def computeSNR(inputFile, window, M, N, H):
"""
Input:
inputFile (string): wav file name including the path
window (string): analysis window type (choice of rectangular, triangular, hanning, hamming,
blackman, blackmanharris)
M (integer): analysis window length (odd positive integer)
N (integer): fft size (power of two, > M)
H (integer): hop size for the stft computation
Output:
The function should return a python tuple of both the SNR values (SNR1, SNR2)
SNR1 and SNR2 are floats.
"""
def energy(x):
e = np.sum(np.abs(x)**2)
return e
fs, x = UF.wavread(inputFile)
w = get_window(window, M, False)
mX, pX = stft.stftAnal(x, w, N, H)
y = stft.stftSynth(mX, pX, M, H)
n = x - y[:x.size]
n2 = x[w.size:-w.size] - y[:x.size][w.size:-w.size]
SNR1 = 10 * np.log10(energy(y) / energy(n))
SNR2 = 10 * np.log10(energy(y) / energy(n2))
return SNR1, SNR2
def computeSNR(inputFile, window, M, N, H):
"""
Input:
inputFile (string): wav file name including the path
window (string): analysis window type (choice of rectangular, triangular, hanning, hamming,
blackman, blackmanharris)
M (integer): analysis window length (odd positive integer)
N (integer): fft size (power of two, > M)
H (integer): hop size for the stft computation
Output:
The function should return a python tuple of both the SNR values (SNR1, SNR2)
SNR1 and SNR2 are floats.
"""
## your code here
w = get_window(window, M, False)
(fs, x) = UF.wavread(inputFile)
(mX, pX) = stft.stftAnal(x, w, N, H)
y = stft.stftSynth(mX, pX, M, H)
noise = x-y[:x.size]
# get energy of x
enX1 = np.sum(abs(x)*abs(x))
enNoise = np.sum(abs(noise)*abs(noise))
enX2 = np.sum(abs(x[M:-M])*abs(x[M:-M]))
enNoise2 = np.sum(abs(noise[M:-M])*abs(noise[M:-M]))
SNR1 = 10*np.log10(enX1/enNoise)
SNR2 = 10*np.log10(enX2/enNoise2)
return (SNR1, SNR2)
def computeSNR(inputFile, window, M, N, H):
"""
Input:
inputFile (string): wav file name including the path
window (string): analysis window type (choice of rectangular, triangular, hanning, hamming,
blackman, blackmanharris)
M (integer): analysis window length (odd positive integer)
N (integer): fft size (power of two, > M)
H (integer): hop size for the stft computation
Output:
The function should return a python tuple of both the SNR values (SNR1, SNR2)
SNR1 and SNR2 are floats.
"""
## your code here
(fs,x) = UF.wavread(inputFile)
w = get_window(window, M)
mX, pX = stft.stftAnal(x, w, N, H)
y = stft.stftSynth(mX, pX, M, H)
n1 = x - y[:x.size] # Get to where signal lies in y: test the dimension of y
n2 = x[M:-M] - y[:x.size][M:-M]
def calculate_energy(x):
e = np.sum(x**2)
return e
SNR1 = 10*np.log10(calculate_energy(x)/calculate_energy(n1))
SNR2 = 10*np.log10(calculate_energy(x)/calculate_energy(n2))
return (SNR1, SNR2)
def main(inputFile = '../../sounds/piano.wav', window = 'hamming', M = 1024, N = 1024, H = 512): """ analysis/synthesis using the STFT inputFile: input sound file (monophonic with sampling rate of 44100) window: analysis window type (choice of rectangular, hanning, hamming, blackman, blackmanharris) M: analysis window size N: fft size (power of two, bigger or equal than M) H: hop size (at least 1/2 of analysis window size to have good overlap-add) """ # read input sound (monophonic with sampling rate of 44100) fs, x = UF.wavread(inputFile) # compute analysis window w = get_window(window, M) # compute the magnitude and phase spectrogram mX, pX = STFT.stftAnal(x, fs, w, N, H) # perform the inverse stft y = STFT.stftSynth(mX, pX, M, H) # output sound file (monophonic with sampling rate of 44100) outputFile = 'output_sounds/' + os.path.basename(inputFile)[:-4] + '_stft.wav' # write the sound resulting from the inverse stft UF.wavwrite(y, fs, outputFile) return x, fs, mX, pX, y
def computeSNR(inputFile, window, M, N, H):
"""
Input:
inputFile (string): wav file name including the path
window (string): analysis window type (choice of rectangular, triangular, hanning, hamming,
blackman, blackmanharris)
M (integer): analysis window length (odd positive integer)
N (integer): fft size (power of two, > M)
H (integer): hop size for the stft computation
Output:
The function should return a python tuple of both the SNR values (SNR1, SNR2)
SNR1 and SNR2 are floats.
