&%PAGE&

-  &%PAGE&  -
Document AVC-317R
CCITT SGXV                                            	Document AVC-317R
Working Party XV/1                        			July 10, 1992
Experts Group for ATM Video Coding				

SOURCE : CHAIRMAN OF THE EXPERTS GROUP FOR ATM VIDEO CODING
TITLE  : REPORT OF THE SEVENTH MEETING IN NEW JERSEY AND RIO DE 
         JANEIRO (July 1-10, 1992) - PART I
Purpose: Report

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

Part I (Sole Sessions)

Contents

		   1.	General
		   2.	Documentation
		   3.	Tape demonstration
		   4.	Review of the previous meetings relevant to 
			the Experts Group
		   5.	Picture format
		   6.	Source coding
		   7.	AAL and other network aspects
		   8.	Multimedia multiplex
		   9.	Hardware verification of H.26X
		  10.	Very low bit rate video coding
		  11.	Preparation for the joint sessions
		  12.	Others


1.	General

The seventh meeting of the Experts Group consisted of two parts; CCITT sole 
sessions in New Jersey and joint sessions with ISO/IEC JTC1/SC29/WG11 (MPEG) 
in Rio de Janeiro.  The list of participants appears at the end of this 
report. 

The first part was held during 1-3 July 1992 at Sheraton Hotel in Eatontown, 
New Jersey, USA, at the kind invitation of AT&T Bell Laboratories and 
Bellcore.  At the opening session, Dr. Jules Bellisio, Bellcore, made a 
welcoming address on behalf of the hosting organizations.  

At the end of the New Jersey sessions, Chairman thanked the hosting 
organizations for the meeting facilities provided and the excellent 
secretarial support.  

The Experts Group also had a short closing session on 10 July in Rio de 
Janeiro.

The following change of the membership of the group was announced;

    -	Participation of Korea had been approved at the SGXV meeting in May 
with its Coordinating Member Mr. J-Y. Nam.
    -	Mr. H. Carbiere had resigned Liaison Representative to CMTT/2.

The Experts Group regretfully accepted Mr. Carbiere's resignation and 
solicited his successor.

2.	Documentation (TD-2)

For this meeting, 61 AVC-numbered documents and 8 temporary documents were 
available as listed in Annex 1.


3.	Tape demonstration (TD-3)

Several video tape demonstrations were given with D1 or U-matic as detailed 
in Annex 2 to present experimental results.  There were also the following 
hardware demonstrations;

    -	AT&T videophone,
    -	GEC-Marconi videophone.


4. 	Review of the previous meetings relevant to the Experts Group

4.1 WPXV/1 in May (AVC-261,262,263,264,265)

Chairman presented the outcome of the WPXV/1 meeting held in May, focusing 
on the following topics;

    -	H-series new and revised Recommendations 
    -	joint work with MPEG
    -	PSTN/mobile video telephone

4.2 CCIR, CMTT in May

Mr. Zedler reported that the May CCIR recognized the work of MPEG and 
decided to prepare requirements from the broadcasting point of view, and 
that an Special Rapporteur (Mr. D. Nishizawa, NHK - Japan) was appointed with 
two assisting members for this purpose.

4.3 SGXVIII in June (AVC-311,312)

Chairman introduced briefly the liaisons statements received from Mr. K. 
Yamazaki, Special Rapporteur SWP XVIII/8-3.


5. 	Picture format (AVC-274,295)

AVC-274 presented simulation results on 50Hz/60Hz mutual conversion, and 
AVC-295 reinforced the flexible spatial resolution approach.  An answered 
question is what picture format become a major on B-ISDN.  The first 
document is assuming two versions of CCIR-601, and the second document is 
assuming window-based multimedia workstations and PCs.  There was a comment 
that US sentiment for HDTV is somewhat in line with the second document.  We 
need application scenarios as recognized in Stockholm.

Since H.26X should be generic and be able to cope with a wide range of 
formats, the current format issue is rather of H.32X terminal.  During the 
discussion, it was stressed that the problem is how to achieve interworking 
between various types of terminals with different source formats which are 
connected through B-ISDN.  One way is to define a default format every 
terminal should have, the other way is to define that every decoder of the 
same class be able to decode and display any format smaller than a defined 
maximum.

