&%PAGE&

-  &%page&  -
Document AVC-22R
CCITT SGXV                                            	     Document AVC-22R
Working Party XV/1                        			     November 16, 1990
Experts Group for ATM Video Coding
         
SOURCE : CHAIRMAN
TITLE  : REPORT OF THE FIRST MEETING OF THE EXPERTS GROUP FOR ATM VIDEO
         CODING IN THE HAGUE (November 13 - 16, 1990)
PURPOSE: Report

-------------


1. General

The first meeting of the Experts Group was held in The Hague (The 
Netherlands) at the kind invitation of PTT-RNL.  On behalf of the hosting 
organization, Mr. Ronald Plompen made a welcome address.  The list of 
participants appears at the end of this report.  The meeting started with 
introduction of each participant.
 
As to the Coordinating Member of The Netherlands, Mr. Dolf A. Schinkel 
advised the meeting that he replaced Mr. Plompen.

At the final session, Chairman thanked the hosting organization for the 
meeting facilities provided and the excellent secretarial support.


2. Documentation (TD2)

For this meeting, 21 normal documents and 13 temporary documents were 
available as listed in Annex 1.


3. Organization and role of the Experts Group

3.1 CCITT (AVC-2)

The meeting confirmed the terms of reference, the working methods and the 
work plan which had been guided by Working Party XV/1 at its July 1990 
meeting.

3.2 Regional coordination (AVC-7,13)

Mr. Wada and Mr. Schinkel presented regional coordination in Japan (AV&ATM 
Video Coding Subcommittee) and Europe (NA5/VCM, TE/AVM, COST, RACE, ESPRIT) 
for this Experts Group using AVC-7 and AVC-13 respectively.  Mr. Schaphorst 
presented North American coordination taking place in T1Y1.1, T1Q1.5 and 
T1S1.  Mr. Biggar presented activities in Australia.


4. State of the art in ATM video coding (AVC-10,13,17,19; TD3)

A number of activities in the field of ATM video coding were briefed by 
participating organizations with indications of essential results.  Some of 
them were accompanied by video tape demonstrations as listed in Annex 2.


5. Applications where ATM video coding is used (AVC-8,15,16)
 
Possible applications in B-ISDN and their requirements were provided in AVC-
8 and 6, which are summarized in general term as;

     - conversational services,
     - distributive services,
     - retrieval services,

with stress on their multimedia nature.  It was also pointed out that Draft 
I.211 gives an extensive list of B-ISDN services.

During the discussion, the meeting recognized that our objective is to 
define a unified coding which can cover the above mentioned services, 
rather than to confine to a specific service.


6. Boundary conditions for video coding

6.1 Network characteristics (AVC-4,17,20; TD12)

ATM network characteristics to which our new video coding should adapt 
were overviewed in AVC-4 and AVC-20, listing up opportunities as well as 
limitations.  AVC-17 provided a list of questions concerning ATM network 
characteristics which affect the design of the video coding.  The meeting 
recognized it necessary to enhance this list for the group's future work, 
and requested Mr. Biggar to formulate questions addressed to SGXVIII which 
will meet at the end of November, 1990.  The outcome is contained in Annex 3.

The meeting also discussed for a while on the networks to which our codec 
will be connected, concluding that we focus on B-ISDN but do not preclude 
such networks as LAN and MAN as far as they are ATM based.  During this 
discussion, it was pointed out that policing is required even in CBR services 
for safety margin reason.

6.2 Technical requirements for ATM video coding (AVC-8,13,15,16,17)

6.2.1 Video signals to be handled

Initially we concentrate on video coding of standard television signals but 
at  the same time try to take into account its extension to EDTV and HDTV.

6.2.2 Picture quality target

The target is defined as a range between conversational service quality and 
distribution service quality, awaiting the quantification in the future.

6.2.3 Processing delay target

Processing delay of the new video coding should be less than that of the 
current systems. 

6.2.4 Average bit rate

It should cover a range from 64 kbit/s to several tens of Mbit/s.

