&%PAGE&
- &%PAGE& -
Document AVC-177R
CCITT SGXV Document AVC-177R
Working Party XV/1 November 29, 1991
Experts Group for ATM Video Coding
SOURCE : CHAIRMAN
TITLE : REPORT OF THE FOURTH MEETING OF THE EXPERTS GROUP FOR
ATM VIDEO CODING IN YOKOSUKA (November 18-29, 1991)
Purpose: Report
-----------------
1. General
2. Documentation
3. Tape demonstration
4. Review of joint sessions with MPEG
5. Discussion
6. Harmonization
7. Work plan
8. Others
1. General
The fourth meeting of the Experts Group consisted of two parts; joint
sessions with ISO/IEC JTC1/SC2/WG11 (MPEG) and CCITT sole sessions. The list
of participants appears at the end of this report.
The first part was held at JVC R&D Center in Kurihama, Japan, during 18-26
November 1991 at the kind invitation of JVC. At the start of the meeting,
Mr. H. Aoike, Senior Managing Director, made a welcoming address. At the
final plenary of MPEG, Chairman thanked the hosting organization for the
meeting facilities provided and the excellent secretarial support.
The second part was held at NTT Yokosuka R&D Center in Yokosuka, Japan, at
the kind invitation of Ministry of Posts and Telecommunications. At the
starting session, Dr. T. Kamae, Director of NTT Human Interface Laboratories,
made a welcoming address on behalf of the the hosting organizations. At
the closing session, Mr. D.G. Morrison, UK, thanked the hosts for the meeting
facilities and excellent services provided on behalf of the participants.
The Experts Group accepted resignation of Mr. Milton Anderson, Bellcore, as
liaison representative between WPXV/1 and MPEG, and thanked his outstanding
efforts (see Section 3.1/TD-7).
2. Documentation (TD2)
For this meeting, 71 AVC-numbered documents and 11 temporary documents
were available as listed in Annex 1.
3. Tape demonstration (TD3)
Several video tape demonstrations were given with D1 as detailed in Annex 2
to present experimental results.
4. Review of the joint sessions with MPEG
4.1 Technical sessions (TD-4,5,6; AVC-176)
Appointed representatives of the Experts Group made the following summary
reports on the joint sessions with MPEG that were held during 18-26
November 1991 at JVC-Kurihama. At the end of this report, Chairman thanked
the proposal submission by the Experts Group members and stressed that one
of the most effective ways to make influence on other groups is through
good technical work.
4.1.1 Test (S. Okubo)
Subjective tests for the thirty algorithm proposals were carried out
according to the double-stimulus continuous quality-scale method defined in
CCIR Rec. 500-3. Three test sequences were used for the 4 Mbit/s test and
four sequences for the 9 Mbit/s test. According to the average scores, rank
orders for 525 and 625 systems were obtained. Furthermore, the top ranking
group whose members are mutually indistinguishable in a statistical sense
was identified by using the Duncan's method. Draft report on the test
results are contained in TD-5.
During the discussion, it was raised whether the test results meet the
objective of "NTSC/PAL level quality at 3-5 Mbit/s" and "close to CCIR
Rec.601 quality level at 8-10 Mbit/s." It was felt necessary to know the
relationship between MOS (Mean Opinion Score) and the testing scales
following CCIR Recommendation 500-3.
It was also felt necessary to know how NTSC/PAL picture quality maps onto
the testing scale. Mr. Tabatabai presented AVC-176 that contains relevant
information showing that pictures coded at 4 Mbit/s are better than NTSC
pictures on average. It was pointed out that the average score values
differ in the testing scale between TD-5 and AVC-176 even for the same coded
picture.
The meeting recognized the desirability to test the coding algorithm for
source signals that are converted from NTSC/PAL/SECAM signals to 4:2:2
digital component signals, because they may include particular artifacts.
There may be some practical difficulties, however, to convert a CCIR-601
digital signal into an analog NTSC/PAL/SECAM signal and then record the
NTSC/PAL/SECAM signal in a D-1 tape.
4.1.2 Requirements (S. Okubo)
The Requirements sub-group assisted the Video sub-group in setting the
guidelines toward defining Test Model 0 (TM0) as summarized in TD-6. The
goal of developing a generic video coding standard which meets the
requirements of various applications was confirmed. Use of TM as a vehicle
to reach this objective was also confirmed. General feeling was that we
could start with MPEG-1 syntax using simple field merging.
