| 摘要 |
PURPOSE:To reduce the block distortion of amination reproduced pictures by linearly converting continuous plural frames or fields while using different conversion object blocks for the respective frames or the fields. CONSTITUTION:The picture buffer 2 of a picture compressing/reproducing means 1 is provided with a capacity capable of storing one frame with the raster data of MXN picture elements as input picture data. A block generation part 8 divides and outputs the picture data stored in the buffer 2 for the respective blocks of 8X8 picture elements and also inversely stores the reproduced picture data to the buffer 2. An orthogonal transformation device 3 executs DCT (discrete cosine transformation) and inverse DCT for the respective picture elements, a quantizer 4 quantizes transformed output corresponding to a quantization table and then, zigzag scanning and the inverse operation are performed. Also, a code/decoder 5 turns first-order quantized data to Huffman codes and performs the inverse operation. Thus, the discontinuous difficulty of watching due to the block distortion of the animation reproduced pictures is reduced. |