The meeting confirmed the previous decision in Stockholm to endeavor to 
reach a common understanding in Ipswich next October.


6.	Source coding

6.1 Coding structure

	6.1.1 TM1 (AVC-260)

TM1 was defined at the Haifa meeting.  The agreed definition of coding 
schemes and experiments have been edited as AVC-260 by the effort of 
Editing Committee.

	6.1.2 Prediction optimization (AVC-280,281,282,285,301)

It has been recognized that comparison between various predictions 
(field/frame adaptive, field, pure field, FAMC) need data on the following;

    -	picture quality,
    -	inherent delay,
    -	hardware complexity.

Documents AVC-280,281,282 provided such information, particularly on FAMC.  
The Experts Group considered this topic from from the low delay mode (see 
Section 6.2).  These and other comparisons will be the major topic in Rio de 
Janeiro.  During the discussion, the following comments were noted;

    -	The number of memorys is one of the most important factors for 
implementation study.
    -	FAMC requires a larger number of memory accesses than other 
predictions; this will affect hardware complexity.
    -	Videoconferencing applications will be benefited by a particular 
prediction.
    -	Since multimedia applications use film and computer graphics sources, 
frame based prediction is preferable.
    -	Decoder complexity does not differ so much even if we allow multiple 
predictors.

Document AVC-301 proposed an generalization of prediction which consists of 
a generic prediction vector, up to four pixels in the reference picture 
which are address by the generic prediction vector, and their weighting 
values to obtain a prediction value (see Annex 3).  The reference pixels and 
their weighting may be down loaded by the video signal itself or pre-defined 
according to the application.  The current proposal covers all the 
predictors contained in TM1.

The meeting recognized that this would be a solution for the prediction 
issue, if we really need a different predictor for each sequence because it 
is significantly advantageous compared to other predictors.  Generally, the 
following predictions are suited;

    -	frame coding for still or slowly moving pictures,
    -	FAMC for moderately moving pictures,
    -	field coding for rapidly moving pictures.

It was pointed out this solution might ease the standardization process and 
that it reduces number of macroblock types.  However, the following items 
require further consideration;

    -	Can one predictor not be practically sufficient?
    -	Impact of picture distance (value of M) and use of picture dropping
    -	How B pictures be processed?
    -	How to cope with adaptive prediction on a macroblock basis, if it is 
necessary?
    -	Loading interval for the tables depends on the application. 

	6.1.3 Adaptation to "tune-in" (AVC-310)

Document AVC-310 proposed inclusion of quantizer matrices in the picture 
header and use of leaky prediction for channel hopping in broadcasting 
applications.  The meeting was aware of that this tune-in feature also 
facilitates switching type multipoint communications.  

The meeting supported the first proposal.  As to the leaky prediction, the 
meeting decided to consider it in the context of cell loss resilience.  
During the discussion, it was clarified that the leaky prediction provides 
graceful build-up compared with the delayed start due to the use of 
periodic intra pictures, and that the loss of coding efficiency is 
compensated by non-use of periodic intra pictures.  

6.2 Low delay mode (AVC-260)

	6.2.1 Performance objectives (AVC-266)

Summary of SGXII study results on the delay in audiovisual services was 
provided by a liaison statement to SGXV.  Recommendation G.114 stipulates 
that one way delay of 150 ms is "acceptable for most user applications" if 
echo is adequately controlled.  Information on longer delay considerations 
are also given in this liaison statement.

	6.2.2 TM1 experimental results and their implications 
             (AVC-275,276,284,291,302)

Several topics regarding the low delay mode were addressed by respective 
contributions as follows;
    
1) coding structure; values of {N,M}, use of forced intra slices

	AVC-275,276,284,291,302

It was pointed out that pure field coding with M=3 may also provide less 
than 150 ms coding/decoding delay in addition to the TM1 low delay mode.

2) optimum prediction; field/frame, field, pure field, dual field, FAMC

      AVC-275,276,302

Document 302 describes an FAMC like field prediction, which indicates that 
the idea of FAMC is also applicable to the field coding.

3) first picture in TM experiments

	AVC-275,291

There were found two different ways to process the first picture in the TM1 
experiment; coding it as a intra picture (field or frame) and then starts 
forced intra slice from the second picture, and coding the first one/two 
slices only at the first picture, then the first two/four slices at the 
second picture and so on. 

