Home Media Server content management
`Andrew Tokmakoff and Harry van Vliet
`
`Telematica Instituut
`P.O. Box 589
`Enschede 7500 AN
`Netherlands
`Phone: +31 (0)53-4850461
`Fax: +31 (0)53-4850400
`Email: [tokmakoff, vanvliet]@telin.nl
`
`
`ABSTRACT
`With the advent of set-top boxes, the convergence of TV (broadcasting) and PC (Internet) is set to enter the home
`environment. Currently, a great deal of activity is occurring in developing standards (TV-Anytime Forum) and devices
`(TiVo) for local storage on Home Media Servers (HMS). These devices lie at the heart of convergence of the triad:
`communications/networks - content/media - computing/software. Besides massive storage capacity and being a
`communications ‘gateway’, the home media server is characterised by the ability to handle metadata and software that
`provides an easy to use on-screen interface and intelligent search/content handling facilities.
`
`In this paper, we describe a research prototype HMS that is being developed within the GigaCE1 project at the
`Telematica Instituut2. Our prototype demonstrates advanced search and retrieval (video browsing), adaptive user
`profiling and an innovative 3D component of the Electronic Program Guide (EPG) which represents online presence.
`We discuss the use of MPEG-7 for representing metadata, the use of MPEG-21 working draft standards for content
`identification, description and rights expression, and the use of HMS peer-to-peer content distribution approaches.
`Finally, we outline explorative user behaviour experiments that aim to investigate the effectiveness of the prototype
`HMS during development.
`
`Keywords: Home Media Server, Personal Digital Recorder, Personalisation, Electronic Program Guide, Content
`Distribution, Video Browsing, T-Commerce
`
`1. INTRODUCTION
`In what the French call ‘the ballet of the electrons’, the worlds of Information and Communications Technology (ICT),
`Internet and broadcast television are coming together, stimulated by the digital revolution and with the mission to
`conquer the home environment. Currently the most prominent weapons in this campaign are the set-top box and the
`Personal Digital Recorder (PDR) [1]. Positioned as a ‘link’ between the home network and the networks (both the
`broadcast network and Internet) that are ‘out there’, it represents the dream of the broadcast, entertainment and telecom
`industry: the ‘connected’ home.
`
`Set-top boxes have already been with us for many years. The first generation of set-top boxes were capable of only
`receiving and unscrambling analogue transmissions and displaying the results on the TV set. But set-top boxes are
`quickly evolving from simple gateways for de-scrambling television signals towards powerful devices for interactive
`services. A recent step toward this has been the introduction of the PDR, a consumer device that includes high capacity
`disk storage. Commercial examples include TiVo, ReplayTV and UltimateTV. PDRs give consumers the opportunity to
`‘time-shift’; i.e. to watch what they want, when they want. More importantly, such PDRs come with specific services for
`personalisation: like an electronic program guide that shows (filtered) information and gives suggestions on possibly
`interesting content. The addition of a back channel provides a mechanism to allow for commercial services including
`
`1 The GigaCE (GigaPort Content Engineering) project (gigace.telin.nl) is a knowledge acquisition project focussing on content engineering, which is
`part of the GigaPort project (www.gigaport.nl). GigaPort prepares the Dutch industry and knowledge institutes for Next Generation Internet. GigaPort
`is a joint initiative of industry, higher education, research institutes and the Dutch government.
`2 http://www.telin.nl
`
`168
`
`Internet Multimedia Management Systems II, John R. Smith, Sethuraman Panchanathan, C.-C. Jay Kuo,
`Chinh Le, Editors, Proceedings of SPIE Vol. 4519 (2001) © 2001 SPIE · 0277-786X/01/$15.00
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/17/2014 Terms of Use: http://spiedl.org/terms
`
`Apple 1011
`
`

