| 摘要 |
An edit point of an earlier GOP1 is placed after an I picture I22. In addition, the edit point is not placed immediately after an I or P picture. Thus, the GOP1 is decoded. A stream placed after the edit point (a P picture and later pictures) is discarded. B pictures B26 and B27 placed after a P picture P25 are re-encoded with a forward moving vector Fw. The remaining pictures other than B pictures B26 and B27 are re-encoded with encoding information used in a decoding process. Since an edit point of a later GOP2 is not placed before the I picture I2, the GOP2 is decoded. A stream placed before the edit point (pictures B0 to P5) is discarded. The first P picture P8 is changed to an I picture. Thereafter, the re-encoding process is performed. The P picture P8 changed from the I picture is used as a predictive reference picture. With the P picture P8, a backward moving vector Bk is obtained. With only the moving vector Bk, the B pictures B6 and B7 placed before the P picture P8 are re-encoded. |
