how to convert the x-axis accelerator signal into spectrogram in python?

Question

my dataset (patient No., time/millisecond, x, y, z, label)

1,15,70,39,-970,0
1,31,70,39,-970,0
1,46,60,49,-960,0
1,62,60,49,-960,0
1,78,50,39,-960,0
1,93,50,39,-960,0
.
.
.

i am trying to to use the spectrogam for x-axis signal in preprocessing stage to use it then as the input data for a machine learning model instead of using the original raw x-axis data

here is what i tried to do

import matplotlib.pyplot as plt
import numpy as np

dt = 0.0005
t = np.arange(0.0, 20.0, dt)

data = np.loadtxt("trainingdataset.txt", delimiter=",")
x = data[:]

NFFT = 1024       # the length of the windowing segments
Fs = int(1.0/dt)  # the sampling frequency

ax1 = plt.subplot(211)
plt.plot(x)
plt.subplot(212, sharex=ax1)
Pxx, freqs, bins, im = plt.specgram(x, NFFT=NFFT, Fs=Fs, noverlap=900)
plt.show()

it gets me the following error

Warning (from warnings module):
  File "C:\Users\hadeer.elziaat\AppData\Local\Programs\Python\Python36\lib\site-packages\matplotlib\axes\_axes.py", line 7221
    Z = 10. * np.log10(spec)
RuntimeWarning: divide by zero encountered in log10

does your code work? if so, isn't it what you expected? what do you expect — Diego Aguado, Dec 05 '17 at 21:01
if it work i expect to get a spectrogram column for each vector (row) — Hadeer El-Zayat, Dec 06 '17 at 07:18

score 0 · Answer 1 · answered Dec 05 '17 at 21:04

0

If x is your signal and you can assume that your sampling rate is the mean of time/millisecond, then probably you can use the librosa library to compute the mel-spectrogram using librosa.feature.melspectrogram, there's also other utils to compute signal related features.

answered Dec 05 '17 at 21:04

Diego Aguado

1,604
18
36

its not an audio signal, its an accelerometer sensor signal – Hadeer El-Zayat Dec 06 '17 at 07:20
That does not matter. If the accelerometer data is evenly sampled (or reasonably so), then you can use librosa to calculate on it. – Jon Nordby Jul 15 '18 at 18:12

how to convert the x-axis accelerator signal into spectrogram in python?

1 Answers1