6.3 Compatibility issues (AVC-5,9,13,14,15,16)

There was some discussion on the balance of achieving compatibility and 
highest coding performance.  The meeting agreed on a guideline that the 
compatibility between the new coding system and existing systems should be 
highly respected.  Exact ways to implement this property requires further 
study.

6.4 System aspects other than video coding (AVC-4,10,20)

The meeting discussed on how to handle such system aspects as multimedia 
multiplex.  Though it is obvious that this Experts Group is responsible for 
video coding aspects, we are in a position to raise requirements to WPXV/1  
which is responsible for systems aspects of audiovisual systems using B-ISDN.

The meeting confirmed that we will study these system aspects to the extent 
that it facilitates video coding study.


7. Video coding problems to be worked out (AVC-10,13,17)

7.1 Coding architecture (AVC-5,10)

An example of video coding architecture to cope with various services and 
performance was presented in AVC-5, based on layered coding.  AVC-10 also 
required versatile coding without specifying a particular method.  The 
meeting agreed that the Group aims at a universal coding algorithm in terms 
of services, quality, resolution, applications and bit rates each of which 
are given as a range.

7.2 Variable bit rate coding vs constant bit rate coding 
    (AVC-4,10,13,15,17,20; TD11)

Since VBR (variable bit rate) for constant quality is considered as one of 
the outstanding features of ATM, the meeting recognized it as an urgent 
study item  to clarify the advantages of VBR video coding against CBR 
(constant bit rate) video coding.  It is also a common understanding that 
applications should be clarified where VBR is effective.

In the course of discussion, the meeting felt it would facilitate our making 
a good progress toward the next meeting to agree on a common network 
model, even if simple and crude, during this meeting.  Mr. Tabatabai 
undertook to coordinate a small group for this task.  The outcome is 
contained in Annex 4.  The meeting agreed to invite comments on this model 
from SGXVIII (see Annex 3).
 
Comparison in video coding efficiency (e.g. one layer coding vs two layer 
coding) based on this model as well as proposals for improvement of this 
model are encouraged.  Elaboration of the model may include cluster type of 
cell loss pattern, one more model appropriate for another QOS and/or higher 
bit rates.  It was also pointed out that the measuring window will be 
critical when discussing VBR.

7.3 Layered coding (AVC-5,13,17,20)

Documents were presented, showing that the intention of layered coding is 
either of to obtain compatibility among different service classes or to cope 
with cell loss (equivalently to make use of statistical multiplexing).  As to 
the latter aspect, see the discussion in 7.2 above.

7.4 Picture formats (AVC-5,10,16)

For the simulation purpose, we will initially deal with QCIF/CIF and CCIR-601 
formats.  The meeting recognized the idea of a single coding format for 
CCIR-601 level pictures, but the details await further study.  Contributions 
are requested to clarify advantages and disadvantages of that idea for 
making a decision.  


8. AAL and multimedia multiplex suitable for video services 
   (AVC-6,11,18; TD12)

Extensive information was presented to understand the relation between AAL 
and multimedia including video was presented.  Mr. Tabatabai made a comment 
that;

     -	SN may or may not be needed.
     -	VPI/multiple VCIs may be used for multimedia multiplex.
     -	Instead of using IT, one bit in AAL can be used as a protocol 
discriminator to indicate whether a pointer for boundary indication 
will follow.

Mr. Elewaut pointed out parity cells may be inserted for error protection 
using CS as implemented in the video codec described in AVC-19.  

After questions and answers, the meeting confirmed that we should 
concentrate on Type 2 and provide our study results for SGXVIII (see Annex 
3).  It is also a common understanding that SAR is service independent while 
CS is service dependent. 


9. Work plan and work method

9.1 Work plan (AVC-9,16)

There was consensus on the following work plan of the group;

     -	Final Recommendation be made official in 1994, taking into account the 
completion of the B-ISDN Recommendations in 1992 and subsequent 
service provision.
     -	Outline Recommendation be produced at the end of the current study 
period, which includes scope, list of contents, such parameters as 
picture formats, framework of coding scheme, etc. to be agreed by that 
time.

