CODEWITHSUNDEEP
pythonnumpymatplotlibscipy
user2262504
user2262504

Reputation: 7297

how to generate rectangular impulse with numpy/scipy

I want to generate a rectangular impulse with python. I believe it can be done with numpy or scipy. But I cannot get it from API. After I generate the rectangular impulse, i will plot it with matplotlib.

Upvotes: 4

Views: 6099

Answers (2)

mins
mins

Reputation: 7514

You can also use the heaviside function (aka unit step function in engineering), e.g. numpy.heaviside. A pulse is created by the difference between two heaviside sequences:

n = np.arange(1-N, N)
n1, n2 = -3, 5 # pulse limits
pn = np.heaviside(n-n1, 1) - np.heaviside(n-n2, 1)

The heaviside function changes the values of the input array to 0, to a constant or to 1, depending on whether they are negative, null or positive. Here the input array contains the index values shifted by an offset corresponding to some index, and the constant is 1.

enter image description here

Full code

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.ticker as tck


# Sample numbers
N = 10
n = np.arange(1-N, N)

# Pulse limits
n1, n2 = -3, 5
pn = np.heaviside(n-n1, 1) - np.heaviside(n-n2, 1)

# Some signal clipped by pulse
xn = 3 * 0.8**n * pn

# Plot everything
fig, axes = plt.subplots(figsize=(6,6), nrows=4, sharex=True, layout='constrained')
axes = axes.flatten()

for ax, p, title in zip(axes[:2], [n1, n2], ['h1', 'h2']):
    ax.set_title(title)
    h = np.heaviside(n-p, 1)
    ax.stem(n, h)

for ax, s, title in zip(axes[2:], [pn, xn], ['pulse', 'signal clipped by pulse']):
    ax.set_title(title)
    ax.stem(n, s)

ax.xaxis.set_major_locator(tck.MultipleLocator())

Upvotes: 1

DarenW
DarenW

Reputation: 16906

To create a 1D array, all zeros but for a stretch of 1.0 values - a rectangular impulse:

import numpy as np
a = np.zeros( (1000,) )   # whatever size. initializes to zeros
a[150:180] = 1.0          # index range sets location, width of impulse

To see a plot:

import matplotlib.pyplot as mp
mp.plot(a)
mp.show()

Upvotes: 6

Related Questions