4.1.3 System (B.G. Haskell)
Normative part of CD 11172-1 (MPEG-1 System) was completed at Kurihama. It
contains specifications for multimedia syntax, which provides CBR and VBR
operations, buffering for each medium as well as multiplexed signals and
timing for audio/video cross media synchronization.
4.1.4 Video (G. Bjoentegaard)
After each proposer had given 10 minute introduction to his/her coding
scheme, three groups reviewed the proposals in details, analyzing coding
methods, characteristics and functionalities. As an outcome of this work,
the summary table was produced as in TD-4.
Considering these analyses and test results, a joint Video/Requirement
meeting discussed how to start the convergence work. Guidelines and
specific action plans toward the next meeting in Singapore were obtained as
in TD-6.
4.1.5 Implementation (D.G. Morrison)
A work to rank the proposals was initiated, but there were some significant
differences between assessor's opinions. This difference will be sorted out
by the first day of the Singapore meeting through correspondence.
Implementation sub-group recommended that coding performance be evaluated
whether backward prediction is necessary or not, since it gives a great
impact on the coding scheme implementation.
4.2 DSM/ATM workshop (AVC-172; AVC-114,131,140)
This workshop was held in the afternoon of November 20. Two speakers
invited by the DSM group presented storage characteristics of the optical
disk and the label format. Then the following four speakers from the
Experts Group presented ATM network characteristics and their impacts on
video coding;
M. Biggar Broadband ISDN evolution
O. Poncin ATM network characteristics
T. Tanaka VBR aspects
G. Morrison Video coding techniques suitable for utilizing ATM network
characteristics
This workshop was found effective in obtaining mutual understanding among
experts participating in the MPEG/CCITT EG joint work. The meeting thanked
these four speakers for informative presentation.
5. Discussion
5.1 Report of the meetings relevant to the Experts Group
5.1.1 CMTT/2 SRG meetings in Torino and Paris (AVC-173,174)
Mr. Carbiere presented the two documents, focusing on items of our concern.
Mr. Sawada provided additional information in response to the questions.
Through the questions and answers, the following were clarified;
- Exact number of lines per field for the 525/60 TV system is discussed
as in the note of p.2/AVC-174.
- The two different HDTV frame rates, 60 Hz and 59.94 Hz, may require
two oscillators in the codec, because frame rate conversion is not
straightforward.
- The pyramidal coding between HDTV and EDTV/CTV formats is still an
example without consensus.
- A 20-30 year long lifetime of consumer equipment implies that the
coming video coding standard should not become obsolete in e.g. five
years time frame, and it should leave room for the encoder to make
performance improvements by maintaining the syntax, thus the existing
receivers can enjoy the improved coding performance.
5.1.2 IVS Co-ordination meeting in Tokyo (AVC-110)
Chairman reported that the objective of this meeting included two aspects;
co-ordination between network and video coding standards and co-ordination
among video coding standards, and that several action points were identified
for the co-ordination purpose. The meeting paid attention to the time
tables of network, video coding and service standards development.
Mr. Carbiere clarified that CMTT/2 assumes and considers as a mandatory
constraint that the same bit-rate reduction algorithm may be used for
secondary distribution in both STM and ATM networks, and in the latter case
for both CBR and VBR modes. This information was sent to relevant groups.
5.1.3 SGXV WPXV/1 meeting in Geneva (AVC-138; TD-7)
Mr. Makoto Yamashita, Vice-Chairman of Study Group XV and Chairman of
Working Party XV/1, presented the outcome of the WPXV/1 meeting held in
Geneva during November 11-15, appreciating the activities of the Experts
Group. He also provided information on the coming meetings of WPXV/1 (May
1992, November 1992, July-September 1993), stressing that Recommendations
can now be produced virtually at any time using Resolution No. 2. He also
touched upon the expected restructuring of ITU leading to integration of
CCITT and part of CCIR into the Standardization Sector.
5.2 Picture format (Annex 3/AVC-106R)
Issues concerning the picture format selection for Recommendation H.26X are
summarized in Annex 3 to AVC-106R. The meeting reviewed several documents
addressing this topic.
5.2.1 Number of formats (AVC-125,134,169,174)
AVC-125
Experimental results were provided to conclude that geometric distortion of
6% is clearly perceptible for the Miss America picture. It was pointed out
that circles are more demanding in geometric distortion. Related to whether
the pel aspect ratio is exactly 16:15 for CIF, it was questioned whether 720
pels per line in CCIR-601 correspond to exactly "4" of 4:3 aspect ratio.
