|
In telecommunications, the data are often split into information units
called packets, which are asynchronously transmitted. For several reasons,
some of the packets may be lost or subject to unnaceptable delays, and
error
concealment becomes a necessity. Considerable effort has been spend
studying
methods for dealing with the lost data. Mathematically, the missing packet
recovery is an interpolation problem, and there are several possible ways
of
looking at it and implementing interpolation methods that lead to improved
performance in the presence of packet errors. We discuss some iterative
and
noniterative techniques for achieving this goal. The need for interleaving
the data prior to transmission and its impact on the numerical stability of
the problem is detailed.
|