Digital Signal Processing Reference
In-Depth Information
introduction of these error control and recovery mechanisms heavily
degrades the coding efficiency. In general, there are four types of data
losses associated with bit errors which occur in wireless video transmission
systems:
1. Least significant data loss - These are the type of errors which occur at less
important portions of the encoded segments. For example, errors in the
texture data of the video will not propagate in the temporal domain, so
they will only deteriorate the quality of the picture which they belong to.
Also, a bit error in texture data will not propagate spatially. Hence, this is
a localized error which causes the least significant data loss.
2. Prediction data loss - If bit errors take place in motion vectors, they will
result in prediction errors. There is a higher tendency for these prediction
errors to propagate in the temporal domain.
3. Data loss due to synchronization loss - Errors which occur in variable length
code words can sometimes cause the decoder to lose synchronization. In
such a case, the decoder is unable to detect the specific error location and
it discards even the error-free portion of the received data until it reaches
the next resynchronization marker. Therefore, this kind of bit error will
turn into a burst error at the decoder.
4. Loss of entire frame - Bit errors occurring in the header data create the
worst damage compared to the above-mentioned three types. When an
error occurs at header data, the decoder completely loses the track of the
encoder and in turn discards a whole video frame. Hence, this is the worst
type of error that can happen in video transmission.
4.2.2 ErrorResilienceTools
In a broad sense there are four types of error resilience tools associated with
the video encoding and decoding. Namely:
1. Localization techniques.
2. Data partitioning techniques.
3. Redundant coding techniques.
4. Concealment-driven coding techniques.
It should be noted that these tools are not mutually exclusive. The introduc-
tion of these techniques will increase the bit-rates of the encoded sequence
and the complexity of the codec, while decreasing the coding efficiency.
4.2.2.1 Localization
In this method the spatial and temporal dependency between the frames
or slices has been removed or reduced to stop further propagation of
the errors.
Search WWH ::




Custom Search