC Conformance Testing & Technical Report
M1625MPEG-2Audio S. R. QuackenbushReport of the Ad-hoc Group on Reducing AAC Prediction Complexity
M1643MPEG-4Audio Y.B. Thoman Kim, S.H. ParkA Proposal for the MPEG-4 Audio T/F based scaleable codec test
M1660MPEG-4Audio Bernhard Grill,
Heiko Purnhagen
Report of the AhG on MPEG-4 Audio Working Draft Editing and VM Software Implementation
M1677MPEG-4Audio Naoya Tanaka,
Koji Yoshida
Results of the Core Experiment m1509
M1697MPEG-4Audio Masayuki Nishiguchi, Kazuyuki Iijima,
Jun Matsumoto
Bit rate controllable/scaleable tools for the parametric speech coder core of the VM
M1725MPEG-4Audio Karlheinz BrandenburgReport of the Ad-hoc Group on core experiments for MPEG-4 audio
M1728MPEG-4Audio Jun Matsumoto, Masayuki Nishiguchi, Kazuyuki Iijima Core experiment results in Sony for NEC proposal
M1753MPEG-4Audio Kenzo Akagiri, Takashi KoikeA description of the MPEG-2 AAC compliant pre-/post- processing tools for the MPEG-4 VM framework
M1755MPEG-4Audio Rakesh TaoriProposal for new features in the MPEG4 wideband coder: bit-rate scalability and complexity scalability
M1760MPEG-4Audio Akio Jin,
Naoki Iwakami, Takehiro Moriya, Satoshi Miki, Kazunaga Ikeda
Proposal of bit-rate scalability tools for MPEG-4/audio VM
M1765AllAudio Gerhard StollReport of Ad-Hoc Group on Audio FAQ for the MPEG Home Page
M1815MPEG-4Audio Bernhard Grill,
Bodo Teichmann
Additional Block Sizes for AAC-based Coding in MPEG-4
M1816MPEG-4Audio Bernhard GrillImplementation of a Scaleable Audio Coder based on AAC and a Core Coder
M1830MPEG-4Audio Lin Yin,
Mauri Vaananen
Proposal for a Core Experiment on Backward Prediction in MPEG-4 Audio
M1853HoDAudio USNBUSNB Comments on DIS 13818-7
M1855HoDAudio USNBUSNB Contribution re MPEG-2 AAC SRS profile
M1878MPEG-4Audio D. MearesComments on MPEG-4 Audio Requirements
M1879MPEG-2Audio D. MearesComments on the AAC DIS to IS progression ballot

The following documents were also noted as relevant to the business of Audio Subgroup and discussed accordingly
NumberGroup SectionSource Title
M1608MPEG-2Systems SC 29 SecretariatSummary of Voting on ISO/IEC 13818-1/PDAM 4
M1609MPEG-2Systems SC 29 SecretariatSummary of Voting on ISO/IEC 13818-1/DCOR 1
M1616MPEG-4SNHC Caspar Horne, Ganesh RajanReport of the AdHoc Group on SNHC VM editing
M1617MPEG-4SNHC Caspar Horne, Ganesh RajanSNHC Verification Model 2.1
M1619MPEG-4Systems A. Eleftheriadis,
H. Kalva
A Proposed Architecture for an Object-Based Audio-Visual Bitstream and Terminal
M1632MPEG-4Test V. BaronciniProposal to host the July '97 MPEG-4 Tests
M1634MPEG-4Implementation Marco MattavelliReport of the ad-hoc group on Computational Graceful Degradation
M1636MPEG-4Implementation Marco Mattavelli, Sylvain BrunettonDraft proposal for the specification of the ISG Computational Graceful Degradation core experiment
M1649MPEG-4SNHC Bill Welsh, Andy Breen, Steve MinnisDescription of Synthetic Talking Head Demonstrator for SNHC
M1695MPEG-4Systems G.FranceschiniElementary Stream to Adaptation Layer interface and Adaptation Layer Headers
M1699MPEG-4Requirements Dominique Yon, CISACReport of ad hoc group on content-related IPR
M1726MPEG-4Test Laura ContinReport of the ad hoc group on MPEG-4 July 97 audio, video and audio-visual Tests.
M1729MPEG-4SNHC Itaru Kaneko,
Shinya Nakajima
Chairman's report of Ad-hoc group on the SNHC Audio
M1730MPEG-4General Itaru KanekoA proposal for content related IPR's
M1733HoDGeneral The National Body of JapanRequests on MPEG document handling
M1740MPEG-4Requirements Sofie OlssonMPEG-4 Broadcast Profile : Revision and Comments
M1741MPEG-4Systems Julien Signes, Dominique Curet, Bernard Peigne An example of audio and 2D/3D content manipulation application: Multiplex, 3D composition and distant interactivity issues
M1752MPEG-4Systems Carsten HERPELReport of adhoc group on Systems multiplex specification and signaling
M1756MPEG-4General Rob Koenen,
Laura Contin
Registrations for July 97 Tests
M1757MPEG-4Requirements Rob KoenenReport of ad hoc group on MPEG-4 Requirements
M1758MPEG-4Requirements Rob KoenenDraft Revision of MPEG-4 Requirements Document
M1759MPEG-4Requirements Rob KoenenDraft Revision of MPEG-4 Profiles Document
M1778MPEG-4Systems Systems VM AHGSystems VM 2.0 (Editor's revision)
M1780MPEG-2Systems Paul HaskellMPEG-2 Systems Test Bitstream Status
M1788MPEG-4SNHC Eric D. Scheirer, Barry VercoeSynthetic Audio Specification Language Proposal for SNHC Audio
M1790MPEG-4Systems Christian Koechling, Gunnar NitscheMSDL VM and Audio
M1828MPEG-4Requirements Jens Spille,
Carsten Herpel
Definition of Audio Objects and proposed changes to the MPEG-4 Requirements
M1833MPEG-4Video Marta KarczewiczReport on Core Experiment P9
M1840MPEG-4Systems Olivier AvaroReport of the AdHoc Group on Systems Working Draft Editing
M1846MPEG-4Systems Carsten HerpelFraming ('Synchronization') Mechanisms in MPEG-4
M1854HoDGeneral USNBUSNB Contribution re 13818-4 and 13818-5