4) scene change handling

	AVC-275

Tape demonstration accompanying Document AVC-275 showed two possible 
solutions for scene change handling in the low delay mode; picture dropping 
and wiping.  The meeting found the former solution less annoying, though it 
was noted that the repeating the frame before the scene change may produce 
motion jittering.  It was also commented that if picture dropping takes 
place in the MPEG1 system multiplex, time stamp for synchronizing audio and 
video may not work.

There was a question whether the picture dropping be used for VBR, but this 
is open at the moment.

5) rate control method

	AVC-284

Visual moving of forced intra slices were shown in the demonstration 
accompanying Document AVC-284, but they were not recognized in the 
demonstration accompanying Documents AVC-302,310.  We need further 
clarification on this aspect. 

	6.2.3 Core experiments for the low delay mode

After discussion, the meeting asked Mr. Bjoentegaard to coordinate a small 
group for drafting further core experiments to clarify the issues identified 
above in Section 6.2.2.  The outcome is contained in Annex 4.

The meeting reviewed this outcome and agreed to forward it to the joint 
sessions in Rio de Janeiro.

6.3 Compatibility and scalability 

	6.3.1 Experimental results and their implications (AVC-277,286,293,313)

These four experiments provide information on the comparison of embedded 
vs simulcast, indicating the embedded coding gives better SNR than the 
simulcast.  AVC-277 also provides experimental results on comparison of 
prediction from the base layer vs prediction of the prediction error, 
favoring the first scheme in SNR performance.  

There was some discussion on necessary hardware for the embedded decoding 
related to the required number of decoder loops for the full resolution 
only decoder.  If we use high speed processor for two or more loops in a 
time shared manner, the number of loops does not count.  This structure 
does not increase coding and decoding delay due to the use of low data rate 
base layer for prediction of the enhanced layer.

It was clarified for the prediction of the prediction error scheme that the 
enhanced layer decoder receives the base layer motion vector but does not 
use it for reconstructing the full resolution picture.

	6.3.2 H.261 compatibility (AVC-278)

AVC-278 clarified H.261 compatibility requirements for H.26X coder and 
decoder by analyzing several communication patterns.  The meeting noted 
that all of these situations should be taken into account when we discuss 
the compatibility issue. 

If embedded scheme is adopted for H.26X, every coder and decoder should 
have the embedded structure.  In a point to point communication between two 
H.26X terminals, embedded coding may degrade performance, thus a mechanism 
to switch off the base layer as necessary should be considered.

	6.3.3 Scalability (AVC-292,299,306)

Three different scalable structures were presented in these three 
documents.  The meeting simply took note of the information at this moment 
because the compatibility, H.261 compatibility in particular, is a more urgent 
issue from the CCITT view point. 

	6.3.4 Core experiments for H.261 compatibility 

After the above discussion, the meeting asked Mr. Parke to coordinate a 
small group to plan H.261 compatibility experiments.  The outcome is 
contained in Annex 5.  The meeting reviewed this report and decided to put 
forward the experiment proposal to the joint sessions.

6.4 Cell loss resilience 

	6.4.1 Experimental results and their implications 
             (AVC-269,279,287,298,308; AVC-310)

The following techniques were experimented as cell loss resilience measures;

1) Structured packing: AVC-269,279

Inclusion of video date address information in a cell and localization of 
damaged area were found very effective for cell loss resilience.  There was 
a question whether this technique is also applicable to multimedia 
multiplexed audiovisual signals such as H.221 structured signals, and a 
comment that the number of bits for the overhead indicating address etc. 
should more conveniently a multiple of 8.   

2) Leaky prediction: AVC-279,310

The meeting desired that the contribution of leaky prediction be separated 
from that of error concealment in AVC-279.  Mr. Sakai and Ms. Reibman 
clarified that the value of leaky factor sensitively affects efficiency and 
recovery time and that the optimum value depends on the input picture.