`

`storage and profile management. In our opinion, it will be this demand for new services like archiving/storage,
`interactivity, personalisation, et cetera that will push the set top box, PDR, or a similar ‘box’, from a simple interface for
`de-scrambling signals or a super-VCR for time-shifting towards a central server device within the home environment. A
`comprehensive overview of the convergence between TV and the next generation Internet is presented in [2].
`
`Such a device, the Home Media Server (HMS), has communications as its main characteristic. In a world where media
`convergence creates a unified environment where content is available from a variety of sources, and where technology is
`erasing the boundaries between televisions, telephones and computers, the HMS is a central device for the consumer to
`control his communications, whether it be telephone, Internet narrowcasting or television broadcasting. Many services
`will evolve support consumers as they struggle to manage this information glut, since we all now live in a world where
`there is so much content and so little time.
`
`Within the GigaCE project at the Telematica Instituut we are developing a research prototype of a HMS that implements
`services that go beyond the current PDRs but which can already be found in other contexts, such as the Internet or
`mobile telephony. For instance, we investigate ‘presence’ functionality in a TV environment, so users can tell who else
`of their peers are watching a specific program. We also investigate annotating TV-content for indexing and search and
`retrieval purposes, along with peer-to-peer content distribution approaches. In this paper we address the framework
`within which this prototype is being developed by describing the use of MPEG-7 for representing metadata and the use
`of MPEG-21 for content identification, description and rights management. Furthermore we will elaborate on the
`specific services which will be part of our own prototype.
`
`2. BASIC FUNCTIONALITY
`If we take the current set-top boxes and PDRs as a starting point, we can say that the offered functionality is basic
`indeed. Through the set-top box we get (conditional) access to analogue and digital broadcast data, which means more
`channels with (sometimes) better picture and sound quality. By way of the PDR we can ‘timeshift’: watch what we want
`when we want. This is something that we have heard before, when the VCR was introduced. Since then, the VCR has
`developed into both a ubiquitous and marginal device, which is primarily used to play pre-recorded (children’s) video
`tapes [3].
`
`Of course the digital format and the instant access of the hard-disk give the PDR some profound extensions to basic
`VCR functionality such as:
`•
`•
`•
`
`if joining a (live) program halfway, the system must be able to capture the content from the beginning.
`continue recording a program even when it may jump channels.
`record a program/advertisement on a basis of now, when it is on for next ‘x’ shows, all episodes, just the highlights,
`record the next trailer, etcetera. Repeats or the same transmission on other channels are resolved.
`pause a live event and return to it later. Related to this, be able to fast forward back to real time or to skip forward
`through index points.
`instant fast-forward and rewind (the latter also in live programs).
`simultaneous record and playback.
`
`•
`
`•
`•
`
`However, we do not believe that it will be this “super-VCR” functionality that will be the driving force for consumers to
`buy a new device. What can distinguish the PDR from the VCR, though, is its usability. How many VCR-displays don’t
`display ‘set clock’, because the consumer has given up on programming the device? The PDR instead can offer an easy
`interface to available content by means of an Electronic Program Guide (EPG). Consumers can search through the EPG
`and record shows by interacting with the EPG grid. It is easy (‘clicking’) and reminds us of using a regular printed TV
`guide. Moreover, if personalisation is incorporated, all kinds of user support services can be envisioned, for instance:
`
`•
`
`build your own TV guide– so there are fewer things to look at. This can be based on genre, time, channel, date, sub-
`genre specific, or any combination thereof, or on already pre-filtered content that the user profile has created.
`• make use of channels of pre-filtered content, third parties offering ‘best-of’ from all content providers they cover, or
`pre-formatted EPGs based on viewer preferences.
`highlight (e.g. use of colours) of favourite programs in the program overview.
`
`•
`
`Proc. SPIE Vol. 4519
`
`169
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/17/2014 Terms of Use: http://spiedl.org/terms
`
`Apple 1011
`
`

