This paper addresses to the analysis new capabilities in parallel execution of Quarter-pixel Interpolation for Luminance data of H.264/AVC onto concurrent processors system-on-chip. We first exploit a parallelism of the Quarter-pixel interpolation spe...
This paper addresses to the analysis new capabilities in parallel execution of Quarter-pixel Interpolation for Luminance data of H.264/AVC onto concurrent processors system-on-chip. We first exploit a parallelism of the Quarter-pixel interpolation specifically to decrease its execution time. Next, we analyze the capability in parallel execution of Quarter-pixel interpolation onto concurrent processors of system-on-chip. By exploiting the data dependencies in calculation 6-tap FIR and Bi-linear filters, that are two principal filters in Interpolation block, for interpolated pixels of each 4x4 sub-macro block, we proposed a new solution to increase the execution speed by combining concurrent execution processing elements and three-stage pipelined scheduling. In addition, we also considered to the reused data and multiplierless in efficient implementation two kind of filters. The proposed solution using parallel techniques in terms of concurrent processing elements and pipelined scheduling can accelerate the speedup of realizing Quarter-pixel Interpolation in H.264/AVC decoder.