3) Concealment: AVC-279,308

Both experiments showed that concealments supported by suitable AAL 
(including indication of errored cells and slice/MB level re-entry) provide 
an effective means to make either one- or two-layered MPEG coding cell loss 
resilient.  Dr. Raychaudhuri summarized that a specific AAL/error 
concealment for MPEG-1 provides reasonable pictures (corresponding
approximately to "entertainment quality" in the US) at the following cell 
loss ratios;

 One-layer:	CLR ~ 10E-4 to 10E-6 range (depending on decoder concealment
capability)

 Two-layer:	low priority CLR ~ 10E-2, high priority CLR ~ 10E-5

He also informed the meeting that his HDTV experiments (with 120 byte cell 
payload) indicate that cell size has a relatively small impact on the 
resilience.

4) Layered coding: AVC-287,298

In response to a question, it was clarified that if CLP is used on a cell by 
cell basis in a Virtual Channel, the network guarantees cell sequence 
integrity, thus there is no need to synchronize the two priority bitstreams.  
The layered coding in AVC-287 uses only the upsampled base layer prediction 
error with a upscaled motion vector if a macroblock in the enhancement 
layer is hit by the cell loss, while the one layered coding uses only 
resynchronization of the next slice.

	6.4.2 Error resilience improvement for TM1 syntax (AVC-288)

As to the proposed addition of the two bit after the extension code, the 
meeting did not understand its effectiveness, thus decided not to support 
this proposal until further clarification is obtained.

There was a comment on the picture slicing that the real time coding may 
utilize this technique depending on the transmission media, but the pre-
recorded signal can not.

The meeting confirmed that H.261 IDCT specifications stand also at higher bit 
rates envisaged for H.26X, because refresh frequency is defined in terms of 
"once every 132 times coded".

	6.4.3 Core experiments (AVC-260,300)

After the above discussion, the meeting asked Mr. Biggar to coordinate a 
small group to design core experiments addressing the following items;

    -	clarification of the effectiveness of those techniques listed in 
Section 6.4.1 above.
    -	Techniques applicable to single layered coding
    -	Techniques applicable to multiple layered coding

The outcome is contained in Annex 6.  After reviewing this report, the 
meeting decided to put forward this proposal to the joint sessions.

6.5 VBR

	6.5.1 Statistics (AVC-267,307,314)

These three documents provided various statistics of the VBR video, which 
were collected for open loop operation of three different coders (BBE 
hybrid DCT, MPEG-1, H.261).  The meeting appreciated this information for 
studying multiplex gain of the VBR traffic.  There was a question whether 
these statistics may change if the average rate is regulated according to 
the preventive policing of the coder.

	6.5.2 Impacts of UPC (AVC-271,272,312)

AVC-272 provided a delay analysis for CBR and VBR and indicated a 
particular case where the sliding window regulation of the average gives 
benefit to the VBR operation; MPEG-1 like periodical variation of the 
information generation.  AVC-271 reported supporting experimental data.

The liaison statement from SGXVIII indicated that the traffic parameter for 
the average is likely to be defined in terms of a reference algorithm as 
currently done for the peak rate in I.371.  The Experts Group is greatly 
concerned with the time constant of the reference algorithm.  If it is short 
in the order of e.g. 1 ms, the coder should regulate the coding rate much 
more tightly than the current CBR system where a transmission buffer of 
several tens of ms is used.  This should be communicated with SGXVIII.

	6.5.3 Cell loss for multiplexed VBR sources (AVC-296)

Long term characteristics of cell loss were presented for multiplexed VBR 
video sources.  An implication is that once congestion takes place, very bad 
condition with CLR=0.01 may continue for tens of frame time due to 
autocorrelation of the video source signal.  Study of visual effects of 
CLR=0.01 was proposed.  There was a general comment whether the result 
would change if the source is regulated in average by a leaky bucket 
mechanism etc.

If the source coding should cope with this high CLR by using e.g. layered 
coding, redundancy should be included in the coded video, thus the number 
of channels to be accommodated decreases.  Another way to cope with this 
situation may be to accommodate a slightly less number of channels 
containing less redundant coded video.

This periodic bunching may happen for non video traffic.  The meeting 
recognized necessity to communicate with SGXVIII on this matter.  The coming 
IVS Technical Session will be a good opportunity to exchange views on this 
and other network related topics.  
 
Possible feedback of this study result on the current network model was 
raised, and it was clarified that this result will supplement the network 
model, but not change.  The exact way of reflection is for further study.


