A thesis proposal submitted in partial fulfillment of the requirement for the degree of Master of Engineering, School of Engineering and Technology

Thesis (M.Eng.) - Asian Institute of Technology, 2003

The high demand for mobile and multimedia applications has emphasized the need for newer methods for the transmission of video over low bit-rate error-prone channels. H.264 video coding is an emerging standard, and provides a versatile video coding format suitable for different network and storage applications. The versatility is exposed through a set of encoder control parameters. In this thesis, a study of the effects of these parameters on the rate, distortion, and algorithm complexity is performed, and a suitable set of parameters determined, keeping the low bit-rate wireless channel in perspective. Rate Compatible Punctured Convolutional coding is used to produce code rates varying from 8/9 to 1/4 using the same encoder and the same Viterbi decoder. A Hybrid FEC/ ARQ protocol is developed to use the rate compatible codes in the transmission of video, which is especially vulnerable to transmission errors because of the high compression using predictive coding. Hybrid FEC/ ARQ codes combine the predictability of FEC and rate flexibility of ARQ, and are suitable for widely varying error conditions and delay sensitive applications. The proposed system is then used to perform joint source channel coding of video and study the effects over an AWGN channel over various Eb/No values. The results are used to develop a channel adaptive system for the transmission of video over low bit-rate channels. Frame-by-frame adaptation of the video to the channel conditions is simulated to demonstrate that channel adaptive joint source channel coding of H.264 video using hybrid FEC/ ARQ codes can utilize the available channel capacity more efficiently than a fixed coding based system when transmitting over low-bit rate error-prone channels.