def plot_pink_noise():
"""Makes a plot showing power spectrums for pink noise.
"""
thinkdsp.random_seed(20)
duration = 1.0
framerate = 512
def make_spectrum(signal):
wave = signal.make_wave(duration=duration, framerate=framerate)
spectrum = wave.make_spectrum()
spectrum.hs[0] = 0
return spectrum
signal = thinkdsp.UncorrelatedUniformNoise()
white = make_spectrum(signal)
signal = thinkdsp.PinkNoise()
pink = make_spectrum(signal)
signal = thinkdsp.BrownianNoise()
red = make_spectrum(signal)
linewidth = 2
# colorbrewer2.org 4-class sequential OrRd
white.plot_power(label='white', color='#fdcc8a', linewidth=linewidth)
pink.plot_power(label='pink', color='#fc8d59', linewidth=linewidth)
red.plot_power(label='red', color='#d7301f', linewidth=linewidth)
thinkplot.save(root='noise-triple',
xlabel='Frequency (Hz)',
ylabel='Power',
xscale='log',
yscale='log',
xlim=[1, red.fs[-1]])
def plot_pink_noise():
"""Makes a plot showing power spectrums for pink noise.
"""
thinkdsp.random_seed(20)
duration = 1.0
framerate = 512
signal = thinkdsp.UncorrelatedUniformNoise()
wave = signal.make_wave(duration=duration, framerate=framerate)
white = wave.make_spectrum()
signal = thinkdsp.PinkNoise()
wave = signal.make_wave(duration=duration, framerate=framerate)
pink = wave.make_spectrum()
signal = thinkdsp.BrownianNoise()
wave = signal.make_wave(duration=duration, framerate=framerate)
red = wave.make_spectrum()
linewidth = 1
white.plot_power(low=1, label='white', color='gray', linewidth=linewidth)
pink.plot_power(low=1, label='pink', color='pink', linewidth=linewidth)
red.plot_power(low=1, label='red', color='red', linewidth=linewidth)
thinkplot.save(root='noise-triple',
xlabel='frequency (Hz)',
ylabel='power',
xscale='log',
yscale='log',
axis=[1, 300, 1e-4, 1e5])
def plot_pink_autocorr(beta, label):
"""Makes a plot showing autocorrelation for pink noise.
beta: parameter of pink noise
label: string label for the plot
"""
signal = thinkdsp.PinkNoise(beta=beta)
wave = signal.make_wave(duration=1.0, framerate=11025)
lags, corrs = autocorr(wave)
thinkplot.plot(lags, corrs, label=label)
def main():
thinkdsp.random_seed(19)
signal = thinkdsp.BrownianNoise()
make_periodogram(signal)
return
signal = thinkdsp.UncorrelatedUniformNoise()
wave1 = signal.make_wave(duration=1.0, framerate=11025)
wave2 = signal.make_wave(duration=1.0, framerate=11025)
print wave1.cov(wave2)
print wave1.cov(wave1)
print wave2.cov(wave2)
print wave1.cov_mat(wave2)
print wave1.corr(wave2)
print wave1.corr(wave1)
print wave2.corr(wave2)
return
signal = thinkdsp.BrownianNoise()
process_noise(signal, root='red')
return
signal = thinkdsp.UncorrelatedUniformNoise()
process_noise(signal, root='white')
return
signal = thinkdsp.WhiteNoise()
white, integ_white = test_noise(signal, 'white-noise')
print white.estimate_slope()
signal = thinkdsp.BrownianNoise()
red, integ_red = test_noise(signal, 'red-noise')
print red.estimate_slope()
return
signal = thinkdsp.PinkNoise(beta=1.0)
pink, integ_pink = test_noise(signal, 'pink-noise')
print pink.estimate_slope()
thinkplot.preplot(num=3)
white.plot(low=1, exponent=2, label='white', linewidth=2)
pink.plot(low=1, exponent=2, label='pink', linewidth=2)
red.plot(low=1, exponent=2, label='red', linewidth=2)
thinkplot.show(xlabel='frequency (Hz)',
ylabel='power',
xscale='log',
yscale='log',
axis=[1, 18000, 1e-5, 1e8])
return
def main():
plot_gaussian_noise()
thinkdsp.random_seed(20)
signal = thinkdsp.UncorrelatedUniformNoise()
process_noise(signal, root='white')
thinkdsp.random_seed(20)
signal = thinkdsp.PinkNoise(beta=1.0)
process_noise(signal, root='pink')
thinkdsp.random_seed(17)
signal = thinkdsp.BrownianNoise()
process_noise(signal, root='red')
plot_pink_noise()
def plot_serial_corr():
"""Makes a plot showing serial correlation for pink noise.
"""
numpy.random.seed(19)
betas = numpy.linspace(0, 2, 21)
corrs = []
for beta in betas:
signal = thinkdsp.PinkNoise(beta=beta)
wave = signal.make_wave(duration=1.0, framerate=11025)
corr = serial_corr(wave)
corrs.append(corr)
thinkplot.plot(betas, corrs)
thinkplot.config(xlabel=r'pink noise parameter, $\beta$',
ylabel='serial correlation',
ylim=[0, 1.05])
thinkplot.save(root='autocorr3')
def plot_pink_autocorr(beta, label):
signal = thinkdsp.PinkNoise(beta=beta)
wave = signal.make_wave(duration=1.0, framerate=10000)
lags, corrs = autocorr(wave)
thinkplot.plot(lags, corrs, label=label)
N = len(wave)
y1 = wave.ys[lag:]
y2 = wave.ys[:N-lag]
corr = np.corrcoef(y1, y2, ddof=0)[0, 1]
return corr
signal = thinkdsp.UncorrelatedGaussianNoise()
wave = signal.make_wave(duration=0.5, framerate=11025)
serial_corr(wave)
signal = thinkdsp.BrownianNoise()
wave = signal.make_wave(duration=0.5, framerate=11025)
serial_corr(wave)
signal = thinkdsp.PinkNoise(beta=1)
wave = signal.make_wave(duration=0.5, framerate=11025)
serial_corr(wave)
np.random.seed(19)
betas = np.linspace(0, 2, 21)
corrs = []
for beta in betas:
signal = thinkdsp.PinkNoise(beta=beta)
wave = signal.make_wave(duration=1.0, framerate=11025)
corr = serial_corr(wave)
corrs.append(corr)
thinkplot.preplot(1)
wave = signal.make_wave(duration=1, framerate=11025)
wave.write('browian.wav')
wave.plot()
thinkplot.show()
spectrum = wave.make_spectrum()
spectrum.plot()
thinkplot.show()
spectrum.plot_power(linewidth=1, alpha=0.5)
thinkplot.config(xscale='log', yscale='log')
thinkplot.show()
print(spectrum.estimate_slope())
# Pink noise
signal = thinkdsp.PinkNoise()
wave = signal.make_wave(duration=1, framerate=11025)
wave.write('pink.wav')
wave.plot()
thinkplot.show()
spectrum = wave.make_spectrum()
spectrum.plot()
thinkplot.show()
spectrum.plot_power(linewidth=1, alpha=0.5)
thinkplot.config(xscale='log', yscale='log')
thinkplot.show()
print(spectrum.estimate_slope())
# White / UG
signal = thinkdsp.UncorrelatedGaussianNoise()
