Content Distribution

DASH technology in itself is quite mature (considering the fact that it has been recently been standardised by ISO/MPEG) because of its principles being used by vendors in the field for several years (most importantly Apple HLS). Hence, the innovation in ICoSOLE will firstly focus on the efficient delivery of non-traditional video sequences using the DASH principles. As mentioned before, spatial segmentation can be used in conjunction with the temporal segmentation of DASH; further advances versus the current state of the art are envisaged when considering competing/overlapping panoramic sequences or, free-viewpoint Video fragments or blended movies.

A second innovation focuses on minimizing the startup delay encountered with all DASH based transmission methods. This is due to the fact that the decoder needs a certain amount of data for its initialization and the time associated with this is typically a multiple of the segment duration time (in the order of a few seconds). While, for typical video consumption, this is not really an issue, it will present problems when fast switching between videos (e.g. in a blended sequence or when switching between multiple UGC movies) is required. As this minimization cannot (or very minimally) be accomplished by optimizing the protocol, a more high level solution is required, instructing the application and its supporting codec(s) to prefetch information based on the semantic information of the application (e.g. based on user profiling or a best-guess approach).

Thirdly, offloading the before-mentioned content generation (conversion of uploaded Video sequences for multiple output formats) to the cloud would yield a solution that could, at the same time, be cost efficient and result in a scalable setup. The innovation purpose in ICoSOLE is to investigate the efficient encoding of multiple quality representations for adaptive Streaming via DASH using a cloud infrastructure. Note again that this encompasses more than traditional video sequences, as in ICoSOLE the content offered/consumed may consist of panoramic/omni-directional or even free-viewpoint video fragments. This variety of content will require additional server capacity (versus the case in which only traditional Video is considered); something that is easily accomplished with a dynamically instantiated Server infrastructure based on a cloud platform.

Precise synchronization of professional broadcast content and UGC using a second Screen device (Hybrid Delivery) is a novel and challenging task, especially in a live System combining both types of content. Artefacts associated with imprecise synchronization will present themselves most in terms of loss of lip-sync or a disparity between on-Screen sequences. These delivery networks (video on demand and generic IP based/Internet for second screen) are typically completely separated and exchanging timing information for subsecond synchronization is a non-trivial challenge.

Finally, the integration of different content types/views into a coherent user experience and providing access to this content through mobile devices is another challenging task within ICoSOLE. Especially the integration of this different views in the multimedia stream description (e.g. via MPEG-DASH MPDs) and the intuitive management thereof on the small screen of the mobile target device, i.e., smartphones and tablets, will be another key issue. Therefore, the challenge is to provide an intuitive client interaction on tablets and smartphones, using, for example, HTML5 technology or native smartphone apps. All this will be done under the precondition of a smooth content consumption, even of multiple streams, via adaptive delivery i.e., adaption to the bandwidth capabilities and user context (e.g. resolution, codec, etc.) of the target device.