The Research Of Error Detection And Rate Control In Video Communication

In modern audio and video exchange processing, the key question of enhancingthe rate of transmitting the audio and video information and guaranteeing thetransmission quality of audio and video transmission quality is the compressionmethods and the technologies with which audio and video are encoded. Afterintroducing the basic theory of video coding, the development and technicalcharacteristics of two series of international video coding, we mainly introducechina's first audio and video coding standard(AVS), its key technologies and thedifferences of internal and international video standard. Then we discuss the errordetection and the rate control. Aiming at the defect of the traditional error detectionand the rate control, we proposed an even-odd check error detection method,aweighted model for linear regression to substitute the standard linear model on basicunit layer and a rate control algorithm on macroblock layer. The essence of thoseproposed algorithms is introduced as follows.One. An even-odd check error detection method: This method first to embedwatermark in the least significant bit of the scanned DCT coefficient in the encoder.Then, we check this exchanged coefficient of this position in the decoder. If thenumber of this position does not conform to the embedding rule, then we can judgethat this mackoblock is transmitted wrongly.Through confirmed on the test model, this method can enhance the errordetection performance greatly as taking the sacrifice of the lower PNSR, and thetransmission efficiency will not be affected even if we increase the extra expenses inthe transmitted bitstream. Two. An improved Linear forecast model: This method has considered the timeand the spatial relevance of the basic unit in the same position of the preceding frame,the preceding basic unit in the current frame and the current being coded basic unit.The experiment indicated that the rectified MAD is more close to the actual MADcompared to the original method and this proposed algorithm could improve thequality and deduce the rate shift.Three. An rate control algorithm on macroblock layer: we exploit MAD tomeasure the complexity of current macroblock and this could auto-adaptatively adjustthe QP of the current macroblock. We use the smaller QP to quantify the morecomplicated macroblock and use the bigger QP to quantify the less complicatedmacroblock. The proposed algorithm can improve the AVS standard algorithm by arate control algorithm on macroblock layer.