Output Documents

The following output documents were produced by the Audio Subgroup. Those shown in Italics were approved for public release.
NumberAuthor Title
N1518Bosi, M.Information on MPEG-2 DIS 13818-7 (MPEG-2 Advanced Audio Coding, AAC)
N1519Kerkhof, L. vdFinal text of ISO/IEC 13818-3 (second edition)
N1520Kerkhof, L. vdDoC to ISO/IEC 13818-3 (second edition)
N1521Meares, D. and Quackenbush, S. Provision of Matrix-downmix stereophony in MPEG-2 AAC
N1522Coleman, M.Working Draft for AAC Conformance
N1523Thom, D.Working Draft and activity plan for MPEG-2 AAC Technical Report
N1524Thom, D.Overview of MPEG-4 audio core experiment
N1525Thom, D., Brandenburg, K. MPEG-4 audio core experiment methodology
N1526Grille, B.Update to the software decoder VM for MPEG-4 Audio
N1527Grille, B.MPEG-4 Audio ftp site information
N1534Akagiri, K.Listening test working plan of MPEG-2 AAC SSR profile multichannel audio
N1535Kerkhof, L.vdA guide to the second edition of ISO IEC 13818-3
N1536Grille, B.MPEG-4 Audio Working Draft version 2.0
N1537Kerkhof, L.vdMPEG Audio FAQs version 2
N1538Brandenburg, K. MPEG-2 AAC mono and stereo verification test plans
N1539Feige, F.MPEG-4 audio test plans for the July 97 tests
N1542Coleman, M.`Status Report for MPEG-4 Technical Report (MPEG-4 Player)
N1543Kerkhof, L.vdContribution to working draft on the Technical Report 13818-5 relating to the second edition of 13818-3.
N1608Thom, D.Concepts for Audio in 1998 and beyond

SNHC Meeting Report

38th WG11 Meeting, Sevilla, Spain

February 17-21, 1997

Pete Doenges

March 18, 1997

1. Introduction

The Sevilla WG11/MPEG meeting was productive for MPEG-4 and SNHC. A better understanding of SNHC requirements and functionality was apparent among contributors. There was more interaction of people across groups with an improving grasp of the system-level challenges and solutions associated with absorbing SNHC into Systems, Visual and Audio. However, people have been working very intensively in their own areas in many cases. The integration of SNHC into the MPEG-4 parts requires that SNHC contributors shift their focus toward exploiting the resources available in Systems and other Groups where appropriate to represent SNHC objects and functionality. Contributors more knowledgeable in Systems, Video, and Audio can assist in this transformation of SNHC. A coherent mapping of SNHC into MPEG-4 parts will otherwise be compromised. SNHC made concrete progress, with hard work by numerous contributors, through the CFP (for SNHC Audio) and Core Experiment processes, and through useful joint meetings. This report provides an overview of progress and achievements during the meeting.

2. Summary

Many additions and refinements were made to the SNHC VM 3.0 for Face Body Animation, Media Integration of Text and Graphics, Texture Coding, Static and Dynamic Mesh Coding with Texture Mapping, Text to Speech, and the new Structured Audio. The texture coding effort formed in Maceió made tremendous progress in a start-up mode, leveraging prior work from the Video group while bringing important new work to SNHC, with numerous Core Experiment results, and the first texture and mesh coding contributions to the SNHC VM. A Structured Audio technology demonstration was given for Audio and SNHC listeners. An overview was also presented for Systems composition and Audio groups on the SA architecture, Orchestration Language, API, Score File, host control, and the implications for constructing audio synthesizers, mixers, spatial auralization, MIDI integration and extension, and user control with interaction. Demonstrations of Core Experiments and related functionalities were given by FBA, MITG, SA, TTS, and texture/mesh-based coding (now SNHC Generic Object Coding). Joint meetings with Systems, Requirements, Video, and Audio supported the process of unifying architecture, distilling tools, and helping SNHC transition into the MPEG-4 VM/WD.

Significant results of the Sevilla meeting were:

Challenges for SNHC are numerous as we approach the Bristol meeting in only six weeks (five to submissions). New Core Experiments must be evaluated and the results distilled to determine what final set of tools is retained in the SNHC VM/PWD as it is transitioned into the main MPEG-4 parts. A clear, shared understanding must be reached on how to exploit Systems/MSDL capabilities fully to reach an integrated and coherent syntax and semantics for SNHC that preserves needed performance in decoding of and terminal operations on streaming and downloaded A/V objects. Significant progress is needed to carefully organize and distill profiles for SNHC, building on existing broadcast and interactive profiles of MPEG-4 wherever possible, and transitioning summaries of viable SNHC applications to the Profiles Appendix. Further improvements in Requirements must be made by SNHC. SNHC VM software components, including source, will be assembled at an SNHC VM software site. A mapping of these software elements into the MPEG-4 VM parts must be made, and then the SNHC VM software reconfigured under MSDL for the July meeting.

3. Technical Management Imperative

SNHC started late in the MPEG-4 process. A compressed work plan (in relation to the overall MPEG-4 work plan) was released last March with the original CFP, and then augmented with the SNHC Audio CFP last September, to produce the initial release of the SNHC WD by the end of the upcoming Bristol meeting. SNHC contributions have been swelling in the last two meetings, and work has reached a fever pitch. This crucial work provides a baseline for useful SNHC functionality that must now be refined and conformed into the Systems architecture, VM, WD, Requirements, and Profiles of MPEG-4 as a whole. The time available to admit new technology into SNHC, consistent with the initial MPEG-4 CD release in late-October of 1997, is rapidly closing. SNHC must spend the next two meetings "finishing what we started" while migrating properly into the MSDL framework, including VM software and documentation. Mike Colman needs help in developing the MPEG-4 Player targeted for July 1997. SNHC can be further refined before IS in November 1998, but SNHC must now focus on a stable foundation whose parts work together as specified without a glitch. Based on a realistic work plan, SNHC must "cap" or "draw a line" around essential functional scope in Bristol.

