In [157]:
%matplotlib inline
import glob
import os
import librosa
import librosa.display
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.pyplot import specgram
In [158]:
def load_sound_files(file_paths):
raw_sounds = []
for fp in file_paths:
X, sr = librosa.load(fp)
raw_sounds.append(X)
return raw_sounds
In [159]:
def plot_waves(sound_names, raw_sounds):
i = 1
fig = plt.figure(figsize=(25, 60), dpi=900)
for n, f in zip(sound_names, raw_sounds):
plt.subplot(10, 1, i)
librosa.display.waveplot(np.array(f), sr=22050)
plt.title(n.title())
i += 1
plt.suptitle('Figure 1: Waveplot', x=0.5, y=0.915, fontsize=18)
plt.show()
In [160]:
y, sr = librosa.load(librosa.util.example_audio_file(), duration=10)
plt.figure()
plt.subplot(2, 1, 1)
librosa.display.waveplot(y, sr=sr)
plt.title('librosa.display.waveplot()')
plt.subplot(2, 1, 2)
plt.plot(y)
plt.title('matplotlib')
Out[160]:
In [161]:
def plot_specgram(sound_names, raw_sounds):
i = 1
fig = plt.figure(figsize=(25, 60), dpi=900)
for n, f in zip(sound_names, raw_sounds):
plt.subplot(10, 1, i)
specgram(np.array(f), Fs=22500)
plt.title(n.title())
i += 1
plt.suptitle('Figure 2: Spectrogram', x=0.5, y=0.915, fontsize=18)
plt.show()
A numeric identifier of the sound class:
0 = air_conditioner
1 = car_horn
2 = children_playing
3 = dog_bark
4 = drilling
5 = engine_idling
6 = gun_shot
7 = jackhammer
8 = siren
9 = street_music
In [164]:
sound_file_paths = ['../../git/urban-sound-classification-keras/data/UrbanSound8K/audio/fold1/101415-3-0-2.wav']
sound_names = ['dog_bark']
raw_sounds = load_sound_files(sound_file_paths)
In [165]:
plot_waves(sound_names, raw_sounds)
In [166]:
plot_specgram(sound_names, raw_sounds)
In [167]:
def plot_log_power_specgram(sound_names, raw_sounds):
i = 1
fig = plt.figure(figsize=(25, 60), dpi=900)
for n, f in zip(sound_names, raw_sounds):
plt.subplot(10, 1, i)
# パワースペクトログラムをdB単位（log）に変換
D = librosa.logamplitude(np.abs(librosa.stft(f)) ** 2, ref_power=np.max)
librosa.display.specshow(D, x_axis='time', y_axis='log')
plt.title(n.title())
i += 1
plt.suptitle('Figure 3: Log power spectrogram', x=0.5, y=0.915, fontsize=18)
plt.show()
In [168]:
plot_log_power_specgram(sound_names, raw_sounds)
In [169]:
def extract_feature(file_name):
X, sample_rate = librosa.load(file_name)
stft = np.abs(librosa.stft(X))
# 特徴量は全フレームの平均を取る
# 1つの音声ファイルが193次元ベクトルで表せる
# MFCC 40次元
# (40 features, 173 frame) => (40 features, )
mfccs = np.mean(librosa.feature.mfcc(y=X, sr=sample_rate, n_mfcc=40), axis=1)
# クロマグラム 12次元
chroma = np.mean(librosa.feature.chroma_stft(S=stft, sr=sample_rate), axis=1)
# メルスペクトログラム 128次元
# (128 features, 173 frame) => (128 features, )
mel = np.mean(librosa.feature.melspectrogram(X, sr=sample_rate), axis=1)
# spectral contrast 7次元
contrast = np.mean(librosa.feature.spectral_contrast(S=stft, sr=sample_rate), axis=1)
# tonal centroid features 6次元
tonnetz = np.mean(librosa.feature.tonnetz(y=librosa.effects.harmonic(X), sr=sample_rate), axis=1)
return mfccs, chroma, mel, contrast, tonnetz
In [176]:
def parse_audio_files(parent_dir, sub_dirs, file_ext='*.wav'):
features, labels = np.empty((0, 193)), np.empty(0)
for sub_dir in sub_dirs:
print('***', sub_dir)
for fn in glob.glob(os.path.join(parent_dir, sub_dir, file_ext)):
try:
mfccs, chroma, mel, contrast, tonnetz = extract_feature(fn)
except Exception as e:
print('Error encountered while parsing file: ', fn)
continue
ext_features = np.hstack([mfccs, chroma, mel, contrast, tonnetz])
features = np.vstack([features, ext_features])
# data/UrbanSound8K/audio/fold1/101415-3-0-2.wav => 3
labels = np.append(labels, fn.split('/')[-1].split('-')[1])
return np.array(features), np.array(labels, dtype=np.int)
In [177]:
def one_hot_encode(labels):
n_labels = len(labels)
n_unique_labels = len(np.unique(labels))
one_hot_encode = np.zeros((n_labels,n_unique_labels))
one_hot_encode[np.arange(n_labels), labels] = 1
return one_hot_encode
In [175]:
parent_dir = '../data/UrbanSound8K/audio'
train_sub_dirs = ['fold1', 'fold2']
test_sub_dirs = ['fold3']
train_features, train_labels = parse_audio_files(parent_dir, train_sub_dirs)
test_features, test_labels = parse_audio_files(parent_dir, test_sub_dirs)
print(train_features.shape, train_labels.shape)
print(test_features.shape, test_labels.shape)
train_labels = one_hot_encode(train_labels)
test_labels = one_hot_encode(test_labels)
np.save('train_features.npy', train_features)
np.save('train_labels.npy', train_labels)
np.save('test_features.npy', test_features)
np.save('test_labels.npy', test_labels)
In [182]:
training_epochs = 50
n_dim = train_features.shape[1]
n_classes = 10
n_hidden_units_1 = 280
n_hidden_units_2 = 300
learning_rate = 0.01
In [186]:
from keras.models import Sequential
from keras.layers import Dense, Activation
from keras.optimizers import Adam
from keras.callbacks import CSVLogger, ModelCheckpoint
model = Sequential()
model.add(Dense(n_hidden_units_1, input_shape=(n_dim, )))
model.add(Activation('relu'))
model.add(Dense(n_hidden_units_2))
model.add(Activation('relu'))
model.add(Dense(n_classes))
model.add(Activation('softmax'))
In [184]:
model.summary()
In [187]:
logger = CSVLogger('train.log')
weight_file = 'train.{epoch:02d}-{val_loss:.3f}.h5'
checkpoint = ModelCheckpoint(weight_file,
monitor='val_loss',
verbose=1,
save_best_only=True,
mode='auto')
In [188]:
model.compile(loss='categorical_crossentropy',
optimizer='adam',
metrics=['accuracy'])
In [189]:
model.fit(train_features, train_labels,
batch_size=32,
epochs=50,
verbose=1,
validation_data=(test_features, test_labels),
callbacks=[logger, checkpoint])
In [ ]: