Other Free Encyclopedias » Online Encyclopedia » Encyclopedia - Featured Articles » Contributed Topics from K-O

Multimedia Content Repurposing - Content Repurposing, Approaches to content repurposing, Architectures, Conclusion

server client based user

Abdulmotaleb El Saddik and Md. Shamim Hossain
School of Information Technology and Engineering
University of Ottawa, Ontario, Canada

Definition: Content repurposing is the process of converting multimedia content from one format to another, and then representing it to handcraft the device, network and user characteristics.

The recent advancements in multimedia technology have enabled content providers and consumers alike to gain access to multimedia content over any type of network, via any device, from almost any where, at any time. However, with this rapid growth of pervasive devices, their associated ubiquitous network connections, and their imposed user preferences, new problems, related to distributing multimedia content have emerged:

  • Heterogeneity of pervasive client devices and their ubiquitous connections: Heterogeneous clients range from desktop to ubiquitous information appliances, such as DTV, HTV, Laptop, PDA, Cell phone, smart phone etc. They vary greatly in their processing powers, storage and display capabilities. Their connection speeds and methods which range from slow speed wireless networks to high speed wired networks.
  • Mobility of clients: Clients may very likely move frequently while accessing multimedia content. Most of the time, there is no QoS guarantee and satisfaction is not even a priority. As network traffic may change with time from high speed to congested. Access networks include a wide range of technologies that vary from Ethernet, Bluetooth, and 3G Mobile, to ISDN, xDSL, and GPRS.

One may, and rightly so, deem the task of developing and handcrafting appropriate multimedia content for some, let alone all of the pervasive devices mentioned above, to be a difficult and expensive one. Considering the ubiquitous network connections and their preferences, as well as each of their combinations, the task becomes even unfathomable. A solution to the problem is content repurposing which refers to the conversion process by which multimedia content that is originally designed for a particular device, platform, or user, is transformed to fit other devices, platforms, or users. By doing so, a single copy of the content is preserved in its original form and is easily separable from its corresponding presentation format. Content repurposing uses existing content to enable its use and reuse in different context depending on the various devices, and user profiles. Hundreds of different device profiles are available for accessing online multimedia content. However, the large variety of existing formats, their proprietary access protocol standards, and the lack of general agreement on how to universalize these multimedia contents make it very difficult to use and reuse content.

Content Repurposing

As mentioned above, content repurposing is the process of converting multimedia content from one format to another, then representing it to handcraft the device, network and user characteristics. This transformation can take place in different ways; some of which are:

  • Conversion of different modes (e.g. Speech to text, voice to image, image to text)
  • Conversion of video coding format (e.g. MPEG1 to MPEG4)
  • Conversion of coding parameter

    • Frame rate (e.g. 30fps to 10 fps)
    • Bit rate (e.g. 6 Mbps TV broadcast to 56 kbps cell phone)
    • Spatial resolution (e.g. CIF to QCIF or 4QCIF)
  • Conversion of Spatio-temporal resolution (e.g. VGA to QVGA)
  • Content Summarization (e.g. 2 hours news in 15 minutes by highlighting some of the news)
  • Visualization of Content (e.g. key fame visualization and browsing for a video).
  • Selection of best variation based on MPEG21 part 7 standards:

    • Network Capabilities (e.g. bandwidth, delay and error characteristics).
    • Terminal Capabilities (coding and decoding capabilities, device profiles, and I/O capabilities).
    • User Preferences (e.g. content preferences, presentation preferences, accessibility, mobility and destination).
    • Natural Enviroments (e.g. location, noise level etc.).

Multimedia content repurposing essentially customizes multimedia content and thereby promotes their reuse. Designing and developing multimedia content in a manner that allows the customization, editing, and adaptability to the user’s needs is vital to providing high quality content that is cost effective and sustainable [ 1 ].

The abstract view of content repurposing is shown in Figure 1.

Approaches to content repurposing

According to content variation, multimedia content repurposing is divided into the following approaches:

Static repurposing: In this approach, a server pre-processes and stores multiple versions of the multimedia content. A version is ‘shaped’ by capabilities of network and client and by user preferences, i.e. these factors determine how the content varies. When a client makes a request, the appropriate version is selected from among the existing versions of multimedia content, without any alteration. Most current websites use this approach to avoid extra processing. The advantage of this approach is that there is no processing required since the desired version is already available. This allows for quick delivering and optimal utilization of bandwidth. On the other hand, there is the need for large storage capacities to house the different formats of the multimedia content. Also, every time new formats/versions are invented, content on the server needs to be converted to the new version, rendering the task of maintaining the content a costly one.