4. SNHC & Related Contributions

Sevilla contributions were presented or discussed from Monday through Thursday in SNHC task group meetings and in joint meetings with other groups, based on an allocation of contributions outlined below:

M1616Horne, RajanReport of the Ad Hoc Group on SNHC VM Editing
M1617Horne, RajanSNHC Verification Model 2.1
M1818NagarajanProposal for Core Experiments in SNHC

4.2 SNHC Face/Body Animation
M1597Mulroy, Welsh, Breen Face Animation and TTS Files for SNHC Demonstration
M1618Wu, Chen, TaoRockwell C Implementation of SNHC FAP Coding
M1649Welsh, Breen, Minnis Description of Synthetic Talking Head Demonstration for SNHC
M1808Petajan, CapinAd Hoc Group Report on Face and Body Animation
M1835Thalmann, Thalmann, Pandzic, Capin, Escher, Boulic, Miccio, Molet MIRALab/LIG Contribution to SNHC in the Project VIDAS
M1839Capin, Boulic, Esmerado, Molet Proposal for Update of Body Animation Section in the SNHC-VM
M1861Chen, Tao, WuCoding of Face Animation Parameters: Further Results

4.3 SNHC Media Integration of Text & Graphics
M1623NagarajanReport of Ad Hoc Group on Media Integration of Text & Graphics
M1654Auore Experiments Report on SNHC-MITG
M1794RajanResults of Core Experiments on Media Integration of Text & Graphics
M1825KattoResults of Core Experiment M1
M1863Fisher, ChenAbstract Bit Stream Syntax for MITG

4.4 SNHC Audio (TTS, Structured Synthetic Audio)
M1729Kaneko, Nakajima Chairmanís Report of Ad Hoc Group on SNHC Audio
M1739Hahn, Lee, Yang Revision of MPEG-4 TTS Interface
M1788Scheirer, Vercoe Synthetic Audio Specification Language Proposal for SNHC Audio
M1848Ostermann, Beutnagel Comments on the Current Definition of TTS IN SNHC

4.5 SNHC Texture Coding
M1629Tekalp, Beek, Toklu Description of CE X4 (2D Mesh-Based Texture Mapping for Synthetic and Natural Content Manipulation)
M1630Tekalp, BeekResults of CE X4 (2D Mesh-Based Texture Mapping for Synthetic and Natural Content Manipulation)
M1667Aono, Saiga, Katata Spatial Scalability with Fine Granularity (in Core Experiment X5)
M1690DeKnuydt, Desmet, Gool, Eycken Coding of Dynamic Texture: Problem and Example
M1706Torres, Gimeno, Garcia Results of Core Experiment X1
M1708Messing, SezanResult of X4 for Augmented Reality
M1722EbrahimiReport of the Ad Hoc Group on SNHC Texture Coding
M1723EbrahimiDescription of Core Experiments in SNHC Texture Coding
M1737Jordan, Horbelt, Ebrahimi Results of Core Experiment X3 for Texture Coding
M1775Schmid, Menagaz Results of Core Experiment X1 for Efficient Coding of Homogeneous Textures
M1782Li, Forchheimer, Ahlberg Efficient Coding of Facial Texture Results of Core Experiment X2
M1792Rajan, Banerjee Results of CE X4 (2D Mesh-Based Texture Mapping for Synthetic and Natural Content Manipulation)
M1793RajanResults of CE X3 (View-Dependent Texture Coding for Transmission of Virtual Environments)
M1798Liang, Moccagatta Results Report on Core Experiment X5: Efficient Coding of Static SNHC Texture to be Mapped in 2D/3D Surface Models
M1858Wu, ChenResults of X3: View-Dependent Texture Coding
M1860Scholl, ChenResults of Core Experiment X5
M1870Sodagar, Lee, Zhang Report of Core Experiment X5: Efficient Coding of Static SNHC Texture

4.6 SNHC Geometry/Mesh Coding
M1744Choi, Lee, AhnGeometry Compression Using Predictive Vector Quantization
M1784BossenEfficient Representation of Triangulations
M1826Nagel, Falkehagen, Graffunder New 3D Graphics Interface Using General Scene Initialization & Update Parameters

4.7 Systems re: SNHC
M1620BarronVRML Node Inclusion in MSDL Model
M1741Signes, Curet, Peigne An Example of Audio and 2D/3D Content Manipulation Application: Multiplex, 3D Composition, and Distance Interactivity Issues
M1771Irwin, SchaferMPEG-4 Terminal Architecture, Introduction
M1774Bouville, Signes 2D/3D Scene Modeling and Presentation
M1776Cooke, Bouville Proposal for 2D & 3D Composition
M1777Cooke, WardProposal for 3D Scene Description and Composition
M1813JavasoftJava Media Framework

4.8 Requirements re: SNHC
M1757KoenenReport of Ad Hoc Group on MPEG-4 Requirements
M1758KoenenDraft Revision of MPEG-4 Requirements Document
M1759KoenenDraft Revision of MPEG-4 Profiles Document
M1828Spille, HerpelDefinition of Audio Objects and Proposed Changes to the MPEG-4 Requirements
M1873Sodagar, Zhang, Talluri, Sezan, Lei, Li, Chen, Rajan, Ogata, Bower Common Requirements for Still Image/SNHC Texture Coding in MPEG-4

5. SNHC Work Groups

The SNHC work plan for Sevilla emphasized separate meetings of SNHC Work Groups to make progress on the VM (including a large number of Core Experiment results from the first round initiated in Maceió). Related joint meetings with Audio, Video, Systems, and Requirements were designed to address requirements, architecture, meeting contributions of shared interest, object representation (streaming & downloaded), and audio/visual composition.

