Model classification rate control algorithmfor video coding

1 .INTRODUCTIONRate control is ani mportant issue althoughit does notbelong to the normative part in video coding stan-dard . Without rate control many video coding sys-tems can not run smoothly because the client buffermay often under-flow and over-flow when a channelused to deliver the compressed streamis of constantbandwidth. There are several rate control ways for aconstant bit rate videotransmissionsystem.One si m-ple solution is shown in MPEG-2 TM5[1],and it isusedto be a benchmar…

Model classification rate control algorithm for video coding

A model classification rate control method for video coding is proposed. The macro blocks are classified according to their prediction errors, and different parameters are used in the rate quantization and distortion quantization model. The different model parameters are calculated from the previous frame of the same type in the process of coding. These models are used to estimate the relations among rate, distortion and quantization of the current frame. Further steps, such as R D optimization based quantization adjustment and smoothing of quantization of adjacent macroblocks, are used to improve the quality. The results of the experiments prove that the technique is effective and can be realized easily. The method presented in the paper can be a good way for MPEG and H.264 rate control.