The MPEG-J activity addresses a flexible system for programmatic control of an audio visual session to adapt to the operating conditions of the terminal. At the Dublin meeting, the existing architecture of MPEG-J was refined. Several companies (Sun, AT&T, Philips, Microsoft, CSELT, SIA, FINSIEL) contributed to this activity, in particular by providing Java API’s to this meeting. Through a process of convergence, a first set of API’s has been defined. Additional API’s proposed by Sony to allow multi-user world to be supported by MPEG-4 are under consideration and should be introduced in the current architecture.
Work continued on the development of the MP4 format, the stored content format for MPEG-4. Further progress was made in integrating the MP4 format with the other portions of the MPEG-4 specification, and development of reference software has begun. Content authors, providers, and other interested parties are encouraged to join the effort at defining this format specification scheduled for December 1998.
The proposed technical specification for IPMP (M 3537) was refined and introduced into the MPEG-4 Systems VM for Version 2. To facilitate this, a dedicated Systems Plenary was held on Tuesday morning. As a result of this meeting, the IPMP specification was refined and passed on to the VM/WD editor. To enable the IPMP specification to be included into the MPEG-4 standard Version 1, a Verification Test, described in the Im1 workplan needs to be conducted.
A competition for Systems logo has been proposed. The best logo will be chosen in Atlantic City upon the vote of the Systems people. The Systems chair promised to allocate a Magnum bottle of Champagne for the happy winner.
Study on FCD
The comments from National Bodies as well as personal input contributions related to the FCD have been collected and discussed during the meeting. The list of comments is available as Study of FCD (N2349).
Only the main issues were discussed. These discussion should continue on the email@example.com reflector (the firstname.lastname@example.org is discontinued). New issue should be brought up as early as possible.
An interim editors meeting is planned to integrated Study on FCD input document for Atlantic City meeting. This document will be the starting point for the creation of FDIS, in anticipation of the resolutions of NB comments.
The main issues concerning the study on Systems FCD were the following :
Handling of upstream control channels (a solution is provided with a control Stream type)
Stream control : play, pause, etc... (a solution is provided with the new BIFS Message node)
Broadcast carousel (a solution is provided in the Systems Sync-layer)
URLs (a solution is provided in the DMIF specification)
IPMP compatibility in Version 1 (a solution is provided in the study of FCD)
Global stream identification (a solution is provided in the study of FCD)
Association of Face and Audio nodes to a TTS stream (specification improved but further work needed to integrate it).
Cross checking of the items above and final drafting of the study of FCD needs now intensive work on the specification reflector : email@example.com.
The activity on conformance clearly insufficient. If the needed resources are not allocated and if the work does not make significant progress, there is a big risk that further MPEG-4 Version 2 activity will stop. Jean-Claude Dufourd has volunteered to be the main editor as well as the Systems editor of what will be MPEG-4 Part 4 “Conformance”.
The activity on video normative composition has been stopped because of lack of contributions.
Methodology for evolution from FCD to FDIS
We have produced in Dublin a Study on Systems Final CD. This document has no official value. The Systems sub-group will now :
Continue to work on the study of FCD in order to provide for clean normative text for the bugs that have been identified and fixed.
Take care that all the comments that we foresee are taken into account by at least one national body, that will report them at the second round of comment in Atlantic City.
In Atlantic City, Systems will produce the FDIS, and no more technical changes to the Version 1 specification (only editorial) will then be possible.
The following items have been treated in advanced BIFS :
The audio nodes have been enhanced.
The user interaction nodes have been enhanced.
The specification of missing VRML nodes has been enhanced.
The nodes related to multi-user world have been introduced.
The relation ship between the current version 2 specification on user interaction node and the new version BIFS node needs to be clarified.
A new node has been proposed to allow the coding of visual objects with existing algorithms like MPEG-1/2, H.261/3, GIF, JPEG, … The node has been introduced as core experiment in the Version 2 BIFS VM. The AHG on BIFS should evaluate with the Requirement group, the Video group and the ISG group if such a functionality should enter the MPEG-4 standard.
The contributions Java API’s have been reviewed by the group and a set of API has been chosen. During a discussion with IPMP, the conditional access API’s are withdrawn. There were no consensus on inclusion of Multi User Interaction (MUI) API’s in MPEG-J as proposed by Sony. A place holder has been created in MPEG-J and the issue shall be resolved within the MPEG-J AHG.
A list of companies have committed themselves for software validation and implementation. For validation, the presentation engine can be proprietary and only made available in the form of an executable. The software associated to the application engine has to be provided in source code according to WG11 rules on software donation. After validation, the API’s go to WD stage. For Part 5 version 2 implementation, a complete and unique software implementation of the presentation engine will be developed jointly by the companies based on Im1. This presentation engine as well as the application engine will be ready at CD stage and be donated according to WG 11 rules.
Finally, a contribution has been proposed to use ECMA Script for the application engine. It has been decided that the best technology will be chosen in accordance to the MPEG procedure of core experiment. In case both ECMA Script and the current MPEG-4 VM choice (Java), satisfy equally the MPEG-4 requirements in terms of performance and functionality, the VM technology will stay as is. A core experiment on ECMA Script for use in the MPEG-4 engine will be introduced in the current VM and published on the MPEG-J and Systems reflector within 10 days after the Dublin meeting.
MPEG-4 File Format
The MPEG-4 File Format Break-out Group met throughout the entire week. The work was chaired by Dave Pawson. All input documents were processed.
The goals of the Subgroup were to review comments since Tokyo, to clarify relationships between systems and audio/visual, to clean up unnecessary open issues and to assign “owners” to remaining open issues.
During the meeting, all AHG suggestions and submissions have been reviewed and the results have been incorporated into the VM. 21 open issues have been reduced to 10, several are already written up and ready for ad-hoc distribution and discussion.
In particular, it has been concluded that the sample table and constituent atoms satisfy our requirements, the carriage of FlexMux has been resolved by storing relevant tables in Sample Description Atom of hint tracks, the appearance of SL-Packets in stored content has been defined.
Results of this meeting have been included for version 2 systems VM 3.0. It has been verified that the file format properly supports audio/visual. An AHG will be created to continue VM development and track progress of open issues. Implementation will continue; QuickTime code will migrate towards “true” MP4 and the merging of this code with IM1 will be investigated.
The Intellectual Property Management & Protection (IPMP) Break-out Group met throughout the entire week. The work was chaired by Niels Rump. All input documents were processed.
The proposed technical specification for IPMP (M 3537) was refined and introduced into the MPEG-4 Systems VM for Version 2. To facilitate this, an exhaustive Systems Plenary was held on Tuesday morning. As a result of this meeting, the IPMP specification was refined and passed on to the WD editor.
To enable the IPMP specification to be included into the MPEG-4 standard Version 1, a Verification Test needs to be conducted. This test will use IM1. A meeting with IM1 was held to discuss issues related to this test. To conduct the test, a detailed test plan will be set up by July 24. It was agreed that the tests results should be available by September 15, 1998. To facilitate these tests, an ad-hoc group needs to be established. An ad-hoc meeting may take place in Sunnyvale, California, in mid-September.
A joint meeting of the IPMP Break-out and MPEG-J Break-out groups was held on Wednesday to discuss IPMP/MPEG-J issues. Further discussion between these groups is needed to reach full understanding of the issues.
The comments of the Japanese National Body on MPEG-4 IPMP (M3681) have been answered.
Methodology for evolution in Version 2
The methodology for evolution of the VM is described on the MPEG home page and follows the Core experiment process. Syntax and semantic specified in the VM will go in the WD only when validated (implemented, tested and exercised).