H.264/MPEG-4 AVC is expected to cover wide range of applications not only for smaller images like CIF/QCIF but also for SDTV/HDTV. As encoding/decoding of larger images requires more number of macroblocks to be processed, any kind of technique for speeding up is indispensable for real-time applications. Pipeline processing is an effective tool for hardware implementation. However, when we apply it to intra prediction, additional constraint will be necessary, which causes a severe loss in coding efficiency. In this paper we propose modifications on intra prediction that enable us to apply pipeline processing with an ignorable level of loss in coding efficiency.
Joint Video Team (JVT) was established to standardize next video coding technology. ITU-T H.26L (TML8) was adopted as the start point of the standard, and Joint Model (JM) was released. High quality video, not only for professional equipments but also for consumer applications, is very important target for JVT codec. However, H.26L was tested relatively lower bit rate. It is found that there are several problems due to quantization in order to support high quality video. In this paper, two extensions of quantization to support high quality video, which are frequency weighting matrix and extension of quantization table, are proposed. The frequency weighting matrix adds flexibility to the transform and quantization process by providing substantial coding efficiency without sacrificing the visual quality. It is possible to change quantization characteristic for each frequency depending on the sensitivity of human visual system. Encoder can compress video more efficiently. By extending the quantization table, encoder can use smaller and larger q_scale. In order to confirm the advantage of the proposal, simulations were performed to evaluate both PSNR and subjective quality. It was verified that proposed method is able to reduce bit rate keeping the same visual quality.
We have developed a Motion Vector (MV) Synthesis algorithm for an MPEG2-to-MPEG4 transcoder. MPEG2 bitstream parameters are utilized to adaptively scale and refine the MPEG2 MVs to synthesize MPEG4 MVs. The simulation results show that our MV Synthesis algorithm produces results that are equivalent to VM Full Search in both coding efficiency and subjective image quality, with significant reduction in complexity.