Research Of Error Resilience Application Of SP Frame In H.264

Video communication is one of the most important applications in multime dia communication, but it is impossible to avoid transmission error as interrupti on and delay always happen in channel. Since highly compressed video inform ation is very sensitive to interference, error resilience becomes significant in vi deo transmission. In order to meet the dual demands of error resilient and codi ng efficiency,a new frame type-SP frame is proposed in H.264.SP frame not on ly can achieve the purpose of error resilient,and its coding efficiency is better than I frame.However, how many SP frame should be inserted in a sequence?I n response to this question,in this paper,a optimized error-resilient program of SP frame is proposed, the program give the relationship of the rate of SP fram e, QP of SP frame and QP of P frame.The program is based on a rate-distorti on model,under limited distortion,seeking the smallest bit-rate.The major works finished are:Make an in-depth study of new feature and the error resilience tools in H.264, such as Intra prediction, Inter prediction, entropy coding, flexible macrobl ock order (FMO), parameter sets, data partition, and so on.The codec and main application of SP frame is analysed.And through the experimental,the coding efficiency of SP-frame, P frame and I frame is compar ed.Then, the error resilience application of SP frame is analysed, and error resi Hence experimental is done.Based on the research of many correlated domestic and foreign work, thro ugh a rate-distortion model, under limited distortion and seeking the smallest bi t-rate,a optimized program is proposed. The optimized program give the relatio nship of the rate of SP frame, QP of SP frame and QP of P frame.Under different packet loss channels, the optimization program and the oth er two programs were compared.Test result is analyzed. According to test results,under certain bit-rate, in most cases,PSNR of the optimized program is higher than the other two cases, Especially at high bit-rate, the effects become more obvious.