def insert_file(fn, word): """ Read the signal in file fn, calculate the fingerprints and insert them into database. """ X, params = readWaveFile(fn, withParams=True) fingerprints = get_fingerprints(X, sr=params[2]) db = Database() db.insert(word, fingerprints) db.logout()
def recognize_file(fn):
"""
Recognize the given audio file fn.
"""
X, params = readWaveFile(fn, withParams=True)
fingerprints = get_fingerprints(X, sr=params[2])
db = Database()
word = db.find(fingerprints)
db.logout()
if word:
print(fn + " is about word \"" + word + "\"")
else:
print("Cannot recognize file " + fn + ".")
def getFingerPrint(path, start=None, end=None):
print("Creating fingerprint of ",path)
X = au.readWaveFile(path)
if start != None:
if end != None:
X = X[(SR*start):(SR*end)]
l = len(X)
i = 0
fp = []
while i <= (l - 4096):
P = realFFT(X[i:(i+WINDOW_SIZE)])
p = pickPeaks(P)
# print(p)
fp.append(sorted(p))
# au.graphSpectrum(X[i:(i+WINDOW_SIZE)])
i += SLIDE_INTERVAL
print("Generated", len(fp), "windows.")
return fp
import numpy as np
import audioUtilities as au
signal = au.readWaveFile("SteelString.wav")
# au.displaySignal(signal)
# print(len(signal))
amp_envelope = au.getEnvelope(signal, 2205, 2205)
# au.displaySignal(amp_envelope)
# # print(amp_envelope)
signal = signal[:44100]
# tx = [ 5 , 6, 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 ]
tx = signal
ta = [0.5, 0.7]
intermediateWindows = []
window = len(ta)
def zeroPadLeft(lst, m):
for j in range(m):
lst.insert(j, 0)
return lst
for i in range(0, len(tx) - 1, 2):
for j in range(len(ta)):
intermediateWindows.append(ta[j] * tx[i])
q = [
def getMaxAmp_Wave(filename):
wave = au.readWaveFile(filename)
maxAmp = max(wave)
return (maxAmp, wave)
if value < 0:
final = value + new_amp
else:
final = value - new_amp
return final
def compress_wav(filename, threshLimit, ratio):
(_sourceMax, _sourceWave) = getMaxAmp_Wave(filename)
threshold = int(_sourceMax / threshLimit)
_new = []
for elem in _sourceWave:
if abs(elem) > threshold:
# use decibel reduce function to find new amp add to list
diff = threshold - elem
dr = decibelReduce(elem, diff, ratio)
_new.append(dr)
else:
_new.append(elem)
return _new
test = au.readWaveFile("Bn.wav")
test_compress = compress_wav("Bn.wav", 2, 2)
#
au.displaySignal(test)
au.displaySignal(test_compress)
# au.writeWaveFile("BnNew2.wav", test_compress)
#
def main(file, displaySignal=False):
filename = file.split(".wav")[0]
wave = au.readWaveFile(file)
if(displaySignal):
displaySignals(wave)
writeAudioFilters(wave, filename)
def main(file, displaySignal=False):
filename = file.split(".wav")[0]
wave = au.readWaveFile(file)
if (displaySignal):
displaySignals(wave)
writeAudioFilters(wave, filename)
s1 = [np.float64(i.real) for i in s1]
s1 = np.asarray(s1)
s2 = np.fft.fft(hanning_window * a2)
s2 = [np.float64(i.real) for i in s2]
s2 = np.asarray(s2)
phase = (phase + np.angle(s2 / s1)) % 2 * np.pi
a2_rephased = np.fft.ifft(np.abs(s2) * np.exp(1j * phase))
a2_rephased = [np.float64(i.real) for i in a2_rephased]
# add to result
i2 = int(i / f)
result[i2:i2 + window_size] += hanning_window * a2_rephased
result = ((2**(16 - 4)) * result / result.max()) # normalize (16bit)
return result.astype('int16')
ninth = au.readWaveFile("Beethoven.Ninth.wav")
window_size = 2**13
h = 2**11
n = 10
factor = 2**(1.0 * n / 12.0)
test = stretch(ninth, 1 / factor, window_size, h)
# test = stretch(ninth, 1, 2048, 1024)
# slower = stretch(ninth, 0.5, 2048, 512)
# faster = stretch(ninth, 1.5, 2048, 512)
au.writeWaveFile("BeethovenNinth10.wav", test)
# au.writeWaveFile("BeethovenNinthHalf.wav", slower)
# au.writeWaveFile("BeethovenNinthOneAndAHalf.wav", faster)
import audioUtilities as au
# SR = 44100
# maxAmp = 2**15 -1
# def exponentialDecay(i,h):
# return 2**(-i/(SR*h))
# amps = [i/SR for i in range(SR)], [exponentialDecay(i,0.6) for i in range(SR)][1]
# bach = au.readWaveFile("Bach.wav")
# def reverb(amps, X):
# for i in range(len(amps)-1,len(X)):
# for j in range(len(amps)):
# out[i] += amps[j]*X[i-j]
# return out
# print(reverb(amps,bach))
signal = au.readWaveFile("BassGuitar.wav")
impulse = au.readWaveFile("ConcertHall.wav")
# print(len(impulse), len(signal))
# convolved = np.convolve(signal,impulse)
# print(convolved[0:100])
au.writeWaveFile("BassConcert.wav", np.convolve(signal,impulse,'valid'))
s1 = [np.float64(i.real) for i in s1]
s1 = np.asarray(s1)
s2 = np.fft.fft(hanning_window * a2)
s2 = [np.float64(i.real) for i in s2]
s2 = np.asarray(s2)
phase = (phase + np.angle(s2/s1)) % 2*np.pi
a2_rephased = np.fft.ifft(np.abs(s2)*np.exp(1j*phase))
a2_rephased = [np.float64(i.real) for i in a2_rephased]
# add to result
i2 = int(i/f)
result[i2 : i2 + window_size] += hanning_window*a2_rephased
result = ((2**(16-4)) * result/result.max()) # normalize (16bit)
return result.astype('int16')
ninth = au.readWaveFile("Beethoven.Ninth.wav")
window_size = 2**13
h = 2**11
n = 10
factor = 2**(1.0*n/12.0)
test = stretch(ninth, 1/factor, window_size, h)
# test = stretch(ninth, 1, 2048, 1024)
# slower = stretch(ninth, 0.5, 2048, 512)
# faster = stretch(ninth, 1.5, 2048, 512)
au.writeWaveFile("BeethovenNinth10.wav", test)
# au.writeWaveFile("BeethovenNinthHalf.wav", slower)
# au.writeWaveFile("BeethovenNinthOneAndAHalf.wav", faster)
import numpy as np
import audioUtilities as au
signal = au.readWaveFile("SteelString.wav")
# au.displaySignal(signal)
# print(len(signal))
amp_envelope = au.getEnvelope(signal, 2205, 2205)
# au.displaySignal(amp_envelope)
# # print(amp_envelope)
signal = signal[:44100]
# tx = [ 5 , 6, 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 ]
tx = signal
ta = [ 0.5 , 0.7 ]
intermediateWindows = []
window = len(ta)
def zeroPadLeft(lst,m):
for j in range(m):
lst.insert(j, 0)
return lst
for i in range(0, len(tx) - 1, 2):
for j in range(len(ta)):
intermediateWindows.append(ta[j] * tx[i])
q = [intermediateWindows[i:i+window] for i in range(0, len(intermediateWindows), window)]
z = []