A study on hybrid motion video coding schemes– block size, comparison of dct and vq schemes, loop filter

To design a highly efficient video codec, comparisons of available techniques from various viewpoints are necessary so that the coding scheme and parameters are chosen appropriately. In recently presented papers and equipment, many systems use motion‐compensated interframe prediction and DCT or VQ schemes for the prediction error coding. A loop filter often is adopted to improve coding efficiency. This paper compares DCT and VQ schemes in an interframe coding scheme, on an optimal blocksize for interfreme prediction error signal coding and on the effect of loop filtering from the viewpoints of coding efficiency and reproduced image quality. Simulations using actual image sequences show that the optimal block size is 4 × 4 or 8 × 8, that the loop filter improves significantly the coding efficiency and reproduced image quality, and that DCT and VQ schemes give nearly the same efficiency with an appropriate loop filter.