| 摘要 |
A technique is proposed to produce coded video data with a high coding efficiency and a high picture quality. A plurality of coding units perform compression coding of a single source stream of video data by using different coding parameters. A control unit (300) controls those coding units in such a way that all the coding units produce respective interframe coded pictures with identical playback times as a first-coded picture in a picture group (e.g., GOP) and produce identical local decoded pictures based on the respective intraframe coded pictures. A selection decision unit (301) makes a selection, for each picture group, of one of the coded video data outputs from the coding units on the basis of their coding results. According to this selection, a selection unit (400) selectively outputs one of the coded video data outputs from the coding units. This output switching operation is performed at a point immediately before a forward interframe predictive coded picture that appears first of those in a picture group included in each of the coded video data outputs being switched. |
