Programmer Guide/Command Reference/EVAL/lpc

lpc

Compute the linear prediction coefficients using a Hamming window.

Usage:

lpc(x, m, p {, out, lfft})

Function:

Computes the linear prediction coefficients or the lpc smoothed spectrum of the signal vector x using the auto-correlation method.

Parameters:

x

An unwindowed signal vector.

m

The number of coefficients.

p

The pre-emphasis factor (0 <= p <= 1).

out

The output selector. See the result for details.

lfft

The fft length for smoothed spectrum.

Result:

A vector y. The type of vector depends on the parameter out:

0 - A smoothed linear amplitude spectrum where length = lfft/2+1 and y[0 .. lfft/2].

1 - The error energy and inverse filter coefficients where length = m+2 and y[0] = alpha and y[1 .. m+1] = AI[0..m].

2 - The error energy and reflection coefficients where length = m+1, y[0] = alpha and y[1..m] = RC[0..m-1].

3 - The error energy and area coefficients where length = m+1, y[0] = alpha and y[1..m] = AR[0..m-1].

4 - The error energy and log. area coefficients where length = m+1, y[0] = alpha and y[1..m] = sqrt(AR[0..m-1]).