9.2 Work method (AVC-10,13,16)

The following methods practiced in the previous Specialists Group for H.261 
were supported;

     -	Study is phased as "divergence" and "convergence",
     -	Step by step using Reference Models, and
     -	Hardware verification at the final stage.

As to the reference model, it was clarified that this time we need two kinds 
of model; one for network aspects study and the other for video coding 
aspects study.  It was also clarified that the latter includes source coding 
as well as channel coding.

The meeting considered when the first Reference Model for video coding be 
defined, and until when various types of candidate algorithm be tried, 
concluding that the both timings coincide with the demarcation between the 
"divergence phase" and the "convergence phase". 

In order to make progress in video coding simulation work as soon as 
possible, the meeting considered that at least test pictures and assessment 
methods should be agreed at this meeting.  The conclusion is as follows;

a) Test pictures

     -	Existing CIF test sequences: SALESMAN, CLAIRE, MISS AMERICA, BLUE 
JACKET, SWING
     -	MPEG phase-2 sequences: FLOWER GARDEN, SUSIE, POPPIE, TABLE TENNIS, 
MOBILE & CALENDAR, TEMPETE 
       	Note: Use of these sequences and distribution among the Experts 
Group members are subject to negotiation with MPEG. 
     -	A long sequence but with average complexity for testing VBR for 
conversational services: Belgium kindly offered to try to generate 
such a sequence in 625/50 and 525/60 formats.

b) Assessment method

Informal observation tests were supported for our current purpose until 
other more effective methods become of use.  Mr. Plompen pointed out that 
there exists an objective measurement method for higher bit rates to assess 
coded pictures.

Since we are going to assess CCIR-601 pictures as well as CIF pictures, a D-1 
machine is necessary for review of processed pictures in addition to a U-
matic machine in future meetings.

9.3 Time table (TD-5,10)                                                    

As a summary of the discussion for the work plan and method, the time table 
as shown in Annex 5 was agreed.

9.4 Status report document 

Mr. Plompen proposed to produce a status report document such as AVC-13 
including the following items;

     - Prose description
     - Terminology
     - Applications
     - Coding models
     - Network model
     - Assessment
     - Reference Model
     - Guidelines for simulations
     - Outstanding questions
     - Action points.

The objective is to form a common ground among participating members and 
clarify what should be worked out.  Due to the lack of time at the meeting, 
Chairman undertook to draft such a document after the meeting based on the 
achievements in this meeting and to circulate it for the members' comments.

10. Harmonization with other groups 

10.1 MPEG and CMTT plans (AVC-3,12; TD-5)

Chairman and Mr. Carbiere presented the organization and relevant work 
plan of MPEG phase 2 and CMTT, respectively to form a background of 
discussion.

10.2 Guideline (TD-6)

Since the decision on how to harmonize with other groups influences 
discussion on all other items, the meeting dealt with it at the earlier part 
of the agenda.

A basic question was whether this Group should work aligned with ISO and 
CMTT.  After having a free discussion, the meeting agreed in principle to 
carry out joint work in order to avoid different standards in the same or 
similar areas and to avoid duplication of standardization work as well.  

As to the details for materializing this "joint work", however, there remain 
many items to be clarified.  The followings are such comments obtained 
during the discussion;

     -	A simple liaison statement to propose joint work is not sufficient.  We 
should indicate our own technical requirements and scope of service 
aspects.
     -	Boundary conditions to the video coding should be debated with ISO.
     -	We should try to keep our work method.
     -	It should be distinguished whether the target is to have a common 
standard or a common coding algorithm.  Commonality should be sought 
to the necessary extent.
     -	Multimedia multiplex aspects should also be considered for joint work.  
MPEG has activities in this area to realize editing of multimedia 
signals, random access, channel switching, etc.
     -	There is not so much relation between ATM transmission and DSM 
applications.
     -	Time schedule should be checked whether compromisable or not.  
Success of the joint work depends on it.
     -	Advantages of B-ISDN should be sought as a long term objective.  
Compatibility with existing standards is a short term objective.
     -	Video coding aspects may need joint work with ISO, but network 
aspects are of CCITT's responsibility.
     -	Initiation of the joint work is the sooner the better.
     -	Submission of technical contributions from this group will give 
impacts on ISO decisions.
     -	We must cope with administrative nightmare.

