spectrum.yulear (Signal Processing Toolbox)

Yule-Walker spectrum object

Syntax

Hs = spectrum.yulear
Hs = spectrum.yulear(order)
Hs = spectrum.yulear(order,FFTLength)

Description

Hs = spectrum.yulear returns a default Yule-Walker spectrum object, Hs, that defines the parameters for the Yule-Walker spectral estimation algorithm. This method is also called the auto-correlation or windowed method. The Yule-Walker algorithm estimates the spectral content by fitting an auto-regressive (AR) linear prediction filter model of a given order to the signal. This leads to a set of Yule-Walker equations, which are solved using Levinson-Durbin recursion.

Hs = spectrum.yulear(order) returns a spectrum object, Hs, with the specified order and the FFTLength determined using NextPow2. The default value for order is 4.

Hs = spectrum.yulear(order,FFTLength) returns a spectrum object, Hs, with the specified order of the AR model and the specified way of determining the FFTLength. Valid values of the FFTLength string are as follows.

FFTLength string
Description

'InputLength'
Use the length of the input signal as the FFT length

'NextPow2'
Use the next power of 2 greater than the input signal length as the FFT length. This is the default value.

'UserDefined'
Use the FFT length provided as an input to the psd method or via a dspopts object. See dspopts for an example.

FFTLength string	Description
`'InputLength'`	Use the length of the input signal as the FFT length
`'NextPow2'`	Use the next power of 2 greater than the input signal length as the FFT length. This is the default value.
`'UserDefined'`	Use the FFT length provided as an input to the `psd` method or via a `dspopts` object. See `dspopts` for an example.

Note See pyulear for more information on the Yule-Walker algorithm.

Examples

Define a fourth order auto-regressive model and view its spectral content using the Yule-Walker algorithm.

randn('state',1);
x=randn(100,1);
x=filter(1,[1 1/2 1/3 1/4 1/5],x);%Fourth order AR filter
Hs=spectrum.yulear;               %Fourth order AR model
psd(Hs,x,'NFFT',512)

spectrum.welch sptool