if __name__ == '__main__':
import PS41mycode
import matplotlib.pyplot as plt
import psaudio
sigma = 0.001
if len(sys.argv)>1:
sigma = float(sys.argv[1])
fs = 8000
if len(sys.argv)>2:
fs = int(sys.argv[2])
audio = psaudio.AudioChannel(fs)
abstr = AbstractChannel(sigma)
(xau,fau,Rau) = PS41mycode.analyse(audio)
(xab,fab,Rab) = PS41mycode.analyse(abstr)
m = len(Rau)
plt.figure(1)
plt.subplot(211)
plt.plot(xau,fau,'b',xab,fab,'r')
plt.grid()
plt.xlabel('Noise PDF: sigma=%f'%sigma)
plt.subplot(212)
plt.plot(xrange(m),Rau,'b.-',xrange(m),Rab,'r.-')
plt.grid()
plt.xlabel('Noise auto-correlation')
plt.show()
def __init__(self, m, h):
self.m = m
self.h = h
def sendreceive(self, x):
return self.m+np.convolve(h,np.random.randn(x.size+20))
if __name__ == '__main__':
for i in xrange(5):
m = np.random.randn(1)
h = np.random.randn(20)
a = A(m,h)
T = random.choice([90000,99999,110000])
n = random.choice(xrange(90,110))
k = random.choice(xrange(5,15))
(x,fx,R) = PS41mycode.analyse(a,T,n,k)
assert type(x)==ndtype, 'x: not ndarray'
assert type(fx)==ndtype, 'fx: not ndarray'
assert type(R)==ndtype, 'R: not ndarray'
assert x.shape==(n,), 'x: wrong shape'
assert fx.shape==(n,), 'fx: wrong shape'
assert R.shape==(k,), 'R: wrong shape'
assert (fx>=0).all(), 'x: not non-negative'
M0 = sum(fx)*(x[1]-x[0])
assert abs(1.0-M0)<0.001, 'x,fx: not PDF samples'
print '\n------\n%f~1.0'%M0
M1 = sum(fx*x)*(x[1]-x[0])
assert abs(M1-m)<0.1, 'x,fx: the expected value does not match'
print '%f~%f'%(M1,m)
M2 = sum(fx*(x**2))*(x[1]-x[0])
v = (h**2).sum()