| 摘要 |
The invention relates to a scalable video (de)coding method for wireless transmission of high definition television signals. Scalable means that the bitstream contains successively smaller quality refinements and that the bitstream can be truncated. The video images are divided in slices, and each slice is divided in blocks of 8×8 pixels. For each block, an optimal encoding method is chosen. Depending on whether the block is found to contain natural or synthetic image content, transform coding (DCT) or graphic coding is applied. Because the different encoding modes have different properties as regards picture quality, the bitstream format has to enable the encoder to very flexibly choose which bits to send first. The bitstream format in accordance with the invention consists of a multitude of scans (31-34) through the coded data of a series of individual blocks, e.g. a slice. In each scan, the encoder decides whether it will include data for natural blocks (BS1,BS2), for synthetic blocks (BS0,BS1), or both. For the DCT blocks, it can further choose between DC coefficient bits (22) and AC coefficient bits (23). For example, the encoder can choose to first send a few scans of DCT DC bits, then send a few scans of graphics bits and then send some DCT AC bits. The decisions that the encoder makes are signaled to the decoder in the form of flags that precede each block part (e.g. encoded bit plane) and/or plurality of blocks (e.g. a slice).
|
