-
Notifications
You must be signed in to change notification settings - Fork 2
/
Compression_2.1.py
115 lines (93 loc) · 4.13 KB
/
Compression_2.1.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
"""
Created on Tue Oct 6 00:15:58 2020
@author: mockingbird01001
"""
import matplotlib.pyplot as plt
import numpy as np
import wave, math, contextlib
from pylab import ceil, arange, log10
from scipy.io import wavfile
# the .wav file to compresse
fname = 'wave.wav'
# the returned .wav file
outname = 'filtered.wav'
# color of the plot at the end
color = 'r' # red
# la frequence de coupure (soit tout ce qui depasse va a la poubelle)
cutOffFrequency = 900.0
def fft_dis(fname):
sampFreq, snd = wavfile.read(fname)
snd = snd / (2. ** 15) #convert sound array to float pt. values
s1, s2 = snd[0::2], snd[1::2] # left & right channels
#get the lenght & take the fourier transform of left channel
n, p = len(s1), np.fft.rfft(s1)
#get the lenght & take the fourier transform of left channel
m, p2 = len(s2), np.fft.rfft(s2)
# coef de moyennage des frequences
taille = 15 # de base 2.0 ~ 5.0
# get the
nUniquePts, mUniquePts = int(ceil((n + 1) / float(taille))), int(ceil((m + 1) / float(taille)))
p, p2 = p[0:nUniquePts], p2[0:mUniquePts]
p, p2 = abs(p), abs(p2)
p, p2 = p / float(n), p2 / float(m) # scale by the number of points so that
p = p ** 2
# we've got odd number of points fft
if n % 2 > 0:
p[1:len(p)] = p[1:len(p)] * 2
else:
# we've got even number of points fft
p[1:len(p) -1] = p[1:len(p) - 1] * 2
"""
Si vous affectez un canal Surround frontal ou arrière gauche à un fichier source stéréo
Compressor redirige le fichier source sur le canal gauche (et ignore le canal droit).
"""
freqArray = arange(0, nUniquePts, 1.0) * (sampFreq / n);
plt.plot(freqArray / cutOffFrequency, 10 * log10(p), color = color)
plt.xlabel('Frequency (kHz)')
plt.ylabel('Power (dB)')
plt.show()
def getFiltredArray(x, windowSize):
cumsum = np.cumsum(np.insert(x, 0, 0))
# on moyenne les frequences
return (cumsum[windowSize:] - cumsum[:-windowSize]) / windowSize
def getTypeWave(raw_bytes, n_frames, n_channels, sample_width, interleaved = True):
if sample_width == 1:
dtype = np.uint8 # 8 bits
elif sample_width == 2:
dtype = np.int16 # 16 bits => 2 bytes
else:
raise ValueError("Only supports 8 and 16 bit audio formats.")
channels = np.fromstring(raw_bytes, dtype=dtype)
if interleaved:
# channels are interleaved, Sample N of channel M follows sample N of channel M-1 in raw data
channels.shape = (n_frames, n_channels)
channels = channels.T
else:
# All samples from channel M occur before all samples from channel M-1
channels.shape = (n_channels, n_frames)
return channels
# 'rb' => for read only
# open the file fname and only read it
with contextlib.closing(wave.open(fname,'rb')) as originWave:
# initialisation
waveRate, amplitudeWidth = originWave.getframerate(), originWave.getsampwidth()
nChannels, nbFrames = originWave.getnchannels(), originWave.getnframes()
# Extract Raw Audio from multi-channel Wav File
signal = originWave.readframes(nbFrames * nChannels)
channels = getTypeWave(signal, nbFrames, nChannels, amplitudeWidth)
# get window size
fqRatio = (cutOffFrequency / waveRate)
N = int( math.sqrt( np.random.randint(0, 1) + fqRatio ** 2) / fqRatio )
# Use moviung average (only on first channel)
filt = getFiltredArray(channels[0], N).astype(channels.dtype)
with wave.open(outname, "w") as filtredWave:
filtredWave.setparams((1, amplitudeWidth, waveRate, nbFrames, originWave.getcomptype(), originWave.getcompname()))
# the values are 'C' & 'F' for C language and Fortran language
filtredWave.writeframes(filt.tobytes('C'))
# the files are closed automaticaly at the end of the with
# montre le plot du fichier de base "wave.wav"
print("Fichier Avant ! {}".format(fname))
fft_dis(fname)
# montre le plot du fichier apres la compression "filtred.wav"
print("Fichier Apres ! {} ~{}".format(fname, outname))
fft_dis(outname)