Transcoding From H264 Bitstream To MPEG-2 Bitstream
MetadataShow full item record
H.264/AVC was recently developed by the JVT (Joint Video Team). This new standard fulfills significant coding efficiency, simple syntax specifications and seamless integration of video coding into all current protocols and multiplex architectures. The H.264 specification represents a significant advancement in the field of video coding technology by providing MPEG-2 comparable video quality at an average of half the required bandwidth. Since widespread use of H.264 is anticipated, many legacy systems including all Digital TVs and home receivers use MPEG-2. This leads to the need for an efficient architecture that significantly employs the lower cost of H.264 video and does not require a significant investment in additional video coding hardware. An H.264 to MPEG-2 transcoder with low computational complexity and comparable quality to an MPEG-2 only encoded decoded sequence is one such effort in the above direction. Algorithm: The proposed research divides the process of transcoding into two parts. The first part deals with the transcoding of I frames where the best method is complete encoding and decoding of the input H.264 bit stream and then re-encoding it in the MPEG-2 format. This is the best method is terms of complexity and quality too due to the advance features of H.264 which is not supported in MPEG-2, like H.264 performs intra prediction in the pixel domain and then applies a 4x4 integer transform across the residual. However in the case of MPEG-2, intra prediction is done in the transform domain and it basically involves the process of dividing every macroblock into 8x8 blocks and apply a 2D-DCT across them to get the transform coefficients. In the case of P frame transcoding, motion vectors are extracted from the H.264 decoding stage and then they are scaled and summed depending on the H.264 block sizes to get an MPEG-2 compatible motion vector for the entire 16x16 macroblock. The obtained motion vectors are then refined over a 2 pixel search window to get the best match and then reused in the MPEG-2 encoding stage. This method produces comparable quality transcoded output to simple MPEG-2 encoded decoded output with a significant saving in the motion estimation time at the MPEG-2 encoding stage.