Acceleration of HEVC transform and inverse transform on ARM NEON platform

HEVC (High Efficiency Video Coding) has recently been published as the next generation video coding standard. Compared with previous standards, the coding efficiency is greatly improved at the cost of much higher codec complexity. On the other hand, ARM with SIMD (Single Instruction Multiple Data) instructions is widely deployed on mobile platform, which makes it feasible to accelerate multimedia applications by software. However, the accelerations are not straight forward, and the modules have to be carefully redesigned to utilize SIMD instructions efficiently. Transform and inverse transform modules are frequently called in HEVC codec. In this paper, we propose an acceleration method of the transform and inverse transform by utilizing the data level parallelism on NEON platform. Experiment results show that the speed of these modules has been increased considerably.