METHOD AND DEVICE FOR DECODING CODED DIGITAL VIDEO SIGNALS

摘要

Conventional MPEG decoders require about 16 Mbits of external memory to contain the input buffer, the two usual reference frames, and the B-information. As proposed earlier, about 2,8 Mbit in memory capacity can be saved by omitting the storage of B-information. According to the invention, a further exchange between memory capacity and internal processing, by MPEG decoding with only one reference frame stored in memory, is proposed: all B and P information required for the reconstruction of display information which cannot directly be retrieved from memory is decoded instantaneously. Three decoding sub-steps are provided in the proposed method and device: a first one for the macroblock parameters, a second one for the picture elements according to a decoding-on-the-fly principle, a third one for the updating of the earlier reference frame after the decoding operation of the later reference frame.