"""
## your code here
def energy(mag):
e = np.sum((10 ** (mag / 20)) ** 2)
return e
(fs, x) = UF.wavread(inputFile)
w = get_window(window, M)
mX, pX = STFT.stftAnal(x, fs, w, N, H)
y = STFT.stftSynth(mX, pX, M, H)
n = x - y[:x.size]
n2 = x[w.size:-w.size] - y[:x.size][w.size:-w.size]
mN, pN = STFT.stftAnal(n, fs, w, N, H)
mN2, pN2 = STFT.stftAnal(n2, fs, w, N, H)
snr1 = 10 * np.log10(energy(mX) / energy(mN))
snr2 = 10 * np.log10(energy(mX) / energy(mN2))
return snr1, snr2
def computeSNR(inputFile, window, M, N, H):
"""
Input:
inputFile (string): wav file name including the path
window (string): analysis window type (choice of rectangular, triangular, hanning, hamming,
blackman, blackmanharris)
M (integer): analysis window length (odd positive integer)
N (integer): fft size (power of two, > M)
H (integer): hop size for the stft computation
Output:
The function should return a python tuple of both the SNR values (SNR1, SNR2)
SNR1 and SNR2 are floats.
"""
## your code here
def energy(mag):
e = np.sum((10**(mag / 20))**2)
return e
(fs, x) = UF.wavread(inputFile)
w = get_window(window, M)
mX, pX = STFT.stftAnal(x, fs, w, N, H)
y = STFT.stftSynth(mX, pX, M, H)
n = x - y[:x.size]
n2 = x[w.size:-w.size] - y[:x.size][w.size:-w.size]
mN, pN = STFT.stftAnal(n, fs, w, N, H)
mN2, pN2 = STFT.stftAnal(n2, fs, w, N, H)
snr1 = 10 * np.log10(energy(mX) / energy(mN))
snr2 = 10 * np.log10(energy(mX) / energy(mN2))
return snr1, snr2
def computeSNR(inputFile, window, M, N, H):
"""
Input:
inputFile (string): wav file name including the path
window (string): analysis window type (choice of rectangular, triangular, hanning, hamming,
blackman, blackmanharris)
M (integer): analysis window length (odd positive integer)
N (integer): fft size (power of two, > M)
H (integer): hop size for the stft computation
Output:
The function should return a python tuple of both the SNR values (SNR1, SNR2)
SNR1 and SNR2 are floats.
"""
# Calculate the SNR after synthesis and analysis STFT
w = get_window(window, M)
# SNR (signal to noise ratio) = 10log10(Energy of signal / Energy of noise)
(fs, x) = UF.wavread(inputFile)
# Do analysis and synthesis
mX, pX = stft.stftAnal(x, w, N, H)
y = stft.stftSynth(mX, pX, M, H)
# Resizing y so we can calculate energy of noise
resized_y = y[:x.size]
# Calculating the noise of part 1 and 2
noise1 = x - resized_y
noise2 = x[w.size:-w.size] - resized_y[w.size:-w.size]
# Analyse both noises
mNoise1, pNoise1 = stft.stftAnal(noise1, w, N, H)
mNoise2, pNoise2 = stft.stftAnal(noise2, w, N, H)
energyInput = energy_computation(mX)
energyNoise1 = energy_computation(mNoise1)
energyNoise2 = energy_computation(mNoise2)
SNR1 = 10 * np.log10(energyInput / energyNoise1)
SNR2 = 10 * np.log10(energyInput / energyNoise2)
return SNR1, SNR2
def computeSNR(inputFile, window, M, N, H):
"""
Input:
inputFile (string): wav file name including the path
window (string): analysis window type (choice of rectangular, triangular, hanning, hamming,
blackman, blackmanharris)
M (integer): analysis window length (odd positive integer)
N (integer): fft size (power of two, > M)
H (integer): hop size for the stft computation
Output:
The function returns a python tuple of both the SNR values (SNR1, SNR2)
SNR1 and SNR2 are floats.
"""
w = get_window(window, M)
fs, x = wavread(inputFile)
mX, pX = stft.stftAnal(x, fs, w, N, H)
y = stft.stftSynth(mX, pX, M, H)
y = y[:len(x)]
noise = y - x
X = abs(fft(x))
Y = abs(fft(noise))
X[X < eps] = eps
Y[Y < eps] = eps
Es1 = En1 = 0
for k in range(0, len(x)):
Es1 = Es1 + X[k] * X[k]
En1 = En1 + Y[k] * Y[k]
SNR1 = 10 * np.log10(float(Es1) / En1)
x2 = x[M:len(x) - M]
y2 = y[M:len(y) - M]
noise2 = y2 - x2
X2 = abs(fft(x2))
Y2 = abs(fft(noise2))
X2[X2 < eps] = eps
Y2[Y2 < eps] = eps
Es2 = En2 = 0
for k in range(0, len(x2)):
Es2 = Es2 + X2[k] * X2[k]
En2 = En2 + Y2[k] * Y2[k]
SNR2 = 10 * np.log10(float(Es2) / En2)
return (SNR1, SNR2)
def computeSNR(inputFile, window, M, N, H):
"""
Input:
inputFile (string): wav file name including the path
window (string): analysis window type (choice of rectangular, triangular, hanning, hamming,
blackman, blackmanharris)
M (integer): analysis window length (odd positive integer)
N (integer): fft size (power of two, > M)
H (integer): hop size for the stft computation
Output:
The function should return a python tuple of both the SNR values (SNR1, SNR2)
SNR1 and SNR2 are floats.
"""
## your code here
fs, x = UF.wavread(inputFile)
#x1 = x
#x2 = x[M:-M]
w = get_window(window, M)
(mX, pX) = stft.stftAnal(x, fs, w, N, H)
y = stft.stftSynth(mX, pX, M, H)
noise = x - y[:x.size]
return (snr(x, noise), snr(x[M:-M], noise[M:-M]))
import time, os, sys
sys.path.append(os.path.join(os.path.dirname(os.path.realpath(__file__)), '../../../software/models/'))
import stft as STFT
import utilFunctions as UF
import matplotlib.pyplot as plt
from scipy.signal import hamming
(fs, x) = UF.wavread('../../../sounds/piano.wav')
w = np.hamming(1024)
N = 1024
H = 512
mX, pX = STFT.stftAnal(x, w, N, H)
y = STFT.stftSynth(mX, pX, w.size, H)
plt.figure(1, figsize=(9.5, 7))
plt.subplot(411)
plt.plot(np.arange(x.size)/float(fs), x, 'b')
plt.title('x (piano.wav)')
plt.axis([0,x.size/float(fs),min(x),max(x)])
plt.subplot(412)
numFrames = int(mX[:,0].size)
frmTime = H*np.arange(numFrames)/float(fs)
binFreq = np.arange(mX[0,:].size)*float(fs)/N
plt.pcolormesh(frmTime, binFreq, np.transpose(mX))
plt.title('mX, M=1024, N=1024, H=512')
plt.autoscale(tight=True)
def computeEngEnv(inputFile, window, M, N, H):
"""
Inputs:
inputFile (string): input sound file (monophonic with sampling rate of 44100)
window (string): analysis window type (choice of rectangular, triangular, hanning, hamming,
blackman, blackmanharris)
M (integer): analysis window size (odd positive integer)
N (integer): FFT size (power of 2, such that N > M)
H (integer): hop size for the stft computation
Output:
The function should return a numpy array engEnv with shape Kx2, K = Number of frames
containing energy envelop of the signal in decibles (dB) scale
engEnv[:,0]: Energy envelope in band 0 < f < 3000 Hz (in dB)
engEnv[:,1]: Energy envelope in band 3000 < f < 10000 Hz (in dB)
"""
### your code here
F1 = 3000
F2 = 10000
(fs, x) = UF.wavread(inputFile)
w = get_window(window, M)
mX, pX = stft.stftAnal(x, fs, w, N, H)
y = stft.stftSynth(mX, pX, M, H)
print "length x = " + str(x.size) + " Shape mX = "
np.shape(mX)
#print mX[0]
mX = np.power(10,(mX/20.0)) # de-convert from decibel
# Note that mX is an array of DFT frames corresponding to each frame in the entire signal.
# Each ROW of mX is a DFT of the particular frame. Toral numebr of rows is the total number of frames.
# Find number of bins in the region 0 - 3k, f = k*Fs/N
k0 = 0
k1 = np.ceil(F1*N/fs) # Bin number/index of first limit
k2 = np.ceil(F2*N/fs)
print "Bin 1 =" + str(k1) + " Bin 2 =" + str(k2)
R1 = [k0,k1]
R2 = [k1,k2]
numFrames = int(mX[:,0].size)
print "Number of Frames = " + str(numFrames)
# Find the energy in the spectrum f > 0 and <3k
e1 = np.zeros(numFrames)
e2 = np.zeros(numFrames)
engEnv = np.zeros((numFrames,2))
for i in range(0,numFrames):
e1[i] = np.sum( mX[i,k0+1:k1+1] * mX[i,k0+1:k1+1])
e2[i] = np.sum( mX[i,k1+1:k2+1] * mX[i,k1+1:k2+1])
# Convert back to dB
e1 = 10*np.log10(e1)
e2 = 10*np.log10(e2)
mX = 20*np.log10(mX)
engEnv[0:numFrames,0] = e1[:]
engEnv[0:numFrames,1] = e2[:]
# By looking at the plot you can see the frame numbers that have more <less 3k or 10k frequency content.
# And by knowign the frame number, you can tell the time location of the change in frequency
plt.plot(e1)
plt.plot(e2)
return(engEnv)
def residue_lpc(audio_inp, params,lpc_order):
"""
Obtains the LPC representation of the Residual Spectral(LPC envelope), and then generates the residual by IFFT'ing this representation with random phase.
Parameters
----------
audio_inp : np.array
Numpy array containing the audio signal, in the time domain
params : dict
Parameter dictionary for the sine model) containing the following keys
- fs : Sampling rate of the audio
- W : Window size(number of frames)
- N : FFT size(multiple of 2)
- H : Hop size
- t : Threshold for sinusoidal detection in dB
- maxnSines : Number of sinusoids to detect
lpc_order : integer
Number of coefficients in the LPC representation
Returns
-------
res_transformed : np.array
Returns the transformed residue(LPC envelope approximation) in the time domain
"""
fs = params['fs']
W = params['W']
N = params['N']
H = params['H']
t = params['t']
maxnSines = params['maxnSines']
w = windows.hann(W)
F,M,P,R = hprModelAnal(x = audio_inp, fs = fs, w = w, N = N, H = H, t = t, nH = maxnSines, minSineDur = 0.02, minf0 = 10, maxf0 = 400, f0et = 5, harmDevSlope = 0.01)
harmonics_recon = sineModelSynth(tfreq = F, tmag = M, tphase = P, N = W, H = H, fs = fs)
# Initializing an empty list to store the residual spectral approximations(LPC)
xmX = []
# Normalize the Residue before analysis(throws a np zero error otherwise)
nf = np.max(np.abs(R))
# nf = 1
# print(nf)
R = R/nf
for frame in ess.FrameGenerator(R.astype('float32'), W, H):
inp = np.pad(frame,[0,N - W],mode = 'constant',constant_values=(0, 0))
env_frame = fe.lpc_envelope(inp,lpc_order,fs,len(inp)//2 + 1)
xmX.append(env_frame)
xmX = np.array(xmX)
XpX = 2*np.pi*np.random.rand(xmX.shape[0],xmX.shape[1])
# xmX,XpX = stftAnal(audio_inp,w,N,H)
# Obtain the audio from the above representation
res_transformed = stftSynth(xmX, XpX, W, H)*nf
# ***Re-normalize the Residual so that it lies in the same range as the original residue***
# scale_init = np.max(np.abs(audio_inp))/np.max(np.abs(R))
# scale_final = np.max(np.abs(harmonics_recon))/scale_init
res_transformed = (res_transformed/np.max(np.abs(res_transformed)))
return res_transformed
import os
import sys
import numpy as np
import math
from scipy.signal import get_window
import matplotlib.pyplot as plt
sys.path.append(
os.path.join(os.path.dirname(os.path.realpath(__file__)),
'../../software/models/'))
import stft
import utilFunctions as UF
eps = np.finfo(float).eps
(fs, x) = UF.wavread(inputFile)
w = get_window(window, M)
mX, pX = stft.stftAnal(x, w, N, H)
y = stft.stftSynth(mX, pX, M, H)
sys.path.append(
os.path.join(os.path.dirname(os.path.realpath(__file__)),
'../../../software/models/'))
import stft as STFT
import utilFunctions as UF
import matplotlib.pyplot as plt
from scipy.signal import hamming
(fs, x) = UF.wavread('../../../sounds/piano.wav')
w = np.hamming(1024)
N = 1024
H = 512
mX, pX = STFT.stftAnal(x, fs, w, N, H)
y = STFT.stftSynth(mX, pX, w.size, H)
plt.figure(1, figsize=(9.5, 7))
plt.subplot(411)
plt.plot(np.arange(x.size) / float(fs), x, 'b')
plt.title('x (piano.wav)')
plt.axis([0, x.size / float(fs), min(x), max(x)])
plt.subplot(412)
numFrames = int(mX[:, 0].size)
frmTime = H * np.arange(numFrames) / float(fs)
binFreq = np.arange(N / 2) * float(fs) / N
plt.pcolormesh(frmTime, binFreq, np.transpose(mX))
plt.title('mX, M=1024, N=1024, H=512')
plt.autoscale(tight=True)
def main(inputFile='../../sounds/piano.wav',
window='hamming',
M=1024,
N=1024,
H=512):
"""
analysis/synthesis using the STFT
inputFile: input sound file (monophonic with sampling rate of 44100)
window: analysis window type (choice of rectangular, hanning, hamming, blackman, blackmanharris)
M: analysis window size
N: fft size (power of two, bigger or equal than M)
H: hop size (at least 1/2 of analysis window size to have good overlap-add)
"""
# read input sound (monophonic with sampling rate of 44100)
fs, x = UF.wavread(inputFile)
# compute analysis window
w = get_window(window, M)
# compute the magnitude and phase spectrogram
mX, pX = STFT.stftAnal(x, w, N, H)
# perform the inverse stft
y = STFT.stftSynth(mX, pX, M, H)
# output sound file (monophonic with sampling rate of 44100)
outputFile = 'output_sounds/' + os.path.basename(
inputFile)[:-4] + '_stft.wav'
# write the sound resulting from the inverse stft
UF.wavwrite(y, fs, outputFile)
# create figure to plot
plt.figure(figsize=(12, 9))
# frequency range to plot
maxplotfreq = 5000.0
# plot the input sound
plt.subplot(4, 1, 1)
plt.plot(np.arange(x.size) / float(fs), x)
plt.axis([0, x.size / float(fs), min(x), max(x)])
plt.ylabel('amplitude')
plt.xlabel('time (sec)')
plt.title('input sound: x')
# plot magnitude spectrogram
plt.subplot(4, 1, 2)
numFrames = int(mX[:, 0].size)
frmTime = H * np.arange(numFrames) / float(fs)
binFreq = fs * np.arange(N * maxplotfreq / fs) / N
plt.pcolormesh(frmTime, binFreq,
np.transpose(mX[:, :(int)(N * maxplotfreq / fs + 1)]))
plt.xlabel('time (sec)')
plt.ylabel('frequency (Hz)')
plt.title('magnitude spectrogram')
plt.autoscale(tight=True)
# plot the phase spectrogram
plt.subplot(4, 1, 3)
numFrames = int(pX[:, 0].size)
frmTime = H * np.arange(numFrames) / float(fs)
binFreq = fs * np.arange(N * maxplotfreq / fs) / N
plt.pcolormesh(
frmTime, binFreq,
np.transpose(np.diff(pX[:, :(int)(N * maxplotfreq / fs + 1)], axis=1)))
plt.xlabel('time (sec)')
plt.ylabel('frequency (Hz)')
plt.title('phase spectrogram (derivative)')
plt.autoscale(tight=True)
# plot the output sound
plt.subplot(4, 1, 4)
plt.plot(np.arange(y.size) / float(fs), y)
plt.axis([0, y.size / float(fs), min(y), max(y)])
plt.ylabel('amplitude')
plt.xlabel('time (sec)')
plt.title('output sound: y')
plt.tight_layout()
plt.ion()
plt.show()
def main(inputFile = '../../sounds/piano.wav', window = 'hamming', M = 1024, N = 1024, H = 512):
"""
analysis/synthesis using the STFT
inputFile: input sound file (monophonic with sampling rate of 44100)
window: analysis window type (choice of rectangular, hanning, hamming, blackman, blackmanharris)
M: analysis window size
N: fft size (power of two, bigger or equal than M)
H: hop size (at least 1/2 of analysis window size to have good overlap-add)
"""
# read input sound (monophonic with sampling rate of 44100)
fs, x = UF.wavread(inputFile)
# compute analysis window
w = get_window(window, M)
# compute the magnitude and phase spectrogram
mX, pX = STFT.stftAnal(x, w, N, H)
# perform the inverse stft
y = STFT.stftSynth(mX, pX, M, H)
# output sound file (monophonic with sampling rate of 44100)
outputFile = 'output_sounds/' + os.path.basename(inputFile)[:-4] + '_stft.wav'
# write the sound resulting from the inverse stft
UF.wavwrite(y, fs, outputFile)
# create figure to plot
plt.figure(figsize=(12, 9))
# frequency range to plot
maxplotfreq = 5000.0
# plot the input sound
plt.subplot(4,1,1)
plt.plot(np.arange(x.size)/float(fs), x)
plt.axis([0, x.size/float(fs), min(x), max(x)])
plt.ylabel('amplitude')
plt.xlabel('time (sec)')
plt.title('input sound: x')
# plot magnitude spectrogram
plt.subplot(4,1,2)
numFrames = int(mX[:,0].size)
frmTime = H*np.arange(numFrames)/float(fs)
binFreq = fs*np.arange(N*maxplotfreq/fs)/N
plt.pcolormesh(frmTime, binFreq, np.transpose(mX[:,:N*maxplotfreq/fs+1]))
plt.xlabel('time (sec)')
plt.ylabel('frequency (Hz)')
plt.title('magnitude spectrogram')
plt.autoscale(tight=True)
# plot the phase spectrogram
plt.subplot(4,1,3)
numFrames = int(pX[:,0].size)
frmTime = H*np.arange(numFrames)/float(fs)
binFreq = fs*np.arange(N*maxplotfreq/fs)/N
plt.pcolormesh(frmTime, binFreq, np.transpose(np.diff(pX[:,:N*maxplotfreq/fs+1],axis=1)))
plt.xlabel('time (sec)')
plt.ylabel('frequency (Hz)')
plt.title('phase spectrogram (derivative)')
plt.autoscale(tight=True)
# plot the output sound
plt.subplot(4,1,4)
plt.plot(np.arange(y.size)/float(fs), y)
plt.axis([0, y.size/float(fs), min(y), max(y)])
plt.ylabel('amplitude')
plt.xlabel('time (sec)')
plt.title('output sound: y')
plt.tight_layout()
plt.show()
def play(self):
""" Play entire file """
data = self.wf.readframes(chunk)
while data != '':
self.stream.write(data)
data = self.wf.readframes(chunk)
#save data to file
file = open("Text File/signal_data.txt", "w")
for item in self.datas:
file.write("%s " % item)
file.close()
'''
#Framing
framerate = self.rates
print framerate #menentukan jumlah frame
frame = round(len(self.datas)/framerate) #mengukur banyak data/frame
hop = 10 #jumlah frame yang diperiksa
overlap = 5 #lompatan frame
a = 0
while a < frame:
f_data = self.datas[a*int(frame):(a+hop)*int(frame)]
f_time = np.arange(a*(f_data.size/hop),(a+hop)*(f_data.size/hop))/float(self.rates)
title = "Frame",a/50+1,"Time-domain"
plt.title(title)
plt.xlabel("Time")
plt.ylabel("Amplitude")
plt.plot(f_time,f_data)
plt.show()
a += overlap
'''
'''
for i in range(hop):
f_data = self.datas[i*int(frame):(i+1)*int(frame)]
f_time = np.arange(i*f_data.size,(i+1)*f_data.size)/float(self.rates)
plt.title("Frame Time-domain")
plt.xlabel("Time")
plt.ylabel("Amplitude")
plt.plot(f_time,f_data)
plt.show()
'''
'''
f_time_sec = f_time[-1]
f_data = np.asarray(())
f_time = np.asarray(())
for i in range(int(frame)):
if len(f_data) == 0:
f_data = self.datas[i*framerate:(i+1)*framerate]
f_time = i*f_time_sec
else:
f_data = np.hstack((f_data,self.datas[i*framerate:(i+1)*framerate]))
f_time = np.hstack((f_time,i*f_time_sec))
if len(self.datas) % framerate > 0:
f_data = np.hstack((f_data,self.datas[(i+1)*framerate:]))
temp = (i+1)*f_time_sec
f_time = np.hstack((f_time,temp))
mod = len(self.datas) % framerate
antimod = len(self.datas) - mod
f_data = np.reshape(f_data[:antimod],(-1,framerate))
sisa_f_data = f_data[antimod:]
print f_data
print sisa_f_data
print f_time
'''
#proses STFT
N = 2048
M = 501 #'''bisa di set'''
H = M / 2 #bisa di set manual'
x = self.datas
x = np.float32(x) / norm_fact[x.dtype.name]
fs = self.rates
w = get_window('hamming', M)
global mX
mX, pX = stft.stftAnal(x, fs, w, N, H)
y = stft.stftSynth(mX, pX, M, H)
file = open("Text File/mX.txt", "w")
for j in range(len(mX)):
for item in mX[j]:
file.write("%s " % item)
file.write("\n\n")
file.close()
plt.figure(figsize=(12, 9))
maxplotfreq = 5000.0
plt.subplot(4, 1, 1)
plt.plot(np.arange(x.size) / float(fs), x)
plt.axis([0, x.size / float(fs), min(x), max(x)])
plt.ylabel('amplitude')
plt.xlabel('time (sec)')
plt.title('input sound: x')
plt.subplot(4, 1, 2)
numFrames = int(mX[:, 0].size)
frmTime = H * np.arange(numFrames) / float(fs)
binFreq = fs * np.arange(N * maxplotfreq / fs) / N
plt.pcolormesh(frmTime, binFreq,
np.transpose(mX[:, :int(N * maxplotfreq / fs + 1)]))
plt.xlabel('time (sec)')
plt.ylabel('frequency (Hz)')
plt.title('magnitude spectrogram')
plt.autoscale(tight=True)
file = open("Text File/STFT frequencies.txt", "w")
for item in binFreq:
file.write("%s\n" % item)
file.close()
plt.subplot(4, 1, 3)
numFrames = int(pX[:, 0].size)
frmTime = H * np.arange(numFrames) / float(fs)
binFreq = fs * np.arange(N * maxplotfreq / fs) / N
plt.pcolormesh(
frmTime, binFreq,
np.transpose(np.diff(pX[:, :int(N * maxplotfreq / fs + 1)],
axis=1)))
plt.xlabel('time (sec)')
plt.ylabel('frequency (Hz)')
plt.title('phase spectrogram (derivative)')
plt.autoscale(tight=True)
plt.subplot(4, 1, 4)
plt.plot(np.arange(y.size) / float(fs), y)
plt.axis([0, y.size / float(fs), min(y), max(y)])
plt.ylabel('amplitude')
plt.xlabel('time (sec)')
plt.title('output sound: y')
plt.tight_layout()
plt.show(block=False)
#print mX.shape
#proses finding peak
minimum = np.min(mX)
maximum = np.max(mX)
t = 0.8
sebaran = np.arange(minimum, maximum)
s_index = int(sebaran.size * (1 - t))
treshold = sebaran[-s_index]
print "treshold =", treshold
ploc = peakdetect.peakDetection(mX, treshold)
global peak_loc
global pmag
peak_loc = []
for i in range(len(ploc) - 1):
if ploc[i] != ploc[i + 1]:
peak_loc.append(ploc[i])
peak_loc.append(ploc[-1])
peak_loc = np.array(peak_loc)
file = open("Text File/peaks location.txt", "w")
for item in peak_loc:
file.write("%s\n" % item)
file.close()
file = open("Text File/peaks magnitude.txt", "w")
for item in peak_loc:
file.write("%s " % item)
for i in range(len(mX[item])):
file.write("%s " % mX[item, i])
file.write("\n\n")
file.close()
pmag = mX[peak_loc]
pl.plot(mX[-10, :])
pl.xlabel('Index')
pl.ylabel('Value')
pl.show()
'''
plt.plot(pX[-10,:])
pl.xlabel('Index')
pl.ylabel('Value')
pl.show()
'''
freqaxis = fs * np.arange(N / 2) / float(N)
plt.plot(freqaxis, mX[peak_loc[0], :-1])
pl.xlabel("Frequency")
pl.ylabel("Magnitude")
pl.show()
'''
file = open("Text File/peak frequencies.txt","w")
file.write("Frequency\tMagnitude\n")
for item in peak_loc:
file.write("%s\t" % freqaxis[item])
file.write("%s\n" % mX[peak_loc[0],item])
file.close()
'''
pl.plot(fs * peak_loc / float(N), pmag)
pl.xlabel("Frequency")
pl.ylabel("Magnitude")
pl.show()
#Menampilkan frequensi pada masing-masing Frame hasil STFT
loc = []
for m in pmag:
loc.append(np.argmax(m))
Freq = freqaxis[loc]
df = pd.DataFrame(Freq)
df.to_excel("Frekuensi Penyusun " +
self.filename.split("/")[-1].split(".")[0] + ".xlsx",
index=False)
#execfile("find_peak_cwt.py")
'''