Test Audio
# Needed imports
import numpy as np
from IPython.display import Audio
from scipy.io import wavfile
We don't really need that, but we can also import our own modules available in GitHub
In this case we will load the calc module to display graphs
Any Python module can be loaded this way
Modules can also be loaded in the virtual machine using the pip installer
# Set to True if we are in colaboratory of False elsewere
colaboratory = True
if colaboratory:
# Load the calc module
!rm calc.*
!wget https://raw.githubusercontent.com/R6500/Python-bits/master/Modules/calc.py
#Import the module
import calc
# Erase output after import
from IPython.display import clear_output
clear_output()
# Check loaded module
try:
print('calc version: ',calc.version)
except:
print('Error loading the calc module')
We can create the audio using code thanks to the numpy math facilities
# Create a sound
framerate = 44100
t = np.linspace(0,5,framerate*5)
data = np.sin(2*np.pi*220*t) + np.sin(2*np.pi*224*t)
The sound can be shown using, for instance, the loaded calc module.
It could also be shown using direct calls the matplotlib module
# Indicate the calc module if we are inside colaboratory
calc.setColaboratory(colaboratory)
# Show waveform
calc.plot11(t,data,"Sound Waveform","Time (s)","Value")
# We can zoom the start of the waveform
calc.plot11(t[0:10000],data[0:10000],"Sound Waveform (Detail)","Time (s)","Value")
The following code plays the sound data in mono
# Generate a player for mono sound
Audio(data,rate=framerate)
Sound can be played automatically after the cell code ends by using the autoplay parameter
This is shown on the following example
# Take a 1s piece of the mono sound
data2 = data[0:framerate]
# Generate a player for mono sound
Audio(data2,rate=framerate,autoplay=True)
Audio data can also be generated in stereo
We will first generate and show the waveforms
# Base sound
def base(t):
return np.sin(2*np.pi*220*t)*np.exp(-t)
tLeft = 4*(t/4 - np.fix(t/4))
left = base(tLeft)
tRight = 4*((t+2)/4 - np.fix((t+2)/4))
right = base(tRight)
# Show sound
calc.plot1n(t,[left,right],"Stereo sound","Time(s)","Values",["Left","Right"])
Now we can play the sound data in stereo
# Can also do stereo or more channels
Audio([left, right],rate=framerate)
The sound data don't need to be generated from code, it can be loade from wav files
In the case of Jupyter notebooks they can be in your computer
In the case of Colaboratory notebooks they need to be available on internet
One way to make a file available on internet is using GitHub
if colaboratory:
# Load sound file from GitHub
!wget https://github.com/R6500/Python-bits/raw/master/Colaboratory/Sounds/Bicycle%20bell%203.wav
# Erase output after import
from IPython.display import clear_output
clear_output()
print("Bicycle bell 3.wav loaded")
else:
print("We are not in colaboratory")
print("The file shall already be in the computer")
Once the file is saved in the virtual machine, it can be read on a numpy array and decoded as sound information
# Load the file on an object
data = wavfile.read('Bicycle bell 3.wav')
# Separete the object elements
framerate = data[0]
sounddata = data[1]
time = np.arange(0,len(sounddata))/framerate
# Show information about the object
print('Sample rate:',framerate,'Hz')
print('Total time:',len(sounddata)/framerate,'s')
The loaded sound data can be shown on a graph, if needed
# Show the sound data
calc.plot11(time,sounddata,"Sound data loaded from wav file","Time (s)","Values")
And, of course, it can be played
# Generate a player for mono sound
Audio(sounddata,rate=framerate)