`

`However, even with these enhancements the PDR still has a limited ‘range’. In a HMS environment we must envision an
`“open” environment, in which (digital) content is widely available from different sources, communication is ubiquitous
`and devices exchange information and content transparently.
`
`In such an environment, new questions arise and new service opportunities become possible. An example of this is
`integrated search facilities for content and an integrated presentation of the search results (see section 5). Currently,
`broadcast content, the Internet and the home network all have their own way of allowing the user search for specific
`content and their own ways of presenting the search results. For broadcasting this is the Electronic Program Guide
`(EPG), for the Internet these are search engines and directory services, and for locally stored digital content this is a PC-
`like file system interface. If we truly want a unified 'portal', an integration of these facilities should be pursued [4].
`Another example is the portability of user profiles, so they are no longer tied to the home but instead, to the actual
`person. We will address these new services opportunities and their possible implementation in the following sections.
`
`3. PROFILING AND PERSONALISATION
`As mentioned above, one of the most promising aspects of the HMS is its capability to provide a personalised user
`experience. In a personalised information system, information is tailored towards the user in a given situation and
`context. The overall effect is to give the user an optimal individual experience, which targets the specific goals of the
`user. Examples of such systems are personalised newspapers, personalised online store catalogues, publication
`recommendation systems, personalised TV guides, and personalised lectures.
`
`In order to personalise such services, there is a need to collect user information and build a user-specific profile. This
`profile forms the basis for personalised services, which utilise the user-specific profile to adjust their system behaviour.
`The “Content Handling and Usage” element of the MPEG-21 Multimedia Framework Standardisation work [5] involves
`identifying approaches for creating, modifying and managing User Profiles. In addition, the ability to interchange
`profiles between applications and systems is an important goal.
`
`3.1.
`Profile creation and maintenance
`We are developing a profile format, based on the TV-Anytime Metadata Model [6] which allows multiple applications to
`use a single profile. The profile consists of both generic user information and also application-specific elements. For
`example, the user’s preferred language is a standard element which can be used by all applications that utilise the profile.
`In contrast, the HMS is an application which adds unique aspects to the profile (such as the user’s favourite TV movie
`genre) which is possibly not of interest to every application using the profile (e.g. an mp3/cd player). We expect that
`MPEG-7 will be useful in this work on developing portable and interoperable profiles.
`
`An important point is that a user profile is (ideally) created once, usable in a specific context and inter-changeable
`between different contexts and application domains. It is also important to be able to consolidate multiple profiles
`created by different applications into a single portable profile. This is currently an area of further work in the project.
`
`When designing a personalised information system it is important to choose the most appropriate method for profile
`initialisation and maintenance.
`
`• Initialisation: When a personalised information system is used for the first time, it has no knowledge about the
`interests of its users and thus, must obtain information from the user to allow a user profile to be initialised. There are
`several possibilities for the initialisation of a user profile. With explicit creation, the user creates the profile explicitly
`(e.g. by adding keywords to the profile). It is also possible that the user is questioned by the system and based on the
`answers, the user’s interests are deduced. Another option is to let the user select one or more pre-defined profiles to
`start with, which function like a template for the user profile. Some systems allow implicit profile creation where the
`user is able to start using the system and the system builds a user profile purely by analysis of the users actions.
`• Maintenance: It is important for personalised information systems to keep user information up-to-date, especially for
`systems that are used more than once. For systems that are only used once it is not necessary to store the user
`interests, as users will not use that same system again (e.g. a museum information system for tourists inside a
`museum). Keeping information about the user up-to-date can be done in two ways: using feedback from the user or by
`allowing the user to manually update the user profile. Concerning the former, there are two main ways that feedback
`
`170
`
`Proc. SPIE Vol. 4519
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/17/2014 Terms of Use: http://spiedl.org/terms
`
`Apple 1011
`
`