10.3 Specific ways of collaboration (AVC-14; TD-7,8,9)

After having heard suggestions on several possible ways of collaboration, 
the meeting asked Mr. Morrison to coordinate a small group to further 
discuss the matter and draft a liaison statement to MPEG.  The outcome is 
contained in Annex 6.  This statement is expected to be discussed at the 
next MPEG meeting to be held during December 4 - 7, 1990 in Berlin.

During this discussion, Mr. Brusewitz undertook to draft a list of 
provisional requirements of ATM video coding as contained in Annex 7, which 
is included in the liaison statement.

A similar liaison statement to CMTT was felt needed.  Taking into account 
that TG CMTT/2 will meet next April, the meeting requested Mr. Carbiere to 
draft such a statement for discussion and decision at the next meeting of 
the Experts Group.


11. Others

11.1 Hardware demonstrations

Chairman raised how to manage voluntary hardware demonstrations which may 
be proposed by participating organizations.  The meeting agreed that we 
would accept such demonstrations as far as they facilitate our deepening 
understanding of the problems.

11.2 Next meeting

The second meeting of the Experts Group will be held during March 12 - 15, 
1991 in Tokyo, but subject to change according to the action taken as 
described in 10.3 above.


END

----------

Annexes 

Annex 1  Available documents 
Annex 2  List of tape demonstrations
Annex 3  Liaison statement to SGXVIII
Annex 4  First simplified network model
Annex 5  Time table of the Experts Group work
Annex 6  Liaison statement to MPEG
Annex 7  Provisional list of H.26x requirements

List of Participants of the first meeting 
of the Experts Group for ATM Video Coding 
(13 - 16 November 1990, The Hague)


FRG          Mr. P. Klein		Siemens AG
		Mr. G. Zedler	 	DBP Telecom			CM

Australia    Mr. M. Biggar	 	Telecom Australia			CM

Belgium      Mr. L. Elewaut	 	Alcatel Bell Telephone
		Mr. O. Poncin	 	University of Louvain
		Mr. W. Verbiest	 	Alcatel Bell Telephone		CM

Canada       Mr. D. Lemay		BNR

USA		Mr. P. Alexander	 	PictureTel
		Mr. B.G. Haskell	 	AT&T Bell Labs
		Mr. R.A. Schaphorst 	DIS					CM
		Mr. A.J. Tabatabai	 	Bellcore				CM
		Mr. C.L. Yeh		Apple Computer
		Mr. J. Zdepski	 	David Sarnoff Research Center

France       Mr. G. Eude		France Telecom
		Mr. J. Guichard	 	France TElecom			CM

Italy        Mr. M. Guglielmo	 	CSELT				CM

Japan        Mr. Y. Kato		Mitsubishi Electric
		Mr. S. Okubo	 	NTT					Chairman
		Mr. T. Tanaka	 	NTT					CM
		Mr. M. Wada		 	KDD					CM
		Mr. M. Yano		 	NEC

Norway       Mr. G. Bjoentegaard	Norwegian Telecom
		Mr. H. Sandgrind	 	Norwegian Telecom			CM

Netherlands  Mr. H. Carbiere		PTT Research			LR (CMTT)
		Mr. A. Koster		PTT Research
		Mr. R. Plompen		PTT Contest
		Mr. D. Schinkel		PTT Research			CM
		Mr. B. Schuurink		PTT Research

UK           Mr. M.D. Carr		BT
		Mr. D.G. Morrison		BT					CM

Sweden	Mr. H. Brusewitz 		Swedish Telecom			CM
		Ms. C. Verreth		Swedish Telecom

Switzerland  Mr. H. Keller		Ascom Tech 			CM


								CM: Coordinating Member
								LR: Liaison Representative