5.1 FBA

The FBA work group met extensively. Body Animation was more widely represented by contributors compared with Maceió, and was better integrated with Facial Animation in Sevilla. Face and Body sections were merged together, and the Body section was made more complete. BAP compression was added to the FBA-related syntax of the SNHC VM. Thoughtful contributor input led to definition of a set of separate FDP feature points in relation to the 3D facial mesh. This provides a level of abstraction in facial animation control, and a framework for future improvements to interface with avatars and more general characters. CE demonstrations of talking heads and expressive avatars were very encouraging. A new Core Experiment was proposed to evaluate lossless and lossy compression of FAPs/BAPs at predefined bit rates with selective quantization and adaptive arithmetic coding. FBA offers very specific challenges to SNHC profiling, multiple stream multiplexing/demultiplexing, and efficient object representation under MSDL to achieve real-time mesh manipulation and A/V object rendering for speech intelligibility.

5.2 MITG

MITG made significant progress with an initial round of impressive CE results to verify basic functionality for static and animated text, combining overlays with backgrounds, coding initial compound scenes, arranging scene layouts into useful temporal and spatial scene compositions, and related functionalities. These drove modifications to the MITG section of SNHC VM 3.0. MITG architecture and object definitions were considered to be fairly stable, while inter-object synchronization and conditional behavior deserve more attention. New Core Experiments were proposed to evaluate and refine functionalities currently provided in SNHC VM 3.0 as well as new functionalities. The first CE was for the media integration of static and animated text and its manipulation. This CE explores: a) static text overlay on images/video with variations in attributes (e.g. language, font, transparency, orientation/placement), b) streaming text overlays with different text attribution, scrolling modes, and multi-layer transparency, and c) animated text including looping text with selective updates, text behavior triggered by object picking, and linkage of text behavior to conditions met by its own or some other object.

A second CE targets media integration of images and graphics including a) variation of placement/transparency attributes of coded still image overlays, b) transformation processing on decoded images before compositing with the A/V backdrop including user or content-provider control of dynamic transformations/warping, c) video overlays on video backdrops, and d) interaction of MITG objects where image object behaviors are conditioned by scene events between objects. 2D graphics overlays will also be discussed for future core experiments. The third CE addresses complex object representations such as grouping of static and animated overlay objects, and correct behavior of animated compound objects consisting of dissimilar subordinate objects. MITG functionality, performance, fidelity, extensibility, portability, modularity, timing, provider/user control, and resource demands will be evaluated. CE partners should begin profiling for implementation complexity of decoders, data caching, and terminal-resident libraries.

The MITG group is turning attention to efficiency, composition, and object representation associated with migrating MITG into MSDL's framework of 2D/3D objects and spatial-temporal scene composition. The Bristol meeting should be used to confirm how MITG fits within or augments current MSDL A/V object class hierarchy, method interfaces, scene composition semantics, timing and synchronization, multiplexing/demultiplexing, and functional/object libraries.

5.3 TTS

Detailed improvements in syntax were developed. PC-based TTS with selectable application of prosodics, gender, etc. was demonstrated in a story teller example, with play control over pitch/rate, and with pre-defined facial animation frames matching visemes with phonemes as the speaker tells the story. Further personalization of the speech with a training set, and linkage of TTS to Facial Animation, are also possible in specific configurations. Applications of this TTS work include story teller on demand, Facial Animation driven by TTS, and dubbing of Motion Pictures with TTS. A new Core Experiment was proposed to verify the usefulness of the TTS interface now specified in SNHC VM 3.0 for the possible improvement of the present TTS interface. The new CE will verify scalability and synchronization of speech synthesis with FA/MP including the use of original or more tailored prosody, trick mode functions, generation of phoneme/viseme symbols needed by FA, synchronization of visual lip-shape changes with speech synthesis, and user control of speech types and qualities. With the Structured Audio initiative, there is also a possibility of using the SA audio compositing functionality to blend speech, signal-type audio coding, and synthetic sound/music coding,

5.4 SA

After the indirect review of proposals from Yamaha and MIT in Maceió and listening to audio samples, Sevilla provided a water shed of useful input on Structured Audio. Key contributions were made in Sevilla as technology demonstrations (no previously defined Core Experiments), presentations, and discussions about supporting simpler MIDI stream applications as well as more complex SA downloads and streams embodied in the SA Orchestration Language, API, Score File, and host control architecture to enable construction of audio synthesizers, mixers, spatial auralization, MIDI extensions, and the potential for user control with interaction. Hard work by very conscientious contributors resulted in a substantial addition to the SNHC VM 3.0, while more work is needed to harmonize the SA architecture fully with Systems/MSDL. Two AHGs were formed, one under N1532 to refine methodology for the conduct of Audio core experiments including the integration and verification of SNHC Audio tools within the Audio VM and WD. A second AHG, formed under N1603 for a number of purposes, will look at the impact of SA with respect to MSDL architecture, and the further related development of the scene description model and the composition of A/V objects as related to the aural environments supported by the SA component of the SNHC VM. This effort is expected to assist the cut-over to MSDL-based SA as the SNHC VM/PWD is mapped into the MPEG-4 parts.

