LMS fit

test_lms.py

import numpy as np
import scipy.signal as sig
from adaptfilt import lms

if __name__ == '__main__':
    import matplotlib.pyplot as plt
    from scipy.io import wavfile

    sigma = 0.1
    order = 100

    fs, y = wavfile.read('shh.wav')
    y = y/2.0**15
    x = sigma*np.random.randn(len(y))

    # y_hat_lms is the vector of estimated samples computed
    # during the optimization process
    # e is the per-sample error
    # h is the estimated filter
    y_hat_lms, e, h = lms(x, y, order, 0.01)

    # y_hat is the "real" model output (noise filtered by the 
    # estimated filter)
    y_hat = sig.lfilter(h, 1, x)
   
    plt.subplot(311)
    plt.plot(y, 'b')
    plt.subplot(312)
    plt.plot(y_hat, 'g')
    plt.subplot(313)
    plt.plot(e, 'r')
    plt.show()