Clarification is needed.
AVC-134
This document discussed several issues regarding the use of a flexible set
of formats, which had been identified in the previous Santa Clara meeting.
During the discussion, the following comments were obtained;
- Since aspect ratio is flexible inside the maximum sized container,
square pixel can be covered in this flexible set of formats using an
indication flag.
- Many filter banks may be required to convert the decoded pictures
into the display format, but display methods are not subject to
standardization.
- CCITT should be concerned with end-to-end services. Leaving the
format conversion process outside the coding standard is not
sufficient.
AVC-174
As part of CMTT Special Rapporteur's Group activities, picture format
related matters were taken up for further review. The following
clarification was made;
- In the secondary distribution, signals of different television
standards will be converted into local television signals in front of
the encoder, thus the decoder needs to handle only regional formats.
- In Figure 1/AVC-174 the common coding format for CTV and EDTV may be
either 2:1 interlaced or progressive, because it has not yet been
standardized whether EDTV format be either 2:1 interlaced or
progressive.
5.2.2 Loss of coding efficiency due to conversion between CCIR-601 and
SCIF (AVC-113,174)
AVC-113
This document reported that no visible degradations due to 480 to 576 to 480
line conversion had been observed. The presentation was supported by a
tape demonstration. Coding loss due to the use of 576 line progressive
format with 30 Hz frame rate was found as less than 3% against 20% increase
in number of pels to be coded.
AVC-126
This document reported that coding efficiency improvement can not be
expected by using optimized 2-d VLC for INTER mode. Mr. Bjoentegaard
pointed out that his experiments indicated that INTRA had not improved but
INTER had improved, and that NTA's coding algorithm proposal uses
coefficient by coefficient adaptation for DCT coefficient coding, leading to
5-10% gain.
AVC-169
This document compared coding efficiency between a SCIF with 576x720 spatial
resolution and a SCIF with square pixel geometry (e.g. 528x704) for the
525/60 signals. The following comment and clarification were obtained
during the discussion;
- As a rule of thumb, 10% coding efficiency difference corresponds to
0.5 dB of SNR. SNR difference of 0.28 dB may not be negligible
compared to loss of coding efficiency by 5.8 %.
- Conversion filter is integer based.
5.2.3 Conclusion
a. We will try to make a decision at the March 1992 meeting as agreed at
the Santa Clara meeting.
b. We need to clarify the color resolution, namely 4:2:2 vs 4:2:0
considering the following;
- picture quality
- use of a 4:2:0 family in existing standards (H.261, MPEG-1)
- interlaced or progressive format as the source coder input
- computer applications requiring 4:4:4
c. Testing environments are not yet ready for progressive format
experiments. A practical problem is that simulation systems including
progressive scan displays are not yet widely available.
d. An idea was raised that side information indicating the origin of the
SCIF signals may be utilized to facilitate selection of coding parameters (or
alternatively coding modes). It may not be used by the decoder if
unnecessary. Concept of the common intermediate format lies, however, in
that encoder or decoder need not worry about the origin of the signals.
e. The following issues are still open;
- single coding format or multiple ones for higher quality video coding
- possible parameter values? how related to CIF? square pixel?
- extension to larger formats such as EDTV and HDTV?
- service integration aspects
- take the initiative to adopt the single format approach (if viable) to
other groups?
- sufficient picture quality is achievable?
- 50/60 Hz frame rate conversion is practical?
- not so much coding loss due to conversion?
5.3 Framework for H.26X video coding
5.3.1 Requirements
1) Current status (AVC-108,173)
Various requirements coming from communication, distribution and storage &
retrieval applications have been identified in AVC-108, Proposal Package
Description.
CMTT/2 provided an updated version of the paper on preliminary functional
requirements for the secondary distribution (see AVC-173) where a "channel
hopping" requirement is included as "a limited response time between
selecting a channel and having the service fully operating, a range of 0.3
to 1 second is being considered."
2) ATM coding specific requirements (AVC-115)
Compatibility, low end-to-end delay and cell loss resilience are the items of
our greatest concern from CCITT point of view. Lower bit rate operation of
the H.26X codec is also of our concern, though the overlapping with H.261
bit rates needs careful consideration. These should be reflected in the
definition of Test Model 0 (see Annex 4).