SA Core Experiments were successfully proposed. Three CEs include: evaluation of a channel-oriented audio decoder using SA, achieving desired maximal functionality in a Java-less terminal (the baseline MPEG-4 capability), and exploring simpler lower-cost scenarios of synthetic audio coding for new SA Profiles. The first CE will investigate how to implement usefully a natural audio decoder, based upon encapsulating natural audio signal coding while providing the benefit of concurrent access to efficient synthetic sound coding and audio compositing. This integration could result in a reduction in overall standardization complexity facilitated by SA. The second CE aims to limit the initial SA specification and implementation by deleting reliance upon Java (for presentation and scheduling of events) upon which the original specification critically depended, and to understand functional compromises necessitated by Java-less implementations of an MPEG-4 SA-equipped terminal. The third CE will evaluate the viability of adopting a subset of the SA tool suite, while preserving enough essential functionality to support lower-power, lower-complexity terminals. This can be compared with the application scenarios in the current SNHC VM. In Bristol, these Core Experiment results can hopefully help make scaling and architectural decisions about integrating SA with MSDL.

5.5 Generic Object Coding

In Work Group meetings and selected joint meetings, contributions were extensively presented or at least summarized on coding of homogeneous texture, facial texture, several forms of view-dependent texture, animated 2D meshes, 3D mesh compression, ZTE wavelets, and methods to leverage the existing Video VM I-frame compression algorithm. Many CE videos were observed. Supporting descriptions were given about applicable comparisons of visual quality achieved, artifacts under motion, PSNR/resolution/bitrate scalability, terminal resources needed in different cases, encoder/decoder complexity and asymmetry, compatibility of texture types with progressive vs. one-shot downloading, and repeatability between different core experimenters.

Contributors realized quickly that current Core Experiments were aligned too strongly with a tool or technique, and not so clearly with a desired functionality or requirement. This led to a redesign of CEs for Bristol along functional lines. Baseline aspects of homogeneous texture, wavelet coding, view-dependent texture, 2D/3D static meshes, and dynamic 2D Delaunay meshes were adopted into the SNHC VM 3.0. This was done with the understanding that competing techniques must be re-examined in CEs for Bristol along functional lines. The goal then will be to consolidate tools under the simplest robust framework that meets requirements and unifies the coding of specific types of still image texture coding embedded in Video and SNHC tools for streaming, progressive, and downloaded image types.

CE definitions for Bristol are extensive. Basically, the CEs are broken down into four categories: a) coding efficiency for high-quality texture, b) texture coding for quality, spatial and view-dependent scalability, c) coding of 3D meshes, and d) coding of animated 2D meshes. The sizes of the original images considered include up to 2048x2048 for generic texture images, and up to 512x512 for homogeneous natural texture, with widely varied target compression ratios depending on the type of scalability. Specific techniques such as Stochastic VQ, eigentextures, DCT, wavelets, motion estimation, quantization and coding methods, or even layered combinations of underlying techniques are then free to address functional objectives for which they are best suited.

5.6 2D/3D Scene Composition (Systems)

Systems has been working this topic long and hard for over a year with cooperating input from SNHC contributors and others who concentrate back and forth as the Systems architecture has settled. Samples of MSDL functionality have been implemented to demonstrate the adequacy and integrity of the underlying class hierarchy, spatial and temporal composition system, multiplexing/demultiplexing, synchronization, and ability to link decoder and rendering facilities together to animate real-time mixed media considered by MPEG-4. This overall design task is complex, and relies on leveraging establish 2D/3D modeling technology and standards where possible. A balance should be struck between coding efficiency, speed of manipulation and rendering of decoded A/V objects, content scalability, and the levels of abstraction supported by MSDL. SNHC functionality needs to achieve high efficiency in coding and manipulating specific A/V object types. Several valuable contributions were presented in Systems with SNHC attendance on the comparative strengths of VRML 2.0 and MPEG-4, and on MPEG-4 strategies for extending (a subset of) VRML.

Much discussion resulted about how this approach can meet SNHC and other MPEG-4 requirements, in what cases specialized node extensions to VRML 2.0 may be needed, and how MPEG-4 can adapt VRML concepts to create a binary format for scenes with the ability to mix downloaded and streaming objects at real-time speeds under the Multiplex conventions provided in MSDL. A decision was made to proceed with this hybrid approach to scene composition that taps VRML technology (and likely elements of VRML content). MPEG-4 will implement the necessary tools in steps that place top priority on MPEG-4 specific nodes included in a group of essential nodes otherwise adapted from VRML as necessary to achieve predictable composition and rendering complexity. The approach taken still allows MPEG-4 to optimize the representation of specific nodes (e.g. 2D animated mesh "attached" to video, geometry compression, streaming geometry and other streaming A/V objects, etc.) native to MPEG-4. Software (MSDL, decoder) portability across different terminal types is still presumably achieved by compiling profile-specific tool kits and A/V object libraries as needed into specific terminal implementations.

An important document on composition methodology was generated by the Systems group (N1544), and conveyed to (a busy!) VRML Consortium meeting just after Sevilla. Relevant consortium Working Groups recognize the difficulty and specialization associated with achieving MPEG-4 real-time mixed-media composition, and appear willing to share expertise with MPEG's evolving needs, including formation of a Streaming Data WG.

6. SNHC & Joint Meetings

The following meetings were held and presentations made on SNHC-related topics:
SNHC PlenaryNew attendees, agenda, contribution allocations, schedule, deliverables
SNHC Work GroupsBreak-outs for SNHC VM/PWD work
As available
Systems, SNHCContributions on 2D/3D composition, animation, object representation
Systems, SNHC2D/3D scene composition for SNHC based upon extensions to VRML
Systems, SNHCMigrating SNHC objects into MSDL architecture (FBA, MITG, etc.)
Systems, SNHC2D mesh, 3D geometry compression
Systems, SNHC2D/3D object composition
Audio, SNHCStructured Audio demo & presentation
Requirements, Video, SNHCReview of common requirements for still images/SNHC texture coding
SNHC VM/PWD Editors Completion of initial SNHC VM 3.0 components, SNHC capability review
by 2100
Systems, SNHCObject libraries, access mechanisms, object naming, download vs. in-place
Systems, SNHC AudioStructured Audio initiative & MSDL architecture/audio composition
Requirements, SNHC Audio & VisualRequirements for SNHC (3D objects, Structured Audio, texture clean-up)
SNHC Work GroupsApplications, profiling, functional grouping of tools, terminals vs. content
SNHC Work GroupsPreliminary VM, PWD, etc.
by 1800
SNHC PlenaryFinal document reviews &approvals (VM, PWD, CEs), resolutions, AHGs
SNHC EditorsDocument preparation for Plenary