7.	AAL and other network aspects

7.1 Clock recovery (AVC-273,315)

AVC-273 addressed the following items regarding the video clock recovery;

    -	supported by AAL or video codec?
    -	what method be used?  what field be prepared for this purpose?
    -	what method if no common clock is available?

Though we could not reach a firm conclusion, there was general support for 
the video codec to have this functionality.  Study for exact way and other 
aspects are to be continued.

AVC-315 provided a solution for timing recovery of audiovisual signals 
transmitted through a variable rate channel where a common clock is not 
available between the encoder and decoder.  This work is based on MPEG 
system study; members of the Experts Group are encouraged to learn its 
wisdom.  It was also pointed out that this solution is useful for the 
database access where pre-recorded video is involved and a common clock 
does not solve the problem even if it is available at both ends.

7.2 AAL for video signal transport (AVC-294,297,312; AVC-298)

Both AVC-294 and AVC-297 provided example AALs for video signal support to 
stimulate the group, while AVC-312, liaison statement from SGXVIII, listed 
required functionalities of AAL for video signal support.  The key issue is 
how generic AAL should be, whether full functionality be in AAL or the user 
data (see also TD-6).  One possibility would be to define a null AAL, and 
contain all necessary fields in the video data. 

We should take the following steps;

    -	To identify required functionalities for audiovisual system, 
particularly for video,  
    -	To identify where each of those functionality be supported; in AAL or 
in the user data.  The guideline should be that the network related 
functionalities should be supported by AAL, but video related ones by 
the user data.

There was expressed some concern about the interface between the AAL and 
the (video) user.  If it is clearly defined, the ATM video codec can also be 
applied to other networks than B-ISDN.

7.3 Support of H.320 terminals on B-ISDN (AVC-312)

SGXIII sent us a question on the subject matter in the liaison statement.  
The meeting felt that the content is almost reiteration of our previous 
question to SGXVIII.  Since there is some time to make a reply (SGXV will 
meet next in January 1993), Mr. Morrison undertook to draft a reply and 
circulate before the next meeting of the Experts Group. 

7.4 LAN (AVC-270)

AVC-270 provided some information on the difference between B-ISDN and 
LANs.  Differing views were expressed whether video signal transport is 
feasible on existing LANs.  Study should be continued on the ground that 
H.26X codecs may be used either on ATM based LANs or on LANs connected to 
B-ISDN as NT2.


8.	Multimedia multiplex (AVC-268; AVC-297)

AVC-268 provided discussion on the possible interface between the media 
control layer and the video codec and raised questions on negotiation 
methods.  

The meeting considered how we can make progress on the multimedia multiplex 
study.  Mr. Tanaka commented that we should study several alternatives in 
parallel until network conditions (cell loss, cost of separate VCs, 
differential delay between separate VCs, B-N interworking) become clearer.  
It is a common understanding of the meeting that a particular solution is 
not wise at this stage.

Chairman raised a study item on the multimedia multiplex; commonality with 
MPEG-2 system which is envisaged as extension of MPEG-1 system.  Analysis of 
the problem is awaited.


9.	Hardware verification of H.26X (AVC-283)

AVC-283 was input as a stimulus to plan hardware trials which is required at 
the later stage of the Experts Group activities.  First of all, at least two 
volunteers are solicited who independently develop hardware for the testing 
purpose.  The following comments were obtained;

    -	Definition of the digital interface between Terminal Adaptor and the 
codec part enables testing in various environments.
    -	The work plan to test at the end of 1993 is not practical.  The 
hardware may be much more difficult to develop compared to the H.261 
Flexible Hardware due to the higher speed and complexity.  Developing 
time of six months is not realistic.
    -	To transport a simulated coded bitstream through the network and to 
decode it at the remote end may be a method of verification.
    -	VBR can be tested at the same time, or CBR is tested first and VBR at 
a later date?
    -	Coordination with MPEG on the hardware verification is suggested.

Members are requested to study various aspects of the testing process to 
materialize the action plan.

Related to this discussion, a view was also expressed that the current plan 
of "freezing specifications" in March 1993 is difficult to achieve if the 
amount of work to do is considered.  Reactions to this view are awaited.  At 
this moment, we should take our efforts to make the "frozen specifications" 
appropriate for our communication systems. 