5.3.2 Architecture
1) Cell loss resilient coding (AVC-120,139)
AVC-120
This document provided a collection of techniques for cell loss resilience,
identifying matters of standardization. The small group for video coding
algorithm study encouraged the contributor to make it more illustrative and
extensive for submission to the Singapore joint sessions.
AVC-139
This scheme consists of one-layer source coding and one-layer or two-layer
transmission coding, which was submitted to FCC for testing of advanced
digital television systems. Cell loss resilience performance was
demonstrated with D1 tape. During the discussion, the following were
clarified;
- The digital television broadcasting medium requires two separate
channels having different transmission error performance for better
radio spectrum usage.
- Payload capacity of 128 bytes was selected in consideration of
overhead penalty and hardware complexity. It can not be 48 bytes due
to overhead penalty.
- Cell loss was made by using a free running random generator.
2) Compatibility with existing and companion standards (AVC-122,123,135)
AVC-122
This document reported experimental results on the comparison of
compatible mode and non-compatible mode for the algorithm proposal #13
(AVC-150). During the discussion, the following were clarified and
commented;
- Difference of 1.2 dB in SNR is visible. Optimization has not yet been
carried out.
- In this configuration, Encoder 2 deals with only differential signals
between input and upconverted ones. However, Encoders 1,2,3 should
better be identical from implementation simplicity point of view,
awaiting further study.
- Horizontal upconversion filter is the one used in SM3.
- There was a comment that the simulcast method had given better
pictures than the pyramidal method. At the same time, there was
another view that the simulcast could be as good as the pyramidal
scheme, but never be better.
AV-123
This document discussed the H.26X compatibilities with existing and future
companion standards in terms of "bit stream level compatibility" and "source
coding level compatibility." It was clarified that H.32X should have
interworking capability in H.221, H.242 as well, but there was expressed an
expectation that network interworking facilities may ease the problem.
AVC-135
This document reported the progress of the ongoing work regarding the
flexible layering approach for backward compatible coding where layering is
based on the DCT coefficient domain decomposition. There was some
discussion whether this decomposition causes inherent artifacts, including a
comment that the "incremental low frequency information" will reduce those
artifacts in a full size picture.
3) Delay analysis (AVC-124; TD-9)
Major delay causing elements were identified in the hybrid coding scheme.
Considering the variety of exactness in presenting the delay values for
proposed algorithms, we concluded it necessary to clearly define
coding/decoding delay including illustration of field-based, frame-based and
backward prediction codings, and put it forward to the Singapore meeting so
that everybody follows a presentation format. Ms. Verreth undertook to
coordinate a small group to draft this contribution. The outcome was given
as in Annex 3 (reproduction of TD-9).
As to the list of delay causing elements, the meeting concluded it better to
make a precise illustration for the common understanding among members
participating in the joint work with MPEG, and to list major inherent
elements for a common data presentation format. Having seen that some
coding algorithm proposals are of multi-pass for better coding rate control,
the meeting also agreed to include a "pre-analysis" box to stress that it is
delay demanding. Representation of delay time is recommended to be both in
millisecond and frame time, such as 20 ms (0.5 frame time).
These will be reflected in a contribution of the Experts Group to the
Singapore meeting.
4) Difference between H.261 and MPEG-1
The meeting recognized that this reality should be considered when we
discuss the compatibility issue. Some preliminary thoughts were expressed
in the course of discussion;
- Some chips include both H.261 and MPEG-1 decoders, hence they can be
used as core in as switchable manner. It will be convenient if picture
formats of H.26X/MPEG-2 have simple relationship with those of
H.261/MPEG-1. Mr. Parke clarified that the pyramidal scheme allows the
use of any format in the core as far as spatial and temporal
upconversion methods are standardized.
- Bitstream should contain standards indicator? MPEG-1 system includes
private bitstream indication. These may suggest consideration of
"open system" in the audiovisual coding area?
- Negotiation can not be used in distribution oriented applications,
requiring a mechanism to indicate the coding algorithm.
- Compatibility and commonality should be distinguished for clarity.
5.4 VBR vs CBR
5.4.1 Characteristics
AVC-116
This document provided buffer size and delay estimation for sliding window
restriction in the network, indicating use of transmission buffer is cheaper
than the use of display buffer. Mr. Haskell stated that AT&T is in favor of
leaky bucket as UPC mechanism because of its simplicity.
