[Minutes Overview] [Workshop Home] [Previous:Security] [Next:Closing Session]
Please refer to the position-papers and slides for authoritative answers. The following minutes are only a snapshot of Presentations and Discussions
See also the [Slides (ppt)] and the Position Paper
Rob Koenen (MPEG/Intertrust) How do you define interoperability?
Answer:We are taking any element of a content and try to use it on any device I want at the moment
Answer:From a publishers view: We like to present content to one or more dr systems and get the same rules and ip asserted in the same way
Question:The descriptors, are they placed in the content or are they in the sound stream?
Answer:No, it listens to the music and trys to work out what artist it is. It looks it up in a database to find the most likely content. The MPEG group is working on metadata to insert in content
Question:How are the MPEG calls for proposals circulated?
Answer:There are different ways. The best way to keep being informed is to look regularly at the MPEG home page
DRM is hard because you are talking about the standardization of trust. MPEG has been dealing with this for 7 years. MPEG would like to develop a Rights Expression Language and a Rights Data Dictionary, and we have RFP out there to do this. Koenen then showed a slide and sketched out a list of the various MPEG standards activities: MPEG-1, -2, -4, -7, and -21.
About interoperability, there are two kinds:
The MPEG constituency is diverse: broadcast, mobile, streaming. We often see people from the same [big] company meeting in the MPEG context for the first time.
MPEG refers to DRM as Intellectual Property Management and Protection (IPMP)
The MPEG standards in general are moving from offering programming hooks to true interoperability definitions. MPEG-2 enables proprietary access systems with copyright descriptor identifier and number (international standard ids). The standard cannot do the enforcement, legislation can enforce. The standard at this time offers hooks to proprietary systems to do the protection. In MPEG-4 the current situation is similar: identification and hooks, at any level granularity (parts of program). Koenen showed a diagram of the MPEG-4 structures, hooks that get the stream to your proprietary IPMP system and then out again to play.
MPEG-7 is about describing content. The describing of content has value,
this is data that is protectable as well. We hate to call it metadata because
it includes descriptors of very low level data, like shapes and colors in a
scene. We call it Metadata ++.
There is also room for high level stuff
in the MPEG-7 structure, e.g., which actors are in a video scene. MPEG-7 uses
XML schema language but we are also making a binary version, broadcasters want
this. It is called BiM Binary format for MPEG-7 descriptions.
MPEG-4 has
places in the stream to put MPEG-7 data.
MPEG-7 will include identification of content and hooks for protection. Also, any solution developed in the IPMP extension for MPEG-4 will apply to MPEG-7 descriptions.
Comment from auditorium: Who wants 5+ different music players? SDMI will not solve the problem. More interoperability is required
Koenen noted that the group was currently extending MPEG-4's IPMP architecture. The second MPEG-4 IPMP call for proposals has a July 2000 deadline. 13 submissions have been received as of October 2000.
There are high level requirements
A rights language is nice but not enough. Trust is needed!!
He showed a Diagram of reference model for IPMP extensions for MPEG-4, included the following and other elements:
About MPEG-21 he said, it serves as a framework for: putting it all
together
:
As the current activity, Koenen listed: calls for proposals for developing
the Rights Data Dictionary and Rights Expression Language. The plan is to open
calls for proposals in July. Koenen concluded by insisting this W3C gathering:
this is a huge set of issues, let's work together.
And indeed, during
the rest of the meeting, the question of how W3C might work with MPEG on these
DRM/IPMP issues did come up and will be pursued in the coming weeks.
See also the [Slides
(ppt)] and the Position Paper
Brian Fleischer was presenting for Jeff Albertson.
About RealNetworks, 170Mio users, 50Mio users of RealJukeBox. RealMedia supports many DRM systems. We provide a platform and thus serve several constituencies, including rights holders and end users. These two have different issues. End users want to find things and have a consistent experience. Rights holders want to secure their rights.
We are interested in W3C work on standardization of metadata and DRMs to try to reduce the proliferation of DRMs. This may also include saying what should not be standardized.
In this context, let's talk about what metadata is and what pieces ought to be in the metadata.
