Convert imshow spectrogram to image

Question

I would like to save just the wavelet image (no ticks nor labels) shown here to a png file.

I tried to follow the solution posted here for saving a spectrogram plot, but this approach is not working for me.

This is what I get:

This is the code that I have used:

import librosa
import librosa.display

import os
import pywt
import matplotlib.pyplot as plt
import soundfile as sf

import skimage.io

from tftb.generators import anasing
from mpl_toolkits.axes_grid1 import make_axes_locatable

import numpy as np
from ssqueezepy import cwt
from ssqueezepy.visuals import plot, imshow

# Set the path to the Binary Class dataset
fulldatasetpath = 'G:/AudioFile'

input_file = (r'G:/Audiofile.wav')


[data1, sample_rate1] = sf.read(input_file)
#[sample_rate1, data1] = wav.read(input_file);
duration = len(data1)/sample_rate1

time = np.arange(0, duration, 1/sample_rate1) #time vector

#%%##############            Take CWT & plot                ##################################################
Wx, scales = cwt(data1, 'morlet')

imshow(Wx,  abs=1)
plt.show()

Wx = abs(Wx)

#%%##############            SAVE TO IMAGE                ###########################################
def scale_minmax(X, min=0.0, max=1.0):
    X_std = (X - X.min()) / (X.max() - X.min())
    X_scaled = X_std * (max - min) + min
    return X_scaled

wave1 = np.log(Wx + 1e-9) # add small number to avoid log(0)

# min-max scale to fit inside 8-bit range
img = scale_minmax(Wx, 0, 255).astype(np.uint8)
img = np.flip(img, axis=0)  # put low frequencies at the bottom in image
img = 255-img # invert. make black==more energy
out = 'out.png'
# save as PNG
skimage.io.imsave(out, img)

what means `"not working for me"` ? Did you get error? Show it in question (not in comment) as text (not image). Did you get wrong result? Show it in question. — furas, Apr 27 '22 at 21:42
your data has almost 50_000 columns and only 300 rows - so it has to create very long image. if you want shorter then you would have to skip some columns - ie, `img = img[ : , ::100]` should reduce to 500 columns and image should be shorter. — furas, Apr 27 '22 at 22:04

score 1 · Accepted Answer · answered Apr 28 '22 at 08:38

You can set the position of the axis to cover the entire figure, and you can also play with figsize. For example:

import matplotlib.pyplot as plt
import numpy as np
from ssqueezepy import imshow

# test image
img = np.zeros((500, 40000, 3), dtype=int)
for i in range(img.shape[1]):
    img[:, i, 0] = int(abs(1 - 2 * i / img.shape[1]) * 255)

# create a figure and set the size
f = plt.figure(figsize=(8, 4))
# add a new axis into which ssqueezepy is going to plot
ax = f.add_subplot()
imshow(img)
# turn off tick labels
ax.axis(False)
# make the axis to cover the entire figure
ax.set_position([0, 0, 1, 1])

f.savefig("result.png")

Convert imshow spectrogram to image

1 Answers1