signal = UncorrelatedGaussianNoise()
wave = signal.make_wave(duration=0.5, framerate=11025)
print serial_corr(wave)
# 布朗噪声中，各个值是前一个值与一个随机“步长”的和，我们预期会有强相关性
from thinkdsp import BrownianNoise
signal = BrownianNoise()
wave = signal.make_wave(duration=0.5, framerate=11025)
print serial_corr(wave)
# 粉噪声介于布朗噪声和UU噪声间，我们预期他会是中度相关性
from thinkdsp import PinkNoise
signal = PinkNoise(beta=1)
wave = signal.make_wave(duration=0.5, framerate=11025)
print serial_corr(wave)
# 粉噪参数为beta=0，为不相关噪声，逐步变化到beta=2，为布朗噪声，对
# 应序列相关性从0变化到非常接近于1.
np.random.seed(19)
betas = np.linspace(0, 2, 21)
corrs = []
for beta in betas:
signal = PinkNoise(beta=beta)
wave = signal.make_wave(duration=1.0, framerate=11025)
corr = serial_corr(wave)
def plot_pink_autocorr(beta, label):
signal = PinkNoise(beta=beta)
wave = signal.make_wave(duration=1.0, framerate=10000)
lags, corrs = autocorr(wave)
plt.plot(lags, corrs, label=label)
import sys
sys.path.insert(1, 'dsp-modulo')
import numpy as np
import thinkplot as plt
from thinkdsp import decorate
from thinkdsp import PinkNoise
pendientes = [0.0, 1.5, 2.0]
for pendiente in pendientes:
senal = PinkNoise(beta=pendiente)
wave = senal.make_wave(duration=0.5, framerate=22050)
#wave.write("ruido_rosa.wav")
#wave.plot()
#decorate(xlabel="Tiempo", ylabel= "Amplitud")
#plt.show()
etiqueta = f'pendiente= {pendiente}'
espectro = wave.make_spectrum()
espectro.plot_power(label=etiqueta, alpha=0.5)
loglog = dict(xscale="log", yscale="log")
decorate(xlabel="Frecuencia", ylabel="Poder", **loglog)
plt.show()