Research On Joint Parallel Algorithm Of H.264Encoders And Decoders

Along with the development of the video encoding and decoding algorithms, using multi-core processors in video codec to achieve parallel speedup gradually becomes a hot spot of international video codec research. At present scholars mainly research parallel encoding algorithms or decoder algorithms separately, we haven’t found the joint method to optimize the speed of encoders and decoders.The paper’s purpose is to optimize the sum of encoding time and decoding time, and improve the efficiency of real-time codec. Which can save computational resources, reduce energy consumption and has great significance in video coding area. This paper compares the optimal slice partition of encoders and decoders in same hardware conditions based on H.264encoder x264and decoder FFmpeg. The experimental results show that the encoder and decoder reach their optimal point at different slice partition. So the joint optimal slice partition for H.264video encoding and decoding is workable.This paper puts forward a joint parallel model and three decoder optimization models aiming at different hardware conditons. Decoding time is transferred to encoder through feedback channel, and the system exists communication delay. We design and realize the adaptive slice partition algorithm for different models based on traditional adapative slice partition algorithm. The experimental results show that the algorithm based on joint parallel model promotes speedup a lot and decreases system load unbalance compared with the algorithm based on encoder model and the algorithm based on decoder models.