Abstract

In order to achieve high efficiency, multimedia source coding usually relies on the use of predictive coding. While more efficient, source coding based on predictive coding has been considered to be more sensitive to errors during communication. With the current volume and importance of multimedia communication, minimizing the overall distortion during communication over an error-prone channel is critical. In addition, for real-time scenarios, it is necessary to consider additional constraints such as fix and small delay for a given bit rate. To comply with these requirements, we seek an efficient joint source-channel coding scheme.

In this work, end-to-end distortion is studied for a first order autoregressive synthetic source that represents a general multimedia traffic. This study reveals that predictive coders achieve the same channel-induced distortion performance as memoryless codecs when applying optimal error concealment. We propose a joint source-channel system based on incremental redundancy that satisfies the fixed delay and error-prone channel constraints and combines DPCM as a source encoder and a rate-compatible punctured convolutional (RCPC) error control codec. To calculate the joint source-channel coding rate allocation that minimizes end-to-end distortion, we develop a Markov Decision Process (MDP) approach for delay constrained feedback Hybrid ARQ, and we use a Dynamic Programming (DP) technique. Our simulation results support the improvement in end-to-end distortion compared to a conventional Forward Error Control (FEC) approach with no feedback.

Library of Congress Subject Headings

Multimedia systems; Data compression (Telecommunication); Coding theory; Rate distortion theory

Publication Date

8-2016

Document Type

Thesis

Student Type

Graduate

Degree Name

Computer Engineering (MS)

Department, Program, or Center

Computer Engineering (KGCOE)

Advisor

Andres Kwasinski

Advisor/Committee Member

Sohail Dianat

Advisor/Committee Member

Shanchieh Jay Yang

Comments

Physical copy available from RIT's Wallace Library at QA76.575 .J37 2016

Campus

RIT – Main Campus

Share

COinS