Lossy joint source channel coding for multicast and multiple description applications.

摘要

The broad objective of this thesis is to tackle the problems of delay intolerant multimedia transmission over wireless networks. Robust, finite block length code design that can operate at different rates to satisfy heterogenous node connections and sources is of interest.;In the first part of this thesis, we propose a joint source-channel coding scheme that combines the advantages and simplicity of standard transform coding with the robustness of modern linear codes. The idea is to combine entropy coding and channel coding into a single linear encoding stage. If the channel is symmetric, the scheme can asymptotically achieve the optimal rate-distortion limit. However, its advantages appear even more clearly in the practical case of finite coding delay and complexity. Besides its effective performance for finite block length, it is also shown that the proposed joint source-channel coding architecture can be easily incorporated into existing networks, while maintaining the advantages and the flexibility of the separated paradigm. To this end we consider multicast and multiple description scenarios.;Multimedia services such as audio and video streaming represent one of the most important Internet applications. Beyond many possible architectures, the fundamental problem consists of multicasting a source to several users subject to some desired fidelity levels. In this thesis this problem is addressed from a source-channel coding viewpoint where a number of relevant system optimization problems are formulated.;Next, the multiple description coding over bit error channels is considered. A novel practical multiple description coding scheme is proposed based on scalar embedded quantizers and rateless codes. A peer-to-peer network model is studied where users can receive different number of descriptions and an arbitrary large number of descriptions can be created and transmitted between peers. For this network model convex system optimization problems are defined and solved.;Moreover, the proposed joint source channel coding scheme is applied to the important case of lossy transmission of digital images. A new coding scheme for image transmission over noisy channel is proposed. Similar to standard image compression, the scheme includes a linear transform followed by embedded scalar quantization. Joint source-channel coding is implemented by optimizing the rate allocation across the source subbands, treated as the components of a parallel source model. Due to the quantizer selection, non-binary code design for the linear encoder is established.;Finally, we investigate the effect of block length in Unequal Error Protection schemes. In these schemes, the source and channel coding parts are separated where the file transmission is done progressively using finite length coding blocks which are not necessarily of the same length. Dynamic programs are established for optimal rate allocations in multicast scenarios.