Merge pull request aldebaro#5 from AiltonOliveir/main

WahWah

python_code/effect_wahwah.py

@@ -1,12 +1,78 @@
 '''
 Demonstrates specific audio effect using digital signal 
 processing (DSP).
-This effect is called XXX
-References:
-Author: 
-Date: 
+This effect is called Wah Wah
+References: https://ses.library.usyd.edu.au/bitstream/handle/2123/10578/Marion%2C%20Bruno%20-%20Wah%20Wah.pdf
+Author: Ailton Oliveira
+Date: 27/03/2022
 '''
+import scipy.io.wavfile as wav
+import numpy as np
+import matplotlib.pyplot as plt
+from sklearn.preprocessing import normalize
+
+def wahwah(signal,fs,dp_factor=0.05,width=2000,min_cutoff=250,max_cutoff=2000):
+    #Function to return a Wah Wah effect on a signal
+    #signal - Input signal that will be modulated (Array);
+    #fs - Sampling frequency (Float - Hertz)
+    #dp_factor - Damping factor (Float) - dp_factor < 0.5
+    #width - effect width (float)
+    #min_cutoff - Lower cut-off frequency (Float - Hertz)
+    #max_cutoff - Upper cut-off frequency (Float - Hertz)
+
+    center_freq = width/fs
+    cutoff_freq=list(np.arange(min_cutoff,max_cutoff,center_freq))
+    while(len(cutoff_freq) < len(signal)):
+        #Add noises until the effect have the same (or bigger) signal size 
+        cutoff_freq.extend(np.arange(max_cutoff,min_cutoff,-center_freq)) #Descending
+        cutoff_freq.extend(np.arange(min_cutoff,max_cutoff,center_freq)) #Ascending
+    cutoff_freq = cutoff_freq[0:len(signal)] #Match the signal size
+    cutoff_freq =  np.array(cutoff_freq)#Convert list to array
+
+    #Filtering coefficients
+    F1 = 2*np.sin((np.pi*cutoff_freq[0])/fs)
+    Q1 = 2*dp_factor
+
+    highpass=np.zeros(len(signal))
+    bandpass=np.zeros(len(signal))
+    lowpass=np.zeros(len(signal))
+
+    #WahWah differential equations
+    for n in range(0,len(signal)):
+        highpass[n] = signal[n] - lowpass[n-1] - Q1*bandpass[n-1]
+        bandpass[n] = F1*highpass[n] + bandpass[n-1]
+        lowpass[n] = F1*bandpass[n] + lowpass[n-1]
+        F1 = 2*np.sin((np.pi*cutoff_freq[n])/fs)
+
+    #Normalized Audio
+    Wahsignal  = bandpass/np.amax(abs(bandpass))
+    return Wahsignal
+
+if __name__ == '__main__':
+    ##Isolated test##
+
+    tst_file = '../test_wav_files/sample-16bits.wav'
+    fs,signal=wav.read(tst_file)
+    normed = wahwah(signal,fs)
+    wav.write("wah_wah_example.wav", fs, normed)
+    plot_signal= normalize(signal.reshape(-1, 1), norm='max',axis=0)
+    _ =input('Want plot the signals ? (y/n)')
+    if _ in ['yes','y','sim',True]:
+        #Plots
+        t1 = np.arange(0,(len(signal)/fs),1/fs)
+        plt.plot(t1,plot_signal,label='original')
+        plt.title('Original Audio')
+        plt.ylabel('Amplitude')
+        plt.xlabel('Length (in seconds)')
+
+        plt.plot(t1,normed.reshape(-1, 1),label='wah wah')
+        plt.title('Wah Wahed Audio')
+        plt.ylabel('Amplitude')
+        plt.xlabel('Length (in seconds)')
+        plt.legend()
+        plt.show()
+    else:
+        print('Finish test')
+
+
 
-def wahwah():
-    #TO-DO
-    return None