Dynamic repurposing: In this approach, after reading multimedia content from the server, it is repurposed on the fly in order to match the client’s capabilities. While this approach requires obviously low storage space because server keeps single version of original multimedia content, it requires significant computing power, as there are many operations involved other than repurposing, such as accepting and dynamically classifying user requests, providing transcoding and personalization, and holding and constantly updating user preference profiles [ 2 ]. This problem could be solved by content summarization, as the repurposing could be referred to as text summarization, format/version changes, reduction of image or audio quality, cutting down number of key frames, audio to text transcoding etc. Due to the heterogeneity and mobility of devices and their ubiquitous networks connections, dynamic content repurposing is critical.

Hybrid Repurposing: This approach, which was introduced by Shin and Koh , uses both the static and the dynamic repurposing approaches selectively in order to preserve bandwidth and storage space. The Static approach is used for the frequently accessed multimedia content, while dynamic approach is used when multimedia content is accessed infrequently. For example, cellular phone accesses multimedia content frequently. In this situation, static repurposing is more appropriate than dynamic, as for each request, the server sends the appropriate version or format to match the client’s capabilities. When the content to be accessed is large and infrequently required, network bandwidth requirements are kept at a low priority. Thus, the hybrid approach, in this case, selects the dynamic repurposing scheme.

Multimedia Content Repurposing Architectures

Repurposing may occur at some point between the creation of the initial content and the final rendering on the client device. This can take place at the client’s end, at the server’s, or at an intermediate proxy between server and client. Based on the architecture, repurposing can be classified into: client-based, server-based, or proxy-based. Each of the mentioned architectures possesses strengths and weaknesses. The main issue is to consider the effectiveness of content repurposing and the efficiency of utilization of the client capabilities in terms of processing power, bandwidth, and storage capacity.

Client Based Repurposing Architecture: This is a raw form of repurposing which is performed mainly on the client devices, without affecting the communication protocol.

This approach helps to avoid transporting large volumes of device and user metadata to the server or the intermediate proxy for repurposing. The repurposing approach is performed on the client device without changing its characteristics, by transferring the entire application from the server to the client. For example, multimedia content and its corresponding style sheets are delivered to the client. Then conversion is performed. Here, multimedia content may include text, images, audio, and video. Another way to client-based repurposing is to deliver the entire multimedia content to the client for the appropriate navigation and selection, based on the client capabilities.

The W3C consortium proposed a client-based repurposing approach including some techniques such as: image resizing, font substitution, transcoding, dedicated rendering and contextual selection.

Used with thin clients such as mobile computing and cellular phones, client-side architecture has significant limitations including restrictions of network bandwidth, device memory, and processing power. However, there are some advantages such as incurring the adaptation overheads at the client site, rather than server end. This architecture is also advantageous when the server is unable to determine device capabilities from the request. Windows CE, e.g., version 3.0 is an example for client-side repurposing. It provides a client-side repurposing for some rich multimedia content by using a color reduction feature which is useful when presenting high quality images on low-color displays of mobile and handheld devices. Client-based repurposing architecture does not depend on the employed communication protocol. The approach helps to avoid the sending of large extents of device and user profile information to the server or intermediate proxies for repurposing. Opera Software is another example for client side architecture. It introduces small screen rendering (SSR) and medium screen rendering (MSR). SSR repurposes contents and generates a user-friendly version of the contents for small mobile browsers, like smartphone, therefore eliminating the need for horizontal scrolling. MSR identifies the Web page’s content and adapts these different elements individually to fit medium-sized screens, ranging from PDAs to low resolution TVs. Original fonts, colors, design and style are left virtually untouched.

Server-based repurposing: In this architecture, repurposing is performed at the server by determining client’s capabilities in terms of storage, network bandwidth, processing power, screen size, media format (e.g. HTML, XML, WML etc), and user preferences (e.g. language, fonts etc.). Both static (off-line) and dynamic (on the fly) repurposing are supported by this architecture. The Static approach generates pre-repurposed variants of the same content off-line in order to match client and user capabilities, while the dynamic approach repurposes content on the fly when a request is received. For example, due to the lack of proper translation engines, multimedia content with different languages is appropriately repurposed by the static approach; the dynamic approach is not useful here. Language preferences can be accepted via content negotiation, using the HTTP requester header field. The server can discover the appropriate content version by parsing the HTTP header as the case in using MPEG-21 digital item adaptation.