AVC-127
This document presented some experimental results on the leaky bucket
occupancy. It was clarified that "when coding rate is rather high" should
read "when coding rate is rather low."
5.4.2 Statistics (AVC-171)
Analysis results were presented for the data that Mr. Verbiest distributed
at the previous meeting in Santa Clara. It was found that the number of
cells per frame follows Gamma distribution even if the coding schemes are
quite different (DCT vs DPCM).
Data on MC based systems are solicited.
5.4.3 Network model (AVC-131)
It should be noted that the CLR equation in Annex 5 to AVC-109 (Status
Report Issue 2, p.50) includes a mistyping and that the corresponding
equation in AVC-138, p.13, has been corrected.
5.4.4 UPC (AVC-128,132; AVC-116,127)
AVC-128 investigated the transmission efficiency between sliding window and
leaky bucket as UPC mechanism, while AVC-132 raised that this Experts Group
should recommend a UPC mechanism appropriate for our video coding purpose.
In the course of discussion, the following comments were obtained;
- Actual video signals are stochastic process, thus we should be careful
of interpreting the results of AVC-128.
- When we request a particular method to the network, additional
network cost should also be taken into account.
- Many of the participants feel in favor of the leaky bucket as UPC, but
some others can not be sure which of leaky bucket or sliding window
is better. Reasons for the support of leaky bucket are transmission
efficiency and ease of coding control in both CBR and VBR.
- Leaky bucket is efficient for video, but may be questionable for other
type of signals.
- It is beneficial for customers if simple designation of "class of
picture quality" is sufficient without specifying many parameters.
Leaky bucket size is better to be defined with one parameter; time
length. If a bit rate is given then the buffer size is determined by
the network. However, it may cause more congestion at higher bit
rates.
- UPC should be the same all through the international connection,
otherwise the terminal can not control its information generation
rate.
- These should be reflected in our liaison statement to SGXVIII.
5.4.5 Open issues
The following issues are still open, awaiting contributions;
- VBR advantage in obtainable picture quality
- Appropriate UPC for video (audiovisual) coding
- Required window or bucket size for average rate monitoring in the
network
5.5 Video coding algorithms
5.5.1 Layered coding (AVC-117,119; AVC-115,135)
Two implementation examples and their characteristics were presented in
AVC-117 and AVC-119, both of which use layering for cell loss resilience.
It was clarified that the scheme in AVC-117 assumes a constant rate, cell-
loss free base layer and a peak rate restricted enhanced layer. It was also
stated that the low priority service is required to guarantee the high
priority service. The meeting was reminded that the tariff structure would
eventually decide the video coding structure, particularly use of the two
channels with different QoS.
The meeting took note of the AVC-119 technique to use two frame memories in
the coding loop for further investigation of cell loss resilience.
5.5.2 New schemes (AVC-118,133,141 through 167; AVC-119; TD-4,5,6)
There were 30 proposals submitted to the Kurihama tests, all of which use
motion compensated interframe or interfield prediction for temporal
redundancy reduction. For spatial redundancy reduction, some of them
adopt DCT, while others adopt sub-band filters or Wavelet Transform.
5.5.3 Conclusion (TD-9)
Having reviewed the work in Kurihama (see Section 4.1 above) and submitted
documents for the sole sessions, the meeting concluded that we can
concentrate on the joint work with MPEG for the H.26X algorithm development
which will be carried out through defining and improving the Test Model.
Though this work may be focused on 5-10 Mbit/s for the time being, we should
bear in mind that our terms of reference cover also extension to the lower
bit rates and to the higher formats.
Regarding the low bit rate operation of H.26X, the meeting confirmed that
H.26X should cover a range of applications including support of lower
quality services and interworking between H.26X and H.261. It was also
pointed out that in layered schemes the choice of base layer bit rate may
affect the total performance, and that 384 kbit/s is a typical bit rate for
videoconferencing services. It was further pointed out that at lower rates
streamlining overhead information should be taken into account.
Based on this conclusion, Mr. Guichard led a small group for establishing
guiding principles for the Singapore meeting, covering low delay,
compatibility with existing standards, scalability or low bit rate operation,
and cell loss resilience. The outcome was presented with TD-9. The meeting
supported this proposal with an editorial clarification, recognizing that the
"core experiments" should be specific enough so that independent
laboratories will obtain the same results to prove or disprove an issue
item. The finalized document is contained in Annex 4.