`

`can be used to maintain a user profile: by analysing the usage behaviour, which is called implicit feedback, and by
`using explicit relevance feedback. In our prototype HMS, a logging component gathers and stores user actions for
`implicit feedback while a Profile Maintenance process analyses this information and performs the necessary updates
`of the user profile. Explicit feedback is gathered directly from the user via the Presentation (user interface) and
`processed by a Profile Maintenance process.
`
`Our prototype HMS system will use a combination of the above approaches; the option to select a pre-defined profile,
`the ability to explicitly modify user profiles and most importantly, automatic profile generation based on user actions.
`
`3.2.
`Personalisation of system functionality
`The prototype HMS uses a centralised personalisation component which is responsible for providing personalisation
`services to other functional components in the system. For example, the Presentation aspect of the User Interface can
`make use of the personalisation component to alter the way it presents information to the user. This can be thought of as
`an adaptive user interface, much in the same style as the adaptive “Start menu” introduced into the Windows 20003 user
`interface. For each of the functionalities in the system which requires personalisation, a new sub-component is required.
`These sub-components include the Search, Presentation and of course, Content Suggestion.
`
`3.3.
`Approaches to Personalisation
`Personalisation, which is the adaptation of a system to the user, can be applied to many different aspects of the HMS. It
`can be used to adjust the user interface by selecting and presenting only specific information to the user and even
`translating a piece of information into the user’s preferred language. All automatic personalisation is based on the
`knowledge the system has about the user which is stored in the user profile. In the demonstrator we are developing,
`personalisation component acts as an advisor for other components. E.g. the search component uses the personalisation
`component to rank search results and to modify the search strings sent to search engines, based upon the user’s profile.
`It is also used to help the system make decisions on what up-coming content in the EPG the user may find interesting
`based upon previous viewing. Based upon the advice of the personalisation component, the system can make a decision
`whether to record the program item or not.
`
`Using a personalisation component as an abstraction layer between the different processes that need personalisation
`advice and the user profile has to following advantages:
`• The different processes do not need to have knowledge about the structure of the user profile;
`• Personalisation techniques can be implemented once and be easily re-used for advising different components;
`• New personalisation techniques can be introduced, without having to update all components that require
`personalisation advice.
`
`Depending on the type of personalisation (user interface adjust or personalised selection of information), there are
`different types of personalisation techniques to do this. Because our focus lies on the personalised selection of
`information, the different techniques available will be discussed here briefly:
`• Category selection: With this technique all information is segmented into categories. The user profile of a user
`contains the categories the user is interested in. In the personalised selection process, only information that belongs
`to the categories the user is interested in is retrieved. e.g. My Yahoo! (http://my.yahoo.com).
`• Query adjustment: With query adjustment the search query (which is either explicitly created by the user when
`querying or implicitly created when browsing or filtering) for a database or search engine is changed based on the
`knowledge the system has about the different influencing factors.
`Information filtering: This is the process in which a system filters a vast amount of information and only delivers or
`recommends information that it considers relevant or interesting to the specific user. The key aspect of information
`filtering is that the system bases its recommendation on the content and its knowledge about the interests of the user.
`Social filtering: These systems (also called collaborative filtering) search for users with similar interests and
`recommends items those users liked. Instead of computing the similarity between items (content) as in information
`filtering, the system computes the similarity between users. How others liked a piece of information is based on how
`
`•
`
`•
`
`
`3 http://www.microsoft.com/windows2000
`
`Proc. SPIE Vol. 4519
`
`171
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/17/2014 Terms of Use: http://spiedl.org/terms
`
`Apple 1011
`
`

`

`they rated these items (explicitly and/or implicitly). An advantage of social filtering is that it works on all media
`types, as no content analysis is needed.
`• Collaborative information filtering: This approach combines techniques from social filtering (finding similar users)
`and techniques from information filtering (filtering based on content). Hirsh, et.al. [7] argue that the combination of
`information filtering and collaborative methods are better than each of them separately.
`
`3.4.
`Demonstration : Portable and Interoperable Profiles
`A small demonstrator that illustrates some of our Profiling & Personalisation ideas is under development and consists of
`the following:
`
`•
`
`Initially, we define and initialise a user profile based upon simple user feedback (a web-based application). This
`profile is stored in the HMS itself.
`• Using this profile, the HMS provides simple content recommendations based upon the user profile.
`• The profile may be transferred from the HMS to a PDA and back (the PDA is only used as portable storage device).
`While on the PDA, the profile is used for other applications (such as an mp3 player) and modified as necessary.
`• When profile is returned to the HMS, the roaming profile is “re-integrated” into the HMS’s local profile, taking into
`account changes in the profile on the PDA.
`• The profile is then shown to be interoperable between devices and applications by showing a “before and after”
`comparison of content suggestion. Before the profile is modified, a set of content is suggested whereas after the
`profile returns, a different set of content is suggested.
`
`An extension of this scenario envisions a group of friends that are having a gathering and who decide to watch some
`content together. In order to come to a simple and fast decision on what to watch, each group member’s personal profile
`(conveniently located on their PDA) is transmitted to the host HMS system, at which point the system recommends TV
`programs for them to watch together.
`
`4. STORAGE OF CONTENT AND METADATA
`With large amounts of multimedia content being captured by the HMS, there is a need for a storage component that
`efficiently manages the storage and subsequent retrieval of content and its associated metadata. The ability to save
`content to external storage (ie CD-RW) is an important new functionality that extends the usefulness of the PDR, but
`must also respect the associated Digital Item (DI) Intellectual Property (IP) rights.
`
`We adopt the “fridge” and “freezer” model in which content that is to be used in the short-term is stored in a so-called
`fridge, whereas content that is to be kept for the long-term is stored in a freezer. Content in the fridge is accessible at all
`times since it is located within the PDR, whereas freezer content is located on external storage. In both cases, however,
`in order to manage vast quantities of stored content, it must be easy to search through and retrieve relevant pieces of
`content. This, of course, requires the use of metadata, which provides information on the content in a searchable form.
`
`4.1. MPEG-7 and Content Metadata
`Metadata is useful for two different kinds of processes: automatic search and automatic personalization of information.
`Also there are two types of metadata: content identification metadata, which describes (multimedia) documents as a
`whole, and content based metadata, which gives access to subparts of (multimedia) documents such as an occurrence of
`a specific keyword. Hence, metadata contribute to reusability and accessibility of content.
`
`MPEG-7, officially named the "Multimedia Content Description Interface", is an emerging standard for content-based
`indexing for digital libraries. It provides standardisation of multimedia content descriptions. The elements standardised
`by MPEG-7 allow a range of applications as broad as possible. Since only the descriptor types are standardised, it is
`generally applicable for all types of content. Thus, all sorts of audio-visual material may be disclosed by the use of
`MPEG-7, including still pictures, graphics, 3D models, audio, speech, video, and information on how these elements are
`combined in a multimedia presentation (scenarios, composition information).
`
`172
`
`Proc. SPIE Vol. 4519
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/17/2014 Terms of Use: http://spiedl.org/terms
`
`Apple 1011
`
`