10.	Very low bit rate video coding

10.1 Background (AVC-263)

At its May 1993 meeting, Working Party XV/1 asked Chairman of the Experts 
Group to make an initial consideration in his group on the very low bit rate 
coding for PSTN and mobile/cordless networks.  The meeting had one hour 
free discussion after having reviewed relevant documents and tape 
demonstrations.

10.2 Experimental results (AVC-289,303,305,309)

Several simulation and experimental results were presented on video coding 
around 8-16 kbit/s.  The following comments were noted in particular;

    -	Modem initialization took very long time (e.g. 20 second) at the call 
set up (AVC-289), though it might be a problem of the particular 
equipment.
    -	Significant improvements have been found by using half-pel accuracy 
motion compensation instead of integer motion compensation plus loop 
filter employed in H.261 (AVC-303).
    -	Improvements by modifying VLCs and video multiplex are rather minor 
(5%), the used picture format is 1/9 CIF, the monitor size must be less 
than 3-4 inches, Susie is a stress material in these low bit rates (AVC-
305). 

10.3 Policy (AVC-290,304)

The following is a summary of views expressed during the free discussion.  
Chairman of WPXV/1 will be consulted with for suggested further actions of 
the Experts Group to prepare for a reply to the November meeting of 
WPXV/1.  

1) Video coding possibility

Algorithms of presented simulation or hardware results were in the category 
of H.261 +/- delta, providing more or less similar performance.  There was no 
indication for possibility of quite new algorithms at this meeting.

2) System and terminal

PSTN videophone standardization should cover all of the following aspects 
for the global terminal;

    -	video source coding
    -	audio source coding  
    -	multimedia multiplex
    -	channel coding
    -	modem, including fall back capability
    -	call set up, delay

PSTN videophone system may be differentiated with cordless system which is 
used as a leg to the ISDN system. 

It was noted that V.FAST modem (28.8 kbit/s) is being studied and 
requirements from conversational services should be reflected.

3) Interworking between PSTN and ISDN videophones

When a PSTN videophone interworks with an ISDN videophone, the ISDN side 
should prepare for interworking capability.  Since existing ISDN videophones 
do not have this capability, we should be careful of making retrospectively 
mandatory  specifications.  Existing ISDN videophones are required to be 
able to interwork PSTN telephones, thus one way to cope with the 
interworking requirement may be to deal with it in the context of ISDN 
videophone and PSTN telephone interworking.

This consideration should be a part of the general PSTN/N-ISDN/B-ISDN 
interworking issue.

4) Standardization program

We discussed necessary standardization in terms of "short term" and "long 
term".  The former means a very quick work and the latter may require 5 or 
more years time.  There was no claim for "middle term" standardization work 
which lasts 2 or 3 years.  Reasoning is;

    -	N-ISDN will prevail in this time frame.
    -	Current proprietary equipment may prevail.

It was a common understanding that longer term standardization of low bit 
rate video coding is required.  Very low bit rate coding significantly better 
than H.261, which may or may not be model based coding, is applicable to 
mobile applications and it will also provide very good pictures for the ISDN 
videophones.

There were expressed opposing views for the short term standardization.  
Supporting views are;

    -	De facto situation should be avoided.
    -	Technical solutions are feasible; video coding close to H.261 but 
optimized at low bit rates which can provide reasonably good pictures 
with reasonable complexity, 5" LCD  without extra cost, around 6 
kbit/s audio coding, V.FAST modem, etc.

Opposing or reluctant views are;

    -	The essential factor in PSTN videophone is rather pricing than 
technical specifications, price related technology is not appropriate 
for standardization.  If short term standardization is necessary, CCITT 
should adopt one of the existing ones, possibly through licensing (?).  
Defining the third one does not help the industry.
    -	It is questionable that the current performance be accepted by 
customers.  Answers may be available in six months or one year.

5) Harmonization with other standardization bodies

It is a unanimous view that CCITT should carry out the global system 
standardization as mentioned 2) above.


11. 	Preparation for the joint sessions

11.1 Documents

TD-5,6,7 are are forwarded as proposal from the Experts Group.  The 
following documents are forwarded for information or discussion with each 
original source; AVC-275,276,278,279,296,297,298,300,307,308,315. 

