Professional Documents
Culture Documents
The Levinson algorithm is a fast recursion to solve the Yule-Walker prediction-error equations that model a random process, or more in general to implicitly invert a Toeplitz matrix. It gives rise to a computational structure (lattice lter) with useful properties. It can be used to analyze an autocorrelation sequence, and to generate a process with this autocorrelation.
It is used for parametric spectrum estimation (based on all-pole or AR modeling), and e.g. speech coding in GSM: instead of transmitting speech samples, the lter coefcients are transmitted so that the receiver can reconstruct the speech.
(These slides: real-valued signals. See book for the complex case.)
October 2, 2011
is uncorrelated to
, etc:
October 2, 2011
is given by
Parametric spectrum estimation is: given the correlation sequence nd . For this we need to nd
and
of
October 2, 2011
as a linear
past samples
and
).
October 2, 2011
model order , we try to nd the best tting parameters minimize the residual noise variance
If the model holds, we can easily derive the Yule-Walker equations, as follows. Multiply by
will not precisely satisfy such a model. For any that will
For
,
October 2, 2011
Yule-Walker equations
For
for
October 2, 2011
Yule-Walker equations
To solve the Yule-Walker equations is straightforward: the lter coefcients are . The noise power (prediction error or innovation power) . multiplications. The Levinson algorithm can
follows from
October 2, 2011
Levinson algorithm
We will derive a recursion, for coefcients For order For order
(
of size
:
. . .
..
. .
. . .
..
.
..
..
. . .
. . .
efciently? 8
the levinson algorithm October 2, 2011
Levinson algorithm
Take as trial solution for order . . .
..
.
..
This is not the correct solution of the YW equations, because in general We need to modify the trial solution such that
October 2, 2011
. . .
..
. .
. . .
..
.
..
..
. . .
. . .
. . .
. . .
10
October 2, 2011
Levinson algorithm
For
. ..
..
. .
..
.
. . .
. . .
..
.
.. ..
. .
..
. . .
. . .
Both trial solutions are not correct solutions of the YW equations of order because
. But we can take linear combinations of the two trial solutions in becomes 0.
11
October 2, 2011
matrix
..
. .
. . .
. . .
. . .
..
.
..
..
. . .
. . .
. . .
. . .
. . .
For
, this will bring the RHS into the required form. The LHS is then the .( is called a reection coefcient.)
the levinson algorithm October 2, 2011
12
. . .
. . .
. We obtain as solution
. . .
This is the update step in the Levinson recursion. The recursion is initiated by
With this, we enter the recursion. In the next step, we obtain the equations
. . .
where
. We set
The th step in the recursion has complexity . Overall, the complexity of solving
13
In particular,
In particular,
. . .
. . .
..
. .
. . .
..
.
..
..
. . .
. Since
. Hence
exists and is also strictly positive denite. This implies for any .
.
the levinson algorithm October 2, 2011
so that
and
be exactly predicted from its past, the process is called deterministic. This can occur only if is singular (thus the matrix is not strictly positive denite). 15
the levinson algorithm October 2, 2011
, so that the
, then
can
with initialization (
) as
16
October 2, 2011
Thus, the resulting system has impulse response If we use this system with input see Slide 4.
(and also
, not used). ,
the levinson algorithm
October 2, 2011
18
October 2, 2011
an IIR lter, the lter structure is recursive, i.e., the lter coefcients of
explicitly computed. It can be shown that the lter is stable as long as all reection coefcients are strictly smaller than 1. This is the case if the original If we replace
sulting output signal is a random process with the same correlation sequence as .
from
. This is
are not
19
October 2, 2011
At the transmitter, the speech signal is split in short frames, each of 20 ms. For each frame, the correlation sequence
Levinson algorithm, the reection coefcients and residual noise power estimated, coded and transmitted to the receiver.
It is not the same as the original speech signal, but has the same correlation sequence, which is good enough for the ear (at least for unvoiced speech).
20
October 2, 2011