In the server-side architecture, authors can control the repurposed result under different client characteristics and user preferences with a lower cost, but in reality, it cannot Page 498  repurpose the content to complex dynamic context. This architecture may depend on the communication protocol used in the content delivery. For example, in Stateful protocols, such as RTP, the server requires the client’s response to perform efficient repurposing. However, in a stateless protocol such as HTTP, it does not happen. In that case, repurposing depends on the client or server interaction. Existing Server-side repurposing solutions include IBM WebSphere Tarnscoding Publisher, BEA Weblogic and AvantGo. All of these solutions are suitable for wired and wireless clients. An example of server-side repurposing is what is used by some news providers such as CNN or Canadian Broadcasting Corporation (CBC). They provide several different media formats on the server representing the original media content. This media content can be selected by the client, based on its preferences and capabilities. For example, three different formats: QuickTime, Real Player, and Windows Media Player, are offered by CBC or CNN, employing three different network connections: Dial-up (56 kbps), GPRS (144 Kbps) and DSL Basic (256 kbps). When a user wants to watch a streaming video from CBC or CNN, he or she should select the format supported by his hardware or software and the network connection that he or she is using.

The advantages of using Server-side adaptation architecture are:

  • Content creators have more control on how the content is presented to the clients.
  • The server offers more processing power than the client devices, this is very important in the case of Video transcoding.
  • In the event of the existence of multimedia content in XML. XSLT is used to transform this content to another appropriate markup language in order to match the browser’s capabilitiespresentation such as using Scalable Vector Graphics (SVG), and filtering HTML documents to WML.

As for the weakness of using server-based content adaptation architecture, they can be summed up by the following:

  • Not all browsers support content negotiation, or the server may not have control over all browsers. So, the server must make assumptions or use default parameters based on the browser’s ability to present the content.
  • Heavy server-side applications may slow down the server.

Proxy-based architecture: In this architecture, the proxy server is located on the network between server and client. The proxy, which makes request to the server on behalf of the client, receives content response from the server, analyzes and repurposes the requested content on the fly, and finally sends the repurposed content back to the client, based on its capabilities. The repurposing performed by the proxy is transparent to the clients and server. The proxy can cache the repurposed results for future use, thus reducing the need for re-repurposing and maintaining multiple variants of the same content. This architecture which supports hybrid repurposing, addresses the heterogeneity and mobility problem of the devices, and reduces the overall end to end response time. It also reduces the size of the multimedia content object while maintaining its semantic value. Existing proxy-based repurposing include IBM’s Web Intermediaries (WBI), UC Berkeley’s Ninja project, RabbiT proxy server etc.

In proxy-based architecture, the proxy can control the delivery by applying appropriate QoS algorithm. Some multimedia content repurposing such as audio and video requires transcoding. There may be one or more transcoders in one proxy. Figure 2 depicts a simple architecture of a proxy based repurposing that uses only one transcoder with different transcoding capabilities to deliver to different devices. The content on the server which requires 256 kbps bandwidth is transcoded into four different formats, for four different clients, over four different connections. The transcoded streams are trafficked over lower bandwidths than the original one; they require 128 kbps, 64 kbps, 32 kbps and 16 kbps respectively. Furthermore, the UMTS, GPRS, DVI, and GSM require less computing power than MPEG-1. The audio/video quality of the transcoded streams is lower than that of the original. However, the clients would probably opt for lower, yet acceptable quality, rather than discontinuous video clip or news due to the lack of bandwidths.

Transcoding also can be defined as a combinatorial process, where multiple transcoders can be chained together in order to obtain a larger transcoding sequence. For example, for the task of transcoding MPEG-2 video to Quicktime, the video could first be transcoded from MPEG-2 to MPEG-1 using one transcoder, then the output could be passed on to another transcoder in order to obtain Quicktime. The result would be exactly the same as if transcoding had been performed in one step. This approach, first of all, ensures that even with a limited number of transcoders, more complicated transcodings can be done, and second of all, allows for the computations to be distributed.

Conclusion

Content Repurposing, or adaptation, has the potential to play a key role in the successful distribution and access of multimedia content. The rapid increase in diversity of multimedia content types and formats, and the ubiquitous nature of accessing them from heterogeneous pervasive limited capability devices and networks are currently major challenges. In order to meet these challenges, the multimedia community proposed MPEG-21, which is an open standard multimedia based framework for multimedia content delivery, adaptation, personalization consumption, and presentation.

Multimedia Design [next] [back] Multimedia Content Modeling in Cosmos-7

User Comments

Your email address will be altered so spam harvesting bots can't read it easily.
Hide my email completely instead?

Cancel or