| 摘要 |
<p>PURPOSE:To prevent the occurrence of the information derived from noise and jitter of a sampling pulse by discriminating more precisely a still area with respect to an input moving picture signal and applying the coarse quantization with respect to a predicted error signal in the discriminated still area. CONSTITUTION:An input moving picture signal is supplied to a frame memory 10, optimum prediction/deciding circuit 12 and substracts 11 and 13. Said circuit 12 selects one predicting function, which minimizes the predicted error, with the aid of the input moving picture signal and a picture signal supplied from the frame memory 10, and supplies it to a variable delay circuit 14, quantizing circuit 15 and anisometric encoder 19. The quantizing circuit 15 quantizes the predicted error supplied from the subtractr 13, where four kinds of quantization characteristics are used. Namely, a norm of an optimum predicting function from the optimum prediction deciding circuit 12 is calculated. After the block is decided to be in the moving area or in the still area, a quantization characteristic with a wide dead zone is selected when the said block is decided to be in the still area.</p> |