`

`MPEG-7 is currently being developed as an international standard, with a great deal of effort coming from both the
`scientific and industrial fields. It is therefore expected that the proposed ISO/IEC standard (to be delivered in September
`2001) will have a sound foundation in both the academic and commercial worlds. With respect to other metadata
`standards, MPEG-7 is the standard that provides the richest and most versatile set of AV feature descriptions since it
`embraces a family of standards such as SMPTE, EBU, TV-Anytime, Dublin Core and more. It thereby aims to be the
`highest interoperable standard among well-known industry standards. We intend to make use of the practical MPEG-7
`knowledge gained in the VIP project [8] when representing metadata and apply lessons learned within that project when
`developing the prototype HMS system.
`
`4.2.
`Automatic Metadata generation
`Consider a PDR hardware platform that is much more powerful than today’s standard set-top boxes which are often
`supplied with cable service subscriptions and thus, tend to be the “cheapest” possible box. We assume a platform similar
`to that of the Nokia Home Media Server4, which uses an Intel x86 architecture with quite significant processing and
`memory resources. With such a hardware platform, it becomes possible to make use of “idle” processing time to
`generate metadata for content already stored within the PDR.
`
`The content may be, for example, personal home videos created by the PDR’s owner. Such content is generally
`completely free of metadata excepting, perhaps, some basic metadata regarding the topic, date and time. The ability to
`automatically segment the video content may be accomplished by providing the HMS with a metadata creation function.
`This function processes content as a background task while the HMS is idle. In the same manner, using a plug-in
`architecture, we plan to allow for many different metadata creation modules to be added. For example, it is envisioned
`that a people-finding module could be used to find specific people, or an advertisement detection module could be used
`to segment advertising content from the rest of the programming.
`
`In many cases, it would also be possible for users to purchase metadata associated with broadcast video content that they
`have recorded. For example, metadata that indicates segments within a cooking show would aid the viewer navigate to
`sections that are of most interest and could be made available for an extra cost.
`
`4.3.
`Storage Technologies
`It is envisaged that in our prototype HMS, a standard file system will be used to store multimedia assets (i.e. mp3, mpg
`and html files). Since MPEG-7 metadata is described using XML and is generally highly hierarchical, making use of an
`XML-enabled database is a good option. This is because an XML-enabled database provides many of the performance
`advantages of traditional Relational Database Management Systems (RDMS) whilst also allowing hierarchical XML
`documents to retain their structure. Currently, many native XML databases suffer from performance issues (which can
`be mitigated through effective indexing) and are still a new/unproven technology.
`
`4.4.
`Demonstration : Sticky Notes
`In order to demonstrate the use of user-generated Metadata and its subsequent use for search and retrieval of content, we
`are developing a demonstrator that enables users to do the following:
`
`•
`
`• whilst watching some content (either from broadcast or locally created content), the user decides that there is a point
`in the content which is worth marking for later reference, much like creating a bookmark or posting a “sticky note”.
`using the PDA as a remote control device, the user is able to enter a textual comment and/or keywords which forms
`the content of the note.
`this note is then stored in a database, along with the specified keywords, the media identifier and media time.
`at a later date, the user may conduct a search through the content (either located in the fridge or the freezer) using
`the keywords and note text as searchable fields.
`
`•
`•
`
`The functionality described above allows the user to take note of interesting points in content and to easily find them
`again at a later time.
`
`
`4 http://www.nokia.com/multimedia/mediaterminal.html
`
`Proc. SPIE Vol. 4519
`
`173
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/17/2014 Terms of Use: http://spiedl.org/terms
`
`Apple 1011
`
`