Results of these meetings are only briefly highlighted in the following. The joint session with Systems/MSDL on SNHC needs in 2D/3D composition, plus other work already underway with Systems, resulted in adoption of a subset of VRML 2.0 with MPEG-4 node extensions and binary format, as discussed above. Another session with Systems explored MSDL compatibility with object libraries that might be stored on servers or temporarily/permanently stored in the terminal, depending on the application. SNHC had questions about partitioning between MSDL and DMIF, access mechanisms, global vs. local object naming, pre-installation of standard A/V object instances commonly used in certain profiles, and related issues. Some small changes in Mux and buffer management within MSDL, and the on-going work with DMIF, were reported to adequately address these questions. Joint meeting with Requirements on post-Maceió MPEG-4 Requirements identified that SNHC is delinquent in supplying adequate consolidation of given requirements and needed improvements for 2D/3D objects/composition and Structured Audio. This will be dealt with before Bristol.

An extensive and well-attended joint meeting was held with Video on current CE results in texture coding and what to do next. Concern was expressed by some attendees that SNHC-specialized image coding tools will make MPEG-4 decoders overly complex, with related concerns that the Video VM is not being sufficiently exploited to serve the largest common set of requirements. This and the desire for objectivity by both SNHC and Video in evaluating submissions resulted in the realignment of CEs for Bristol along functional lines, as discussed before, with the Video VM adopted as a reference for relevant comparisons. Questions were also raised about the prospect for I-frame coding to use a common technique (whatever is technically superior) for all MPEG-4 still image coding (by improving on the existing tools). A joint meeting with Video/Requirements discussed a proposal for common still image requirements (M1873) within MPEG-4. The effort was identified as very useful, and should help unify tools for common purposes. However, specific concerns were expressed that the document may embody a simultaneous super-set of requirements that are not all needed in specific profiles. A suggestion was made that object-based and spatial scalability should be reviewed to determine if the quantitative requirements can be stratified or reduced. A meeting with Systems on geometry streaming and mesh compression provided insights into VRML vs. MPEG-4 special mesh extensions, and segregating the ASCII/binary formats involved. A meeting with Systems on Structured Audio helped start an architectural review and a related AHG.

7. SNHC Results & Recommendations

SNHC documents, AHG recommendations, CE definitions, and related AHGs and important documents were created:

7.1 VM, PWD & transition to WD
N1545SNHC, HorneMPEG-4 SNHC VM Version 3.0
N1546SNHC, Horn, Rajan Guidelines for Progressing from SNHC VM to SNHC WD
N1547SNHC, HorneMPEG-4 SNHC Preliminary Working Draft

7.2 Formation of SNHC AHGs
N1597Horne, RajanAd Hoc Group on SNHC VM/WD Editing
N1548Petajan, Capin Ad Hoc Group on Face/Body Animation
N1549Nagarajan, Rajan Ad Hoc Group on Media Integration of Text & Graphics
N1550Ebrahimi, Chen Ad Hoc Group on Generic SNHC Object Coding
N1551DoengesAd Hoc Group on SNHC VM/WD Integration into MPEG-4 VM/WD

7.3 Next-round Core Experiment
N1552Petajan, Lavegetto, Chen Core Experiment on FAP/BAP Definition & Compression Evaluation
N1553Rajan, Nagarajan Core Experiments on Media Integration of Text & Graphics
N1554HahnCore Experiment on TTS
N1555Torres, Jordan, Horbelt, Ebrahimi, Wu, Chen, Moccagatta, Liang, Sodagar, Ahn, Tekalp Core Experiments on Generic SNHC Object Coding
N1609Fujii, Vercoe, Scheirer Core Experiments on Structured Audio

7.4 Publication of SNHC document
N1545SNHC, HorneMPEG-4 SNHC VM Version 3.0 with modifications (after two weeks)

7.5 Other SNHC resolutions

CD-ROM - A resolution was made to produce, subject to WG11 approval in appropriate steps, a CD-ROM showcasing the best work of MPEG-4 Core Experiments and other demonstrations, from Video, Audio, Systems, and SNHC, that are worthy of public exposure to convey the potential of MPEG-4.

Standards Tutorial - A resolution was made to enlist an individual and meeting resources to conduct a seminar about the MPEG standardization process to be presented at the next WG-11 meeting in Bristol.

7.6 Related Ad Hoc Groups & Documents
N1532Brandenburg, Moryia AHG on MPEG-4 Audio Core Experiments (including SNHC Audio)
N1544Signes, WardMPEG-4 Systems Methodology and Workplan for Composition
N1594Requirements Group MPEG-4 Profiles Document (Version 2)
N1595Requirements Group MPEG-4 Requirements Document (Version 2)
N1603Tabatabai, Signes, Spille, Petajan AHG on (Systems) Composition
N1610Audio, SNHC, Systems, MPEG-4 Technical Overview Video, Rob Koenen

7.7 Documents & Meetings Needed in Bristol

FAQs and answers for SNHC are to be developed.