11.2 Representatives

The following is appointed as representative to the joint session;

	Requirements/Test	S. Okubo
	Video				G. Bjoentegaard
	System			B.G. Haskell
	Implementation		D.G. Morrison


12.	Others

12.1 IVS Technical Session (AVC-311)

1) Speakers

Messrs S. Okubo, D.G. Morrison and A. Tabatabai will speak on behalf of the 
SGXV Experts Group.

2) Participants

Those who intend to attend are requested to advise Chairman by August 10.

12.2 Update of the "status report" AVC-109

1) Time schedule

The meeting agreed to complete revision by the end of September so that we 
can distribute it at the IVS Technical Session.  Cooperation of the editors 
are requested.

2) Editors

As listed in AVC-109.

12.3 Future meeting plan

    -	8th meeting: September 28 - October 1, 1992 in Tarrytown, USA.

	Sole sessions: September 28 (half day)
	Joint sessions: September 29 - October 1

    -	9th meeting: October 28 - 30, 1992 in UK

	Sole sessions: October 28 - 30 in Ipswich
	Joint sessions: November 2 - 6 in London

    -	10th meeting: January 1993?

    -	11th meeting: March ?? - April ?, 1993 in Australia 

	Sole sessions: after the joint sessions?
	Joint sessions: March 29 - April 2

Note: IVS Technical Session is planned on October 26-27, 1992 in Ipswich.


END

*   *   *

Annexes

    Annex 1	Documentation
    Annex 2	List of tape demonstration
    Annex 3	Generic predictor in AVC-301
    Annex 4	Report form the small group on low delay mode
    Annex 5	Report of the small group on H.261 compatibility
    Annex 6	Cell loss resilience: Issues and core experiments
Participants of the seventh meeting of
Experts Group for ATM Video Coding 
(1-10 July 1992, New Jersey and Rio de Janeiro)


										N R 

FRG		Mr. M. Kuehn		DBP Telekom			  X
		Mr. F. May			Daimler-Benz			X X	CM
		Mr. G. Zedler		DBP Telekom			X X	CM

Australia	Mr. M. Biggar		Telecom Australia		X X	CM

Belgium	Mr. O. Poncin		RTT Belgacom			X X	CM
		Mr. B. Voeten		Bell Telephone		X

Korea		Ms. Sang-Mi Lee     	ETRI                	X X	(CM)

USA		Mr. B.G. Haskell		AT&T Bell Labs		X X
		Mr. S. Kumar 		Wiltel              	X
		Mr. A. Luthra     	Tektronix         		X X
		Mr. D. Klenke      	CLI               		X
		Mr. D. Hein      		VideoTelecom      		X
		Mr. N. Randall		DIS				X 	(CM)
		Mr. D. Raychaudhuri	David Sarnoff		X
		Ms. A. Reibman		AT&T Bell Labs		X
		Mr. R. Schaphorst		DIS				  X	CM
		Mr. A. Tabatabai		Bellcore			X	CM
		Mr. F. Tobagi		Starlight Networks		X 
		Mr. J. Zdepski		David Sarnoff		  X
		Mr. Y. Xiancheng  	PictureTel          	X

France 	Mr. G. Eude			CNET				X X
		Mr. J. Guichard		CNET				X 	CM

Italy		Ms. L. Conte		CSELT				X X	CM

Japan  	Mr. S. Okubo		NTT				X X	Chairman
		Mr. T. Murakami		Mitsubishi			X X
		Mr. K. Sakai		Fujitsu 			X X
		Mr. Y. Takishima		KDD				X 	(CM)
		Mr. T. Tanaka		NTT				X 	CM
		Mr. H. Ueno         	Toshiba			  X
		Mr. T  Yukitake		Matsushita Communication	X X

Norway		Mr. G. Bjoentegaard 	NTA				X X	(CM)

Netherlands	Mr. D.A. Schinkel		PTT Research   		X	CM
		Mr. A. Koster		PTT Research			  X

UK      	Mr. I. Parke		BT				X X
		Mr. D.G. Morrison		BT				X X	CM

Sweden		Ms. C. Verreth		Telia Research		X X

							  CM: Coordinating Member
							(CM): Substitute for CM