Font Size: a A A

Improvement Of HEVC Encoder Based On GPU

Posted on:2017-11-07Degree:MasterType:Thesis
Country:ChinaCandidate:X L ZhangFull Text:PDF
GTID:2348330518993460Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
With the development of the Internet,the life of people is changing.The application of video data is becoming more and more popular,also the requirement of related video encoder performance is becoming high.Due to the demand of high definition video and the emergence of ultra high-definition video,the traditional video coding standards have been difficult to adapt to the various requirements.In order to solve the new problems in the field of digital video technology,the video coding standard HEVC(High Efficiency Video Coding)is introduced.HEVC aims to reduce the half mount of bit rate on average and keep the coding performance.But the compression performance of HEVC is at the expense of coding complexity.The complexity of the algorithm making the current hardware platform is difficult to achieve real-time encoding.In the other side,the computing speed of CUDA based GPU achieves a high speed of development and makes significant achievements in many fields.Based on the present situation of the HEVC and the high computing performance of GPU,this paper aims to do a research on the improvement of HEVC based on parallel computing.The main content of this paper is to come up with some optimization methods for High Efficiency Video Coding Standard encoder based on the GPU parallel programming.In this paper,the development of video coding and an overview of the HEVC is introduced.The key technology of HEVC encoder is described,and a detail analysis is made on the part of the inter prediction of encoding.Then combined with the principle of GPU programming knowledge and the basic principle of HEVC encoder,the difficult points of GPU based HEVC encoder optimization methons are analyzed.In this paper,a CPU and GPU platform based HEVC encoder framework is designed.On this basis,two methods are proposed for the encoder.In addition,in order to speed up the processing of CPU and reduce the idle time of GPU,a fast algorithm for CU partition is proposed.The proposed algorithm is implemented in the official test model HM16.0,and the effectiveness of the proposed algorithm is demonstrated by comparing the experimental data.Finally,the overall optimization encoding framework is tested,the experimental results prove that the optimization of the HEVC encoder baesd on GPU can maintain the encoding performance and reduce the encoding time respectively.
Keywords/Search Tags:HEVC, video coding, GPU, CUDA, heterogeneous computing
PDF Full Text Request
Related items