System and method for error-resilient video coding
Error resilient video coding schemes that may be employed at a transmitter or transcoder to limit error propagation at the receiver. Embodiments may include the use of Inhomogeneous Temporal Multiple Description Coding (ITMDC), cross-description error concealment, and cross-description reference picture reset (RPS) as well as homogeneous and inhomogeneous temporal/spatial MDC.
1. A video encoding method comprising:
sub-sampling input video, the input video comprising a plurality of sequential frames, to generate at least first and second video streams, each stream comprising different frames of the input video; and
encoding each of the at least first and second video streams to generate at least first and second video descriptions;
wherein the frames contained in at least one of the video streams comprises temporally inhomogeneous frames of the input video, wherein the temporal inhomogeneity is characterized by an interleave factor, k, wherein k is a number of consecutive frames of an original video signal between gaps in consecutive frames in a stream; and
wherein k is a function of quantization distortion.
2. The method of claim 1 further comprising:
receiving error feedback identifying lost video data in the first description; and
responsive to receipt of error feedback, using an error propagation reduction technique to reduce error propagation in the first description.
3. The method of claim 2 wherein the error propagation reduction technique comprises Intra/IDR refresh.
4. The method of claim 3 wherein the Intra/IDR refresh comprises:
encoding a frame as an I frame or IDR frame.
5. The method of claim 4 wherein the error propagation reduction technique comprises a reference picture selection (RPS) technique.
6. The method of claim 5 wherein the RPS technique comprises:
encoding a frame as a P frame using prediction from a frame prior to the lost frame.
7. The method of claim 6 wherein the RPS technique comprises:
encoding a frame as a P frame using prediction from a frame in the second description preceding the lost frame.
8. The method of claim 7 wherein the frame in the second description used for prediction of the P frame in the first description comprises the temporally nearest preceding frame in the second description that was confirmed via feedback as having been correctly received.
9. The method of claim 7 wherein the frame in the second description used for prediction of the P frame in the first description comprises a frame in the second description that is temporally nearest to the lost frame.
10. The method of claim 1 wherein k is varied over time by the encoder.
11. The method of claim 10 wherein k is varied as a function of a condition of a transmission channel of the video descriptions.
12. The method of claim 11 wherein k is varied as a function of packet loss rate (PLR).
13. The method of claim 1 wherein k is not equal for all of the descriptions.