| 摘要 |
<p>The invention relates to a method of removing blocking artefacts from a frame of a digital video signal in a video encoder (10), the video encoder comprising a motion compensated prediction loop (11, 12, 14, 15, 16, 17, 18, 19) in which a prediction (Pn(x,y)) for a subsequent frame of the digital video signal is formed, the motion compensated prediction loop comprising encoding (12), decoding (14) and motion compensated prediction (17, 18, 19) operations. According to the method of the invention, the frame is encoded by blocks, decoded and stored in a frame memory (16) for use as a reference frame (Rn(x,y)) in motion compensated prediction (17), a certain number (n) of pixels are selected for filtering from both sides of a block boundary (30) within the frame and the values of the pixels are corrected by applying a filtering operation (43) within (33, 34) the motion compensated prediction loop prior to use of the frame as a reference frame (Rn(x,y)) in motion compensated prediction (17) of a subsequent frame. The invention also relates to a corresponding method applied in a video decoder (20) and to a video encoder (10) and a video decoder (20) which apply the respective methods according to the invention. <IMAGE></p> |
