Unequal error protection of H.264 Scalable Video Coding using discrete wavelet transfrom within a lognormal shadowing environment

Abstract

In this paper, Unequal Error Protection (UEP) is applied to a proposed Scalable Video Coding (SVC) amendment to the H.264/AVC standard. The proposed SVC framework uses Discrete Wavelet Transform (DWT) as a hierarchically signal decomposition for multi-resolution analysis and H.264/AVC as a source coding and compression technique. The output streams are divided into four streams with different priorities. UEP is applied in two different dimensions, the first dimension is defined under the video scalability and the second dimension gives higher priority to I-frames over P-frames. UEP is realized using a combination of Hierarchical Quadrature Amplitude Modulation (H-QAM) and Forward Error Correction (FEC) based on Rate Compatible Punctured Convolutional Codes (RCPC). A method to find the optimal selection of parameters for the H-QAM and the rate used by the RCPC is proposed to achieve a certain predefined quality of service based on the received SNR distribution due to lognormal shadowing within the network cell. Simulation results show that the proposed UEP scheme outperforms the Equal Error Protection (EEP) for low SNR values while minimizing as much as possible the loss for higher SNR users and thus achieving a better quality of service.