`

`5. SEARCH AND RETRIEVAL
`The unified content environment encountered in the TV home environment is made up of three basic sources: broadcast
`content, the Internet and the home network, which consists of locally stored digital content (music libraries, family
`photos, documents, stored video, etc). We propose an integrated approach to Searching which provides a single
`interface through which the user may locate content on multiple transmission media. It also provides an integrated result
`presentation interface which links results from different sources to each other.
`
`Having obtained content and stored it in the HMS, the obvious next step is to retrieve it when it is to be viewed at a later
`date. This involves searching through locally stored content but may also include searching through other mediums such
`as the broadcast channels, Internet sites and even other HMS’s (peers – see Section 6.3) for related content. We make
`use of a PocketPC5 based PDA a remote control. Such a PDA offers more opportunities for designing a usable interface
`that corresponds more closely to the EPG on the screen than current remote controls that come with a regular VCR or
`PDR. In Figure 1, a mock-up of the search-function on the PDA remote-control which forms the basis for further
`development is shown. Figure 2 shows the results of the search.
`
`Figure 1: Searching with the PDA
`
`Figure 2: Results of the search
`
`5.1.
`Demonstration : Browsing Video Content
`As storage costs drop and the amount of available storage capacity in HMS systems increases, it is likely that HMS’s
`will host many gigabytes of storage capacity. When making use of PDR functionality, it follows that it is possible for
`the HMS to automatically record many hours of content for a user over the space of a few days. This leaves the user
`with a dilemma – how decide what to watch in the small amount of time available. If the user must waste time wading
`through many hours of recorded content, then the time-saving aspect of the PDR is lost. There is a need for fast and
`effective browsing of the recorded content.
`
`Consider a user that goes on holiday for 5 days and returns to find 10 hours of recorded content (a total of 15 shows)
`stored on the HMS. In order to sort the programs effectively, the user first sorts them by genre – e.g. cooking. By
`reading the program descriptions (in the associated metadata), the user is able to make initial decisions on what to
`keep and what to delete. Making use of the “skip” functionality, the user is then able to skip over the program’s
`introduction credits and get quickly to the welcome scene. This approach allows the user to get a feel for the program
`and its contents rather than be limited to only a textual description of the program. Since we integrate Internet content
`with broadcast video content, the user is also able to access related material sourced from the Internet and displayed
`alongside the program. For each program that has been retained from the first phase of selection, the user may open a
`“storyboard” which displays the important key-frames associated with each scene in the program. Each of these
`storyboard elements may be “activated” which starts them playing. This is useful since the user may watch specific
`
`5 http://www.microsoft.com/pocketpc
`
`174
`
`Proc. SPIE Vol. 4519
`
`Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/17/2014 Terms of Use: http://spiedl.org/terms
`
`Apple 1011
`
`

`

`sections of the material while also viewing the structure of the entire program. For an even quicker summary of the
`program, the user may view a 10% summary which is a slideshow of frames from the program. This summary may
`be paused and re-started as normal or double-speed motion video when an interesting point in the content is reached.
`
`These techniques allow the user to make fast and effective decisions on what content is worth time to watch.
`
`6. DIGITAL RIGHTS MANAGEMENT AND CONTENT SHARING
`
`All current PDR implementations are “sealed” in that they do not allow content to get out of the “box” once it has
`entered. In addition, quite strong measures are taken (e.g. TiVo and UltimateTV) to ensure that the content inside such
`systems is protected from tampering by external sources (e.g. by encryption). These measures are of course, driven by
`the need to protect the Intellectual Property Rights associated with the Digital Items (DI) created by the PDR.
`
`In contrast to the more exclusionary approach taken by current PDRs, we take a cue from the lessons learned from
`widespread Napster usage. Instead of trying to protect the IP of DI’s by ensuring scarcity [9], we promote the use and
`movement of content between end-users whilst still encouraging IP rights to be respected and new business models to
`emerge.
`
`6.1.
`Content Declaration, Identification and Description
`The f