CODEWITHSUNDEEP
pythonnumpysequence
Mathieu
Mathieu

Reputation: 5746

A better way to create a timeline of repeated event at a given frequency

I have a square (pulse) signal which has a frequency fq, and is played between t0 and tf. To create the timeline, I simply do:

import numpy as np

t0 = 0
tf = 200
fq = 20
timeline = np.round(np.arange(t0, tf, 1/fq*1000), 3)

A frequency of 20 Hz means a pulse is delivered every 50 ms. The timeline outputted is:

timeline
Out[284]: array([  0.,  50., 100., 150.])

Now, I would like the same think, but with a burst of n_pulse at frequency fq_burst repeated at the frequency fq.

Example: A burst of 4 pulses at 500 Hz repeated at 20 Hz should give me the timeline:

timeline
Out[282]: 
array([  0.,   2.,   4.,   6.,  50.,  52.,  54.,  56., 100., 102., 104.,
       106., 150., 152., 154., 156.])

My current solution works but is far from being nice and optimal:

start_burst = np.round(np.arange(t0, tf, 1/fq*1000), 3)
timing = list()
for t in start_burst:
    for n in range(n_pulse):
        timing.append(t + n*(1/fq_burst*1000))
timeline = np.round(np.asarray(timing), 3)

Any suggestion on how to improve this timeline generation? I'm thinking of a way to create directly the values in a numpy array and then to reshape it, but I can't figure out how to do it.

Thanks!

Upvotes: 2

Views: 238

Answers (1)

javidcf
javidcf

Reputation: 59681

A simple way:

import numpy as np

def burst_pulse(t0, tf, fq, n_pulse, fq_burst):
    timeline = np.arange(t0, tf, 1 / fq * 1000)
    burst = np.arange(n_pulse) / fq_burst * 1000
    timeline_burst = (timeline[:, np.newaxis] + burst[np.newaxis, :]).reshape((-1,))
    return np.round(timeline_burst, 3)

print(burst_pulse(0, 200, 20, 4, 500))

Output:

[   0.    2.    4.    6.   50.   52.   54.   56.  100.  102.  104.  106.
  150.  152.  154.  156.]

Upvotes: 2

Related Questions