VAD

VAD/VAD01.py

@@ -0,0 +1,55 @@
+import wave
+import numpy as np
+import matplotlib.pyplot as plt
+import Volume as vp
+
+def findIndex(vol,thres):
+    l = len(vol)
+    ii = 0
+    index = np.zeros(4,dtype=np.int16)
+    for i in range(l-1):
+        if((vol[i]-thres)*(vol[i+1]-thres)<0):
+            index[ii]=i
+            ii = ii+1
+    return index[[0,-1]]
+
+fw = wave.open('../sounds/sunday.wav','r')
+params = fw.getparams()
+nchannels, sampwidth, framerate, nframes = params[:4]
+strData = fw.readframes(nframes)
+waveData = np.fromstring(strData, dtype=np.int16)
+waveData = waveData*1.0/max(abs(waveData))  # normalization
+fw.close()
+
+frameSize = 256
+overLap = 128
+vol = vp.calVolume(waveData,frameSize,overLap)
+threshold1 = max(vol)*0.10
+threshold2 = min(vol)*10.0
+threshold3 = max(vol)*0.05+min(vol)*5.0
+
+time = np.arange(0,nframes) * (1.0/framerate)
+vols = np.arange(0,len(vol)) * (nframes*1.0/len(vol)/framerate)
+index1 = findIndex(vol,threshold1)*(nframes*1.0/len(vol)/framerate)
+index2 = findIndex(vol,threshold2)*(nframes*1.0/len(vol)/framerate)
+index3 = findIndex(vol,threshold3)*(nframes*1.0/len(vol)/framerate)
+end = nframes * (1.0/framerate)
+
+plt.subplot(211)
+plt.title("VAD01 using volume")
+plt.plot(time,waveData,color="black")
+plt.plot([index1,index1],[-1,1],'-r')
+plt.plot([index2,index2],[-1,1],'-g')
+plt.plot([index3,index3],[-1,1],'-b')
+plt.ylabel('Amplitude')
+
+plt.subplot(212)
+plt.plot(vols,vol,color="black")
+plt.plot([0,end],[threshold1,threshold1],'-r', label="threshold 1")
+plt.plot([0,end],[threshold2,threshold2],'-g', label="threshold 2")
+plt.plot([0,end],[threshold3,threshold3],'-b', label="threshold 3")
+plt.legend()
+plt.ylabel('Volume(absSum)')
+plt.xlabel('time(seconds)')
+plt.savefig("VAD01")
+plt.show()

VAD/VAD02.py

@@ -0,0 +1,52 @@
+import wave
+import numpy as np
+import matplotlib.pyplot as plt
+import Volume as vp
+from ZeroCR import ZeroCR
+
+fw = wave.open('../sounds/sunday.wav','r')
+params = fw.getparams()
+nchannels, sampwidth, framerate, nframes = params[:4]
+strData = fw.readframes(nframes)
+waveData = np.fromstring(strData, dtype=np.int16)
+waveData = waveData*1.0/max(abs(waveData))  # normalization
+fw.close()
+
+frameSize = 256
+overLap = 128
+vol = vp.calVolume(waveData,frameSize,overLap)
+zcr = ZeroCR(waveData,frameSize,overLap)
+threshold1 = max(vol)*0.10
+threshold2 = min(vol)*10.0
+threshold3 = max(vol)*0.05+min(vol)*5.0
+threshold12 = max(zcr)*0.10
+threshold22 = min(zcr)*10.0
+threshold32 = max(zcr)*0.05+min(zcr)*5.0
+
+time = np.arange(0,nframes) * (1.0/framerate)
+vols = np.arange(0,len(vol)) * (nframes*1.0/len(vol)/framerate)
+zcrs = np.arange(0,len(zcr))*(nframes*1.0/len(zcr)/framerate)
+end = nframes * (1.0/framerate)
+plt.subplot(311)
+plt.title("VAD02 using volume and ZeroCR")
+plt.plot(time,waveData,color="black")
+plt.ylabel('Amplitude')
+
+plt.subplot(312)
+plt.plot(vols,vol,color="black")
+plt.plot([0,end],[threshold1,threshold1],'-r', label="threshold 1")
+plt.plot([0,end],[threshold2,threshold2],'-g', label="threshold 2")
+plt.plot([0,end],[threshold3,threshold3],'-b', label="threshold 3")
+plt.legend()
+plt.ylabel('Volume(absSum)')
+
+plt.subplot(313)
+plt.plot(zcrs,zcr,color="black")
+plt.plot([0,end],[threshold12,threshold12],'-r', label="threshold 12")
+plt.plot([0,end],[threshold22,threshold22],'-g', label="threshold 22")
+plt.plot([0,end],[threshold32,threshold32],'-b', label="threshold 32")
+plt.legend()
+plt.ylabel('Zero-Crossing Rate')
+plt.xlabel('time(seconds)')
+plt.savefig("VAD02")
+plt.show()

VAD/Volume.py

@@ -0,0 +1,27 @@
+import math
+import numpy as np
+
+# method 1: absSum
+def calVolume(waveData, frameSize, overLap):
+    wlen = len(waveData)
+    step = frameSize - overLap
+    frameNum = int(math.ceil(wlen*1.0/step))
+    volume = np.zeros((frameNum,1))
+    for i in range(frameNum):
+        curFrame = waveData[np.arange(i*step,min(i*step+frameSize,wlen))]
+        #curFrame = curFrame - np.median(curFrame) # False
+        curFrame = curFrame - np.mean(curFrame) # zero-justified
+        volume[i] = np.sum(np.abs(curFrame))
+    return volume
+
+# method 2: log10 of square sum
+def calVolumeDB(waveData, frameSize, overLap):
+    wlen = len(waveData)
+    step = frameSize - overLap
+    frameNum = int(math.ceil(wlen*1.0/step))
+    volume = np.zeros((frameNum,1))
+    for i in range(frameNum):
+        curFrame = waveData[np.arange(i*step,min(i*step+frameSize,wlen))]
+        curFrame = curFrame - np.mean(curFrame) # zero-justified
+        volume[i] = 10*np.log10(np.sum(curFrame*curFrame))
+    return volume

VAD/ZeroCR.py

@@ -0,0 +1,14 @@
+import math
+import numpy as np
+
+def ZeroCR(waveData,frameSize,overLap):
+    wlen = len(waveData)
+    step = frameSize - overLap
+    frameNum = int(math.ceil(wlen*1.0/step))
+    zcr = np.zeros((frameNum,1))
+    for i in range(frameNum):
+        curFrame = waveData[np.arange(i*step, min(i*step+frameSize,wlen))]
+        #To avoid DC bias, usually we need to perform mean subtraction on each frame
+        curFrame = curFrame - np.mean(curFrame) # zero-justified
+        zcr[i] = sum(curFrame[0:-1]*curFrame[1:]<=0)
+    return zcr