Project Team Dionysos

Overall Objectives
Scientific Foundations
Application Domains
New Results
Contracts and Grants with Industry
Partnerships and Cooperations
PDF e-pub XML

Section: New Results

Scalable Video Coding (SVC) transmission over IP and Broadcast networks

Participants : Majd Ghareeb, Adlen Ksentini, CĂ©sar Viho, Yassine Hadjadj-Aoul.

One of the multimedia market trends is audiovisual service (TV or VoD) anywhere, at any time. To support such service, a Video Service Provider has to manage, store, and distribute content towards multiple kinds and scales of terminals, and over different and transient access technologies to reach the end user. To solve such issues, video scalability seems to be the most relevant solution. It encodes the video in multiple separated layers, which enable a large number of users with heterogeneous capability to view any desired video stream, at anytime, and from anywhere. One of the most well known scalable standards is the Scalable Video Coding (SVC) extension of H.264/MPEG-4 AVC video compression. Our researches in this topic are related to how to optimize and enhance SVC transmission over IP and broadcast networks.

With the aim at keeping a high perceived video quality using SVC, MultiPath Video Streaming (MPVS) over Video Distribution Network (VDN) comes as a promising solution to overcome the limitations of the classical single path and IP-level video streaming approaches. In [45] and [43] we proposed different approaches that couple the three SVC scalability modes (Spatial, Temporal, SNR), with the path diversity provided by VDN. Our method adapts to both the heterogeneity of end-users using the scalable video coding as well as to network bandwidth fluctuations by observing the changes of the available bandwidth over the multiple overlay paths, and updating the streaming strategy accordingly. In [44] we enhanced the precedent solutions by using the PSQA-SVC version [42] in order to get the end-user feedback in terms of QoE, which helps adapting the streaming strategy. In [48] we designed a new protocol optimizing the energy consumption when transmitting video streams. We propose to exploit the SVC coding to adapt dynamically the received video quality to the instantaneous wireless nodes’ characteristics. This is achieved through determining the number of the transmitted/received enhancements layers of an SVC video based on the wireless node context.

In [49] we proposed to support SVC over DVBT2 networks, by associating the layering architecture of both technologies in order to tackle users mobility. This association allows mobile receivers with good physical channels to decode all the SVC layers and benefit from high video quality. Meanwhile, users with poor channel conditions can at least decode the base layer and benefit from acceptable video quality. Further, we introduced a novel QoE-based adaptive mechanism for SVC layers decoding. The proposed approach selects dynamically the number of layers to decode, at the receiver side, so as to maximize the users' perceived quality. Thus, no feedbacks or signaling messages are needed to implement the proposed algorithm. This makes it compliant with unidirectional technologies such as DVB-T2.