产品
结果展示
结果图
信号滤波处理对比
代码对比
上图中左边是matlab代码，右侧是python代码，同一个框内的实现的作用是一样的可以对比来看，在写法上只有些许差异，大致还是一样的。主要是因为python中调用的numpy库对矩阵的处理方式与matlab相似，画图库matplotlib也是参考matlab而实现的。
有差异的地方：
在矩阵/数组的处理上，
在for循环的处理上，
索引的处理上（matlab索引从1开始，python索引从0开始）
matlab代码
echo on
T=1;
Fs=2/T;
Ts=1/Fs;
c_opt=[-2.2 4.9 -3 4.9 -2.2];
t=-5*T:T/2:5*T;
x=1./(1+((2/T)*t).^2); % sampled pulse
equalized_x=filter(c_opt,1,[x 0 0]); % since there will be a delay of two samples at the output
% to take care of the delay
equalized_x=equalized_x(3:length(equalized_x));
% Now, let us downsample the equalizer output
for i=1:2:length(equalized_x),
downsampled_equalizer_output((i+1)/2)=equalized_x(i);
echo off;
end;
figure(1)
stem(equalized_x)
figure(2)
stem(downsampled_equalizer_output)python代码
import numpy as np
from scipy import signal
import matplotlib.pyplot as plt
T=1
Fs=2/T
Ts=1./Fs
c_opt=np.array([-2.2,4.9,-3,4.9,-2.2])
t=np.linspace(-5*T, 5*T, 21)
x=1./(1+(2/T*t)**2)
x_ = np.hstack((x,np.zeros((2,))))
print(c_opt,np.array([1]),x_)
equalized_x=signal.lfilter(c_opt,np.array([1]),x_);
equalized_x=equalized_x[2:equalized_x.size]
print(equalized_x)
downsampled_equalizer_output = np.zeros((int(equalized_x.size/2+1),))
#% Now, let us downsample the equalizer output
for i in range(0,equalized_x.size,2):
downsampled_equalizer_output[int((i+1)/2)]=equalized_x[i];
print(downsampled_equalizer_output)
plt.figure(1)
plt.stem(equalized_x)
plt.figure(2)
plt.stem(downsampled_equalizer_output)
plt.show()
155-2731-8020