Short discussion with the auditorium: Should price be in metadata? Brian says no, person in group says metadata could be a formula to calculate the price, with many variables (including who is asking)?
RealNetworks would propose W3C work on a standard metadata format for describing attributes:
See also the [Slides (ppt)] and the Position Paper
The DRM issues in the mobile are, are mostly related to MITA (Mobile Internet Technical Architecture).
Nokia and mobile are about phones today, but tomorrow we are the Web. We
need DRM. Our concept for end users is to give them a mobile device and say:
this will be your trusted device.
You need to trust the thing in your
hand. Also, on the system, you are anonymous.
DRM must deal with small pieces of content. Media purchases will become smaller and more frequent, time is fragmented, get little pieces. Media is distributed wirelessly, always online.
Durand explained use-cases:
How was intellectual property protected in past? Copyright is a legal instrument. Legal instruments coexist with social and business requirements (privacy, fair use, and accessability). Legal tools in the past needed to have detection of unauthorized use, e.g., walk through Chinatown and look for counterfeits. It leads to prosecution.
How do we protect intellectual property today? Surely, DRM plays a role. There is a legal market that competes with the black market. So the mission is to compete well and thus reduce the impact of piracy. Also we should use watermarking and tracing.
Nokia would like to see the following DRM requirements:
What is the Role for the W3C? An Activity should support interoperability with a standard rights language. This should be content agnostic.
Janina Sajka (AAB): My people like to listen to streams, but getting the real RealMedia player off their Web site using their HTML form is very hard. Is providing a more accessible copy of the registration form on another Web site a copyright violation, if it leads people to the RealNetworks player?
Julian Durand (Nokia): Digital Rights Management in Mobile Media Mobile Information Society brings Internet and wireless together:
John Erickson (HP): Regarding wireless environments and content transformation, what can you add about the wireless channel requirements? I thought you would address transformation and packaging for different target devices. You mentioned end device but not wireless channel
Julian Durand (Nokia): The primary requirements come from the devices, not from the channels. We just have multiple content formats that need to be protected. The backend content system could use the same content and it is formatted for the device profile. Well, there is more transformation needed for devices than for channels. I'm a great believer in XML.
Norman Paskin (DOI): With regard to the MPEG call for proposals: If several proposals are submitted what is the subsequent process, is just one selected?
Rob Koenen (MPEG): We never rubberstamp the best proposal, we try to select the best parts of all proposals. We will try to combine complementary features from all the best proposals. We can join parts of complementary proposals and cause a team to work together.
Melissa Smith-Levine: With regard to the suggestion that the Internet is the repository of all information: That is the role that Libraries thought they had.
Poorvi Vora (HP): We talk about infrastructure of trust. what is the W3C role in this compared to IETF? What would be the best place for standardising that type of thing (as opposed to a language)
Rob Koenen (MPEG): None of the existing fora can simply declare themselves to be a trusted organization. Banks have built trust over many years, the Web has only been around for 10 years. Maybe we will see how to inherit trust from older organizations, from the physical world. But these older organizations may lack technical expertise. Right now, one outfit trusts another, then you begin to build a trust network.Build a trust network over time, not just a trust tree or forest.
Danny Weitzner (W3C): Answer with regard to W3C involvement. Rob and Norman have pointed out the boundaries of the web and the formats are slightly orthogonal. We might have work that we can do together. I appreciate the invitation
Peter Schirling (IBM): We can build a new technology around existing institutions to build trust. Each industry has it's own institution for trust. If someone builds the framework and mechanism by which the process of trust can be enabled, then those industries can use their existing trusted institutions. It's a process of building a new technology.
John Erickson (HP): There is an organization called the TCPA - Trusted Computing Platform Alliance. Members are who are cross industry and major tech vendors. They are working on a trusted hardware reference platform and a trusted BIOS. See trustedpc.org. It's a starting point for us, in terms of managing trust.
[Minutes Overview] [Workshop Home] [Previous:Security] [Next:Closing Session]
Created by Rigo Wenning February 2001
Last update $Date: 2001/04/18 18:24:36 $ by $Author: rigo $