Video editing apparatus and method

摘要

Picture quality loss is minimized and buffer underflow and overflow are prevented when extracting and recombining a plurality of consecutive scenes from an MPEG stream. The smallest number of pictures (equivalent at most to 1 second) that must be re-encoded so that the scene can be independently reproduced is detected to identify the re-encoding target picture. Then the picture type (I, P, B) after the pictures are re-encoded is determined. Buffer fullness is calculated, and the code allocation to the re-encoding range is calculated from buffer fullness and target picture count. The code allocation to each target picture is then calculated, and the target pictures are re-encoded and connected. <IMAGE>