As to the cell loss experiment specification, the meeting requested Mr.
Parke to organize a correspondence group for making a written proposal to
the Singapore meeting. The group consists of Messrs Brusewitz, Eude, Parke
(Chair), Sakai, Schinkel, Tabatabai, Yukitake, Wada, Zdepski.
There was some discussion on the availability of proper test sequences to
check lower bit rate operation of Test Model. "Susie" was suggested for the
moment, but other sequences may be needed in the future.
5.6 Multimedia multiplexing (AVC-129,136)
5.6.1 Comparison of possible methods
Documents AVC-129 and AVC-136 presented comparison of the following three
methods of multimedia multiplexing for the Experts Group discussion;
- cell based approach
- packet based approach
- H.221 based approach
It was confirmed that the first approach had been chosen as reference to
help the progress of our work, taking into account of the B-ISDN long term
objectives. A view was expressed, however, that the multiplex method should
be chosen independent of the network characteristics because audiovisual
services may be supported by not only B-ISDN but also by other networks.
Additionally, it was pointed out whether the H.221 based approach is
appropriate for VBR services as well.
Mr. Zdepski clarified that his multimedia multiplex scheme in AVC-139
corresponds to Figure 2/AVC-129 (packet based approach).
As a conclusion of the discussion, Chairman requested of the members to
propose a methodology to quantify the comparison and to further study this
multimedia multiplexing matter.
5.6.2 Interworking between N-ISDN and B-ISDN terminals (AVC-136)
This document provided several possible scenarios of B-ISDN and N-ISDN
interworking, focusing on multimedia multiplexing methods. Though support
of the in-channel protocol conversion by Network Adaptor (NA) is assumed in
some scenarios, there was expressed a strong view that it is unlikely in
some countries. It was also pointed out that it may not be so difficult for
H.32X terminals to have two multimedia multiplex methods if they are
implemented with software.
5.6.3 "Media Control Layer" functions (AVC-129)
Mr. Hibi clarified that his idea of virtual media control layer is to maintain
the same logical interfaces between respective media sources and this media
control layer, thus to allow different multimedia multiplex methods without
need to change the media source configuration.
5.7 AAL (AVC-112)
This document provided recent work on AAL Type 1 by Rapporteurs group of
SGXVIII, which will be elaborated at the SGXVIII Melbourne meeting.
Since the H.32X terminal is assumed to use AAL Type 2 and the H.320 terminal
is assumed to use AAL Type 1, it was pointed out that we need clear
understanding on the use of AAL Type 1 and Type 2 for VBR and CBR support.
It was said that Type 1 should always fill the cell, thus can support
continuous bits, while Type 2 supports message start/stop, thus half filled
cells are allowed. This aspect should also be clarified.
5.8 Network aspects
5.8.1 Timing synchronization (AVC-121,130)
This document raised a question whether H.26X should provide some means to
synchronize encoder and decoder video sampling clocks. It was clarified
that CCIR Rec. 723 includes this functionality.
After some discussion, the meeting reached a common view to include
necessary information in the coded video bitstream so that the decoder can
display the encoded pictures without frame dropping or frame repeating
being required. This information can be in picture layer, GOB (slice) layer
or even macroblock layer.
To decide a specific method to achieve the video clock synchronization,
more information is required about what support of the network can be
obtained. In particular, if we seek commonality between AAL Type 1 and Type
2, we must accelerate our study.
This timing recovery issue was included in the liaison statement to SGXVIII
(see Annex 5).
5.8.2 CLP (AVC-170)
This document raised some questions concerning ATM cell header fields to
the Experts Group's attention; CLP availability to the receiving end user,
tagging of compliant cells, inter-dependence between CLP and user
information, QoS values.
These were included in the liaison statement to SGXVIII (see Annex 5).
5.8.3 Liaison to SGXVIII (AVC-130; AVC-136,139; TD-10)
The meeting decided to send a liaison statement on the network related
requests and questions proposed in AVC-130 and other documents to SGXVIII
which meets during 2-13 December 1991 in Melbourne. Mr. Biggar undertook
to coordinate a small group, which made a draft as in TD-10. The meeting
approved this draft and sent the content in Annex 5 as addendum to the
previous liaison statement produced at the Santa Clara meeting.
6. Harmonization
6.1 CCIR SG11 (AVC-111)
Chairman requested the members to pay attention to this new player in video
coding area.