Implementation Complexity - Paul Fellows/Marco Mattavelli/Serge Gourrier:
1. Purpose: identify complexity or performance hot-spots
2. Two tools available:
a. Complexity analysis - implementation/platform-dependent
b. Complexity analysis - independent of platform
3. IC helps with tools & reporting, SNHC does profiling & testing
4. Must have people commitment - SNHC responsible for results
5. No AHG for Bristol, but plan an IC orientation session of SNHC

Testing - Michael Zeug:
1. Verification testing
2. Subjective testing

8. Elements of Future SNHC Workplan

This list highlights some key elements of SNHC work planning to be addressed as SNHC features are frozen:

Report of MPEG/Test meeting

Source: Laura Contin, Chairman


The MPEG Test Subgroup met in Sevilla during the 37th meeting of WG11.

The following items were addressed:

  1. July '97 competition tests
  2. Verification tests
  3. Audio-visual library
  4. Potential test sites

July '97 competition tests

As indicated in documents WG11/N1442 and WG11/N1492, Organizations that were intending to submit a proposal for the July '97 competition tests had to pre-register by the 10th of February.

The MPEG doc. N.1756 provides the list of the pre-registered proposals: seven proposals for audio tests and five for video. A further pre-registration for video was accepted during the Sevilla meeting.

In order to simplify the evaluation procedures and speed up the inclusion of the new proposals in the VM, both Audio and Video Subgroups discussed the possibility of assessing the new proposals by core experiment. At the end of the Sevilla meeting only two proposals were left for audio test and three for video test. Proponents, that would have decided after the Sevilla meeting to include their algorithms in the Core Experiment process, still could withdraw their pre-registration by the 3rd of March.

For Audio proposals it was agreed that a pre-screening will be carried out before the Bristol meeting, to provide information that could help the proponents to decide whether to pursue the formal test approach. Procedures for pre-screening are given in an appendix of doc. WG11/N1557, while the workplan is specified in doc. WG11/N1539.

Both for audio and for video formal tests the proponents have been required to cover the entire cost of the preparation and conduction, that was estimated around 40,000 USD for audio and 12,000 USD for video.

Verification tests

During the Maceiò meeting it was agreed to investigate the feasibility of two verification tests:

  1. audio-visual test for the combination of Low Delay Video profile (former real-time communication profile) and Speech audio profile
  2. video test for error robustness

Concerning the first test, the main problems were related to the availability of suitable anchor and test conditions. During a joint meeting with the Low Delay Video Task Group, in Sevilla, it came out that anchor conditions could be available by July '97, but no optimised rate control of the VM would be ready by that time. In order to avoid unfair comparisons it was decided to delay this verification until suitable rate control strategies are defined.

Concerning the error robustness video verification test, in order to achieve test conditions closer to real situation and to simplify the production of the test conditions, in Maceiò it was suggested to use a real-time video decoder and an implementation of the DEMUX. In Sevilla it came out that real-time video decoders could be available by July '97, but the characterization of a suitable DEMUX still requires further discussion with the System Subgroup.

Three solutions can be taken into account:

  1. Do not use any MUX and just evaluate the impact of residual errors on the decoded sequences
  2. To use the FlexMux
  3. To use both FlexMux and TransMux

The first two solutions would not give much more information with respect to what can be obtained by core experiments, while the third one cannot be implemented by July '97.

Considering that the Error Resilience Task Group is still evaluating, by Core Experiments, techniques to extend error resilience to higher bitrates, it was agreed to postpone this verification test until these techniques can be considered stable and implementations of suitable MUX layers will be available.

Concerning the methods to be used in verification tests, a new assessment method was jointly proposed, during the Tampere meeting, by the European Projects Tapestries and MoMuSys. A test to validate this method was carried out by CNET, in the framework of Tapestries. The results of this test are given in the input doc. N.1604, presented at the Sevilla meeting.

Finally, the Audio Subgroup proposed verification tests for AAC in mono and stereo modes. A preliminary test plan was specified in document WG11/N1538.

Taking into account the outcomes of the discussions with Audio, Video and Systems Subgroups, the Test Subgroup has defined a new plan for Verification Tests (WG11/N1557 and WG11/N1558), that will be discussed in an ad hoc group (WG11/N1556) before next meeting.

Audio-visual library

During the Maceiò meeting a number of new audio-visual sequences were proposed to MPEG (WG11/N.1444). Since the duration of these sequences is between 20 s. and a few minutes and most of the test methods are applied to sequences lasting no more than 15 s., it was necessary to select short subsequences from the whole audio-visual sequences provided.

This selection was made by the Test Subgroup and the results are reported in an annex of document WG11/N.1557.

For video source downsampling, the recommendations given in the input document N. 1791 will be taken into account.

Potential test sites

Fondazione Ugo Bordoni (FUB) has expressed in document N.1632 the intention of hosting both video and audio-visual July '97 Tests. FUB is also volunteering to edit the D1 test tapes.

No potential test sites have been yet identified for July '97 Audio Tests.

List of input documents

1604T. Alpert, L. Contin DSCQE experiment for the evaluation of the MPEG4 VM on error robustness functionality
1632V. BaronciniProposal to host the July '97 MPEG-4 Tests
1700Michael ZeugReport of Ad Hoc Group on MPEG4 Low Delay Evaluations
1726Laura ContinReport of the ad hoc group on MPEG-4 July 97 audio, video and audio-visual Tests.
1756Rob Koenen, Laura Contin Registrations for July 97 Tests
1791Andreas HutterConsiderations on the video input library for the July `97 tests

List of output documents

1556ConvenorAdhoc group on evaluation methods and procedures for MPEG-4 tests
1557TestEvaluation methods and procedures for MPEG-4 tests
1558TestWorkplan for MPEG-4 test activities
1538AudioMPEG-2 AAC mono and stereo verification test plan
1539AudioMPEG-4 audio test plans for the July 1997 tests

Annex 11

Report of MPEG Implementation

Source: Marco Mattavelli, Acting Chairman


  1. Review of Computational Graceful Degradation activity.
  2. Review of padding complexity evaluations.
  3. Extension of the ISG complexity analysis to memory requirements
  4. Review of System architecture proposal.
  5. Joint ISG video: -evaluation of the video VM complexity, -evaluation of the complexity of alternative shape coding techniques.
  6. Joint ISG-requirements.

