| 摘要 |
Accelerated video decoding makes use of FEC-repaired media packets that become available through FEC decoding later than their intended decoding time, so to re-establish the integrity of the prediction chain between predicted pictures. The decoder state may be stored at the time of reception of an erroneous packet or at the time of identification of a lost packet, and decoding continued. After FEC repair, the last known state of the decoder is restored after the lost/damaged packet(s) is (are) resurrected through FEC, and accelerated decoding accordingly is used. Cycles "reserved" for decoding of a sub-sequence may be utilized. By freezing the decoded frame at the begin of a sub-sequence and decoding coded pictures of the main sequence that are part of the previous FEC block the integrity of the main prediction chain may be established again. Alternatively, cycles from enhancement layer decoding may be used.
|