6.2 CMTT/2 (AVC-174)
After having briefly reviewed our previous actions and expected interaction
at the Singapore meeting, the meeting concluded to see the progress in
Singapore without sending new liaison statements at this time.
6.3 SGXVIII (AVC-110)
Chairman requested cooperation of the members at the IVS workshop that
would take place next autumn, perhaps in Europe.
The meeting endorsed that Mr. Biggar kindly act as Representative of this
Experts Group at the SGXVIII Melbourne meeting.
6.4 MPEG (TD-6)
We will seek the success of the joint work as concluded in Section 5.5.4
above.
6.5 ISO/IEC JTC1/SC29 (AVC-175)
Dr. Hiroshi Yasuda, Chairman of newly founded SC29, presented the outcome
of its first meeting held in Tokyo during 21-23 November. The meeting took
note of the fact that SC29 is seeking formal relationship with CCITT Study
Group XV for collaborative interchange.
7. Work plan (AVC-137)
This document suggested a staged approach of the video coding standard
development in correspondence with the B-ISDN standard development.
The Experts Group will consider this suggestion to elaborate its future
work plan. One area needing consideration may be splitting and integration
of video coding aspects and network related aspects.
8. Others
8.1 Status report (AVC-109)
Updating by appointed editors will be carried out on a continuous basis.
The current version is a good summary of necessary work. Inclusion of more
answers are desirable in the future.
8.2 Future meetings
- 5th meeting: Sole session in the afternoon of 6 January 1992, joint
sessions with MPEG during 7-9 January in Singapore.
- 6th meeting: Sole sessions during 18-20 March 1992 in Stockholm and
joint sessions with MPEG during March 23-27 in Haifa.
END
*************
Annexes to AVC-177R
Annex 1 Documents for the fourth meeting of the Experts Group
Annex 2 List of tape demonstrations
Annex 3 Definition of end-to-end delay
Annex 4 Guiding principles for the Singapore meeting
Annex 5 Addendum to liaison statement from SGXV Experts Group to
SGXVIII
Participants of the fourth meeting of
Experts Group for ATM Video Coding
(18-29 November 1991, Yokosuka)
FRG Mr. G. Zedler * DBP Telecom CM
Australia Mr. M. Biggar Telecom Australia CM
Mr. G. Smith AUSSAT
Belgium Mr. O. Poncin RTT Belgium (CM)
Canada Mr. R. O'Shaughnessey BNR (CM)
USA Mr. K. Dallas National Semiconductor
Mr. B.G. Haskell AT&T Bell Labs
Mr. K. O'Connell Motorola
Mr. R.P. Rao + Compression Labs
Mr. A.J. Tabatabai * Bellcore CM
Mr. J. Zdepski David Sarnoff
France Mr. G. Eude France Telecom
Mr. J. Guichard France Telecom CM
Mr. G. Nocture + L.E.P. (Philips)
Japan Mr. K. Hibi Sharp
Mr. K. Matsuda Fujitsu
Mr. K. Matsuzaki Mitsubishi Electric
Mr. S. Okubo NTT Chairman
Mr. K. Sawada * NTT
Mr. Y. Takishima KDD
Mr. M. Takizawa Hitachi
Mr. T. Tanaka NTT CM
Mr. H. Tanihara * Ministry of Posts and Telecom.
Mr. M. Tsujikado Oki
Mr. A. Tsuzuku * Ministry of Posts and Telecom.
Mr. H. Ueno Toshiba
Mr. M. Wada KDD CM
Mr. M. Yamashita * NTT
Mr. M. Yano NEC
Mr. H. Yasuda * NTT
Mr. T Yukitake Matsushita Communication
Norway Mr. G. Bjontegaard Norwegian Telecom (CM)
Mr. J. Bording + Norwegian Telecom
Netherlands Mr. H. Carbiere * PTT Research LR-CMTT
Mr. A. Koster PTT Research
Mr. D.A. Schinkel PTT Research CM
UK Mr. I. Parke BT
Mr. D.G. Morrison BT CM
Sweden Ms. C. Verreth Swedish Telecom
Mr. H. Brusewitz Swedish Telecom CM
Korea Mr. Y-H. Kim ETRI Observer
CM: Coordinating Member
(CM): Substitute for CM
LR: Liaison Representative
* Sole sessions (27-29 November) only
+ Joint sessions (18-26 November) only