List of Participants

Jan Bormans, Imec, Belgium
Ming-Chien Lee, Microsoft USA
Wei-Ge Chen, Microsoft USA
Young-Kwoon Lim, ETRI, Korea
Serge Gourrier, Philips LEP, France
Marco Gandini, CSELT, Italy
Boon C.S., Matsushita, Japan
Pallavi Shah, Sun Microsystems, USA
Marco Mattavelli, EPFL. Switzerland

Meeting report

  1. Review of Computational Graceful Degradation activities.

Contributions M1636 and M1745 have been presented and reviewed with the ad-hoc group activity report (M1634). Document M1745 proposes a syntax modification so as to enable user defined graceful degradation priorities for multiple video object. The ISG considered favorably the proposal but suggested to define a smaller number of bits for the number of hierarchies. A core experiments will demonstrate the functionality in terms of complexity reduction and image quality, and will define the number of hierarchy levels. The draft description of the CGD core experiment conditions, described in document M1636, have been discussed and approved by the ISG. A final version will be distributed to the CGD ad-hoc group so as to begin the core experiment work. Two companies will participate to the activity. The continuation of the ad-hoc on CGD techniques was recommended.

  1. Review of padding complexity evaluations.

Two contributions evaluating the complexity of the 6 padding techniques under test in the ad-hoc group have been presented (M1786,M1814). The results of the complexity analysis were contradictory since two different approaches have been used to evaluate the complexity. In one case a complete description of the approach was not reported and the ISG group was unable to verify the correctness of the made assumptions. In the other case results have been provided by simulations. The ISG suggested to provide complexity results using objective techniques (iprof profiler or using the instrumentation tools) and to produce comparable results accompanied by a full description of the experimental conditions. The continuation of the ad-hoc activity on padding complexity evaluation was recommended.

  1. Extension of the ISG complexity analysis to memory requirements.

The ISG discussed the need to evaluate memory requirements for the VM. An activity considering the extension of the complexity evaluation tools to memory evaluations or the development or use of new tools was volunteered by IMEC. IPR issues have to be checked first.

  1. Review of System architecture proposal.

An overview of the status of the system architecture was presented to the ISG. The ISG discussed the definition granularity of the clock reference and time stamps. The ISG noticed the lack of mechanisms able to ensure synchronization and the lack of a hierarchy for the handling of objects in the scene. The possibility of defining a feedback information to the encoder to acknowledge the correct behavior of the decoder or the occurrence of computational overloading situations has been discussed. Further analysis is needed before discussing the issues with the System group and to give specific recommendations. It was agreed with the System group to put the items as discussion points for the next Bristol meeting.

  1. Joint ISG video: -evaluation of the video VM complexity, -evaluation of the complexity of alternative shape coding techniques.

In a joint Video-ISG meeting, the problem of the complexity evaluation of the VM has been discussed. The ISG presented the two approaches developed and suggested for the complexity analysis: the iprof profiling tool and the instrumentation tool as developed within the Emphasis project. Main characteristics and limitations of both tools have been discussed. A smaller joint meeting with the shape coding video subgroup was held. The candidate alternative shape coding techniques have been presented to the ISG group. Unfortunately no advise about the complexity of such techniques can be given without the use of the suggested tools. Theoretical analysis or worst case analysis are not able to consider and to measure the statistical occurrences of the various coding modes and consequently the overall average complexity. The establishment of an ad-hoc for the evaluation of VM and candidate shape coding techniques was decided. This ad hoc will be chaired by Simon Winder (Microsoft).

6) Joint ISG-Requirements.

A short joint meeting with the Requirements group discussed the decoder complexity of the two possibilities: downscaling or scalability/simulcast. The ISG expressed the opinion that from the point of view of decoder complexity scalability/simulcast are certainly preferable than downscaling a bitstream to a lower resolution. This has to be discussed with Video at the Bristol meeting.

Annex 12

Report of MPEG Liaison

Source: Barry Haskell - Liaison Chairman

The Liaison group considered input documents

Liaison from IEEE concerning cancellation of IDCT specification IEEE1180 The IEEE claims the standard is flawed and needs to be revised. MPEG contains enough IEEE members to undertake this task if there is sufficient interest. A reply to IEEE was drafted.

Liaison from SMPTE concerning Registration Authority for MPEG-2 format_identifier This will be considered by SC29.

Liaison from DVB concerning DVB TM 1214 rev 7 Their digital VCR interface bitstream continues to be non compliant to MPEG-2 Systems. We request a copy of ISO/IEC 100C/1883 from the SC29 secretariat for possible comment.

MPEG97/1802 from ITU-T AVC group expressing interest in DMIF. A reply may be drafted after the April MPEG meeting.

MPEG97/1803 from ITU-T T.130 group expressing interest in DMIF. T.130 aims to provide facilities to manage interactive multipoint communications on a variety of networks. A reply may be drafted after the April MPEG meeting.

W3C Activity: Real Time Multimedia W3C would like to be a forum for exchange of ideas. A reply may be drafted after the April MPEG meeting.

The Internet Multimedia Conferencing Architecture The IETF activity on Multiparty Multimedia Session Control (MMUSIC) aims at multipoint communication on the Internet. A reply may be drafted after the April MPEG meeting.

H.Loose Issues This ITU-T activity is aimed at conferences/broadcasts to loosely connected participants. A reply may be drafted after the April MPEG meeting.

Report of ad hoc Group 1489 on MPEG-4 Systems Evolution This group recommended collaboration with SC24. It was decided that collaboration directly with VRML, which is the corresponding technical working group, would be more productive.

The following output liaison documents were produced:

WG11/N1568 Liaison statement to VRML

WG11/N1569 Liaison statement to IEEE Standards Board

WG11/N1570Liaison statement to ITU-T SG16 LBC Group