Tuesday, November 3, 2009

MPEG Media Transport is taking shape

Achievements at the 90th MPEG meeting in Xi’an, China with respect to MPEG Media Transport (MMT):
The first document, context & objectives, describes the scope of the standard based – as usual – on a number of use cases. These are adaptivity, delivery, QoS/QoE-awareness, convergence, and transparency which are briefly highlighted in the following:
  • Adaptivity is referred to adaptability of multimedia content towards the usage environment context in which the content is being used. Ultimately, this includes the end user and his environment but also the delivery and other areas.
  • Delivery differentiates between (traditional) IPTV, progressive download, and peer-to-peer delivery.
  • QoS/QoE-awareness refers to several optimization techniques which roughly can be clustered into application-level and cross-layer optimizations. The former is further divided into retransmission and buffer management as well as forward error correction (FEC). The latter can be categorized into bottom-up, top-down, and integrated approaches well-known from the research literature.
  • Convergence includes topics related to network convergence, i.e., media-aware devices, and service convergence.
  • Transparency is mainly referred to as transparency with respect to different content protection schemes and rights management systems.
The current requirements document, however, focuses on two main areas which are adaptive delivery and cross-layer optimization. A couple of general requirements (i.e., IMHO a wish list we all pretty much agree on) precede the requirements pertaining to these two selected topics which are listed below followed by my comments:

Anyway, all the technologies that will be submitted as a response to a call for proposals – most likely this will happen in April 2010 – will have to show evidence that they are better than their predecessors (i.e., MPEG-2 TS) but it is pretty much unclear how this evaluation is going to be performed.

Some people argue that the functionality of adaptive, progressive download could be achieved with (minor) extensions of the MP4 file format but is something the industry definitely needs right now. Therefore, solutions for this part of MMT shall be standardized without any delay, i.e., at maximum warp speed! In my view, such a requirement should be included in the list above although it is kind of obvious…

Cross-layer (XL) optimization is a research topic which is around for quite some time with pretty cool results. However, the results mainly stem from XL stuff being done at the lower layers, i.e., link layer and so on. So what MPEG can offer is to standardize an interface – parameter names & semantics – to XL optimization modules whereas these modules should be out-of-scope and left open for industry competition. A kind of interface that could be used for this purpose is Supplemental Enhancement Information (SEI) – originally standardized during the course of the AVC (MPEG-4 Part 10 | H.264) developments – which nowadays can be included in MPEG-2 TS, file formats based on the ISO Base Media File Format, and also RTP. Thus, what is needed is simply a registration authority for SEI messages in the context of XL optimization. However, in order to represent advanced XL models, further standardization might be needed but we have already demonstrated how this could be accomplished existing MPEG-21 Digital Item Adaptation tools [1][2][3]. As in the case of SEI message, the only missing link is some kind of registration authority for parameter names and semantics which have not been considered when the standard has been finally approved – that was already in 2003.

Finally, I’d like to mention the Ad-hoc Group (AhG) where discussions between meetings take place and the workshop in January 2010 on the Wednesday during the Kyoto meeting focusing on “adaptive progressive transport” and “cross-layer design”. Details about the workshop you can find here. Note that usually such a workshop is open to everybody, i.e., you don’t need to be a member of ISO/MPEG in order to attend this workshop.

[1] I. Kofler, C. Timmerer, H. Hellwagner, and T. Ahmed: “Towards MPEG-21-based Cross-layer Multimedia Content Adaptation”, Proc. 2nd International Workshop on Semantic Media Adaptation and Personalization (SMAP 2007), London, UK, Dec. 2007.
[2] C. Timmerer, V. H. Ortega, J. M. González, and A. León: “Measuring Quality of Experience for MPEG-21-based Cross-Layer Multimedia Content Adaptation”, to appear in Proceedings of the 1st ACS/IEEE International Workshop on Wireless Internet Services (WISe’08), Doha, Qatar, April 1-4, 2008.
[3] I. Kofler, J. Seidl, C. Timmerer, H. Hellwagner, I. Djama and T. Ahmed, “Using MPEG-21 for cross-layer multimedia content adaptation”, Journal on Signal, Image and Video Processing, Springer, vol. 2, no. 4, Dec. 2008.

No comments: