Instrument_Classification/eda.py at master · pradyunkumar/Instrument_Classification · GitHub

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
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
import os
from tqdm import tqdm
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from scipy.io import wavfile
from python_speech_features import mfcc, logfbank
import librosa

def plot_signals(signals):
    fig, axes = plt.subplots(nrows=2, ncols=5, sharex=False,
                             sharey=True, figsize=(20,5))
    fig.suptitle('Time Series', size=16)
    i = 0
    for x in range(2):
        for y in range(5):
            axes[x,y].set_title(list(signals.keys())[i])
            axes[x,y].plot(list(signals.values())[i])
            axes[x,y].get_xaxis().set_visible(False)
            axes[x,y].get_yaxis().set_visible(False)
            i += 1

def plot_fft(fft):
    fig, axes = plt.subplots(nrows=2, ncols=5, sharex=False,
                             sharey=True, figsize=(20,5))
    fig.suptitle('Fourier Transforms', size=16)
    i = 0
    for x in range(2):
        for y in range(5):
            data = list(fft.values())[i]
            Y, freq = data[0], data[1]
            axes[x,y].set_title(list(fft.keys())[i])
            axes[x,y].plot(freq, Y)
            axes[x,y].get_xaxis().set_visible(False)
            axes[x,y].get_yaxis().set_visible(False)
            i += 1

def plot_fbank(fbank):
    fig, axes = plt.subplots(nrows=2, ncols=5, sharex=False,
                             sharey=True, figsize=(20,5))
    fig.suptitle('Filter Bank Coefficients', size=16)
    i = 0
    for x in range(2):
        for y in range(5):
            axes[x,y].set_title(list(fbank.keys())[i])
            axes[x,y].imshow(list(fbank.values())[i],
                    cmap='hot', interpolation='nearest')
            axes[x,y].get_xaxis().set_visible(False)
            axes[x,y].get_yaxis().set_visible(False)
            i += 1

def plot_mfccs(mfccs):
    fig, axes = plt.subplots(nrows=2, ncols=5, sharex=False,
                             sharey=True, figsize=(20,5))
    fig.suptitle('Mel Frequency Cepstrum Coefficients', size=16)
    i = 0
    for x in range(2):
        for y in range(5):
            axes[x,y].set_title(list(mfccs.keys())[i])
            axes[x,y].imshow(list(mfccs.values())[i],
                    cmap='hot', interpolation='nearest')
            axes[x,y].get_xaxis().set_visible(False)
            axes[x,y].get_yaxis().set_visible(False)
            i += 1

def envelope(y, rate, thresh):
    mask = []
    y = pd.Series(y).apply(np.abs)
    y_mean = y.rolling(window=int(rate/10), min_periods=1, center=True).mean()  # creates a rolling window to check mean of vals
    for mean in y_mean:
        if mean > thresh:
            mask.append(True)
        else:
            mask.append(False)

    return mask

def calc_fft(y, rate):
    n = len(y)
    freq = np.fft.rfftfreq(n, d=1/rate)
    Y = abs(np.fft.rfft(y)/n)
    return Y, freq

# creating the initial dataframe
df = pd.read_csv('instruments.csv')
df.set_index('fname', inplace=True)

# calculates lenghth of audio
for f in df.index:
    rate, signal = wavfile.read('wavfiles/'+f)
    df.at[f, 'length'] = signal.shape[0]/rate

# creates list and mean amount of data for each instrument
classes = list(np.unique(df.label))
class_dist = df.groupby(['label'])['length'].mean()

fig, ax = plt.subplots()
ax.set_title('Class Distribution', y=1.08)
ax.pie(class_dist, labels=class_dist.index, autopct='%1.1f%%',
        shadow=False, startangle=90)
ax.axis('equal')
plt.show()
df.reset_index(inplace=True)

signals = {}
fft = {}
fbank = {}
mfccs = {}

for c in classes:
    # captures first file of each instrument
    wav_file = df[df.label == c].iloc[0,0]

    signal, rate = librosa.load('wavfiles/'+ wav_file, sr=44100)
    mask = envelope(signal, rate, 0.0005)
    signal = signal[mask]
    signals[c] = signal
    fft[c] = calc_fft(signal, rate)

    bank = logfbank(signal[:rate], rate, nfilt=26, nfft=1103).T
    fbank[c] = bank
    mel = mfcc(signal[:rate], rate, numcep=13, nfilt=26, nfft=1103).T
    mfccs[c] = mel

plot_signals(signals)
plt.show()

plot_fft(fft)
plt.show()

plot_fbank(fbank)
plt.show()

plot_mfccs(mfccs)
plt.show()

if len(os.listdir('clean')) == 0:
    for f in tqdm(df.fname):
        signal, rate = librosa.load('wavfiles/'+f, sr= 16000)
        mask = envelope(signal, rate, 0.005)
        wavfile.write(filename='clean/'+f, rate=rate, data=signal[mask])