Research On Motion Estimation In AVS Video Encoder

As an important pillar of the electronic information industry, thedevelopment of audio and video industry for the healthy development ofnational economy importance is self-evident. To develop its own technologystandards help ensure industrial development is not controlled by others. Tobetter meet the needs of industrial development and technological progress,digital audio and video codec technology national standard AVS standard haveemerged. AVS standards and core technology from the patent for our audio andvideo industry bigger and stronger to provide a standard guarantee. The secondpart of the AVS standard for high-definition video, satellite broadcasting,multimedia communications transmission, video transmission, networkstreaming moving image compression technology developed video codecstandard. AVS standard since its inception, through domestic research anddevelopment staff for their tireless perfect for AVS+video portion, as well asrepresentatives of the international advanced level AVS2has launched a newgeneration of standards, is providing a rare historical opportunity for thedevelopment of audio and video-related industries. This study focused on AVS video encoder motion estimation is the most time-consuming to carry outin-depth research and analysis, mainly to complete the work includes thefollowing three aspects:First of all, the motion estimation for AVS video coding using a variety ofclassic search algorithms, in-depth analysis and research, in the AVS encoderselection of diamond search algorithm search template, adding judge to reducethe number of search points. According to the characteristics of various types offast search algorithm in a typical video test sequence by repeated tests, theperformance comparison of various types at different test sequences for fastsearch algorithm.Second, we propose a macroblock mode based on the determination methodof range of motion video sequences. By the way, you can compare different sizerange of motion video sequences, giving contrast relative to other sequencesvideo sequences provide a reference for the choice of the optimal motionestimation search algorithm.Finally, the starting point to determine the search strategy by studying theAVS encoder, an adaptive search starting position and range of motion in videosequences based search algorithm choice. The search algorithm for differentvideo sequences adaptively select a different search models. Experimentalresults show that compared with the fixed search algorithm used AVS encoder,this paper based on initial search point and range of motion in video sequencesadaptively select the search algorithm SNR essentially unchanged under the premise rate considerably, to some extent, improve the search speed.