Skip to content

2
2021-005
Commit 29ddc2e0 authored by Dinushe Jayasekera's avatar Dinushe Jayasekera
Browse files
Options

python notebook for audio recognition

parent 5091c0b3
Hide whitespace changes
Inline Side-by-side
Showing with 768 additions and 0 deletions
Audio.ipynb 0 → 100644
View file @ 29ddc2e0
{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import keras\n",
"from keras.layers import Activation, Dense, Dropout, Conv2D, \\\n",
" Flatten, MaxPooling2D\n",
"from keras.models import Sequential\n",
"from keras.utils import np_utils\n",
"import librosa\n",
"import librosa.display\n",
"import numpy as np\n",
"import pandas as pd\n",
"import random\n",
"\n",
"import warnings\n",
"warnings.filterwarnings('ignore')"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>slice_file_name</th>\n",
" <th>fsID</th>\n",
" <th>start</th>\n",
" <th>end</th>\n",
" <th>salience</th>\n",
" <th>fold</th>\n",
" <th>classID</th>\n",
" <th>class</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>100032-3-0-0.wav</td>\n",
" <td>100032</td>\n",
" <td>0.0</td>\n",
" <td>0.317551</td>\n",
" <td>1</td>\n",
" <td>5</td>\n",
" <td>3</td>\n",
" <td>children_playing2</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>100263-2-0-117.wav</td>\n",
" <td>100263</td>\n",
" <td>58.5</td>\n",
" <td>62.500000</td>\n",
" <td>1</td>\n",
" <td>5</td>\n",
" <td>2</td>\n",
" <td>children_playing</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>100263-2-0-121.wav</td>\n",
" <td>100263</td>\n",
" <td>60.5</td>\n",
" <td>64.500000</td>\n",
" <td>1</td>\n",
" <td>5</td>\n",
" <td>2</td>\n",
" <td>children_playing</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>100263-2-0-126.wav</td>\n",
" <td>100263</td>\n",
" <td>63.0</td>\n",
" <td>67.000000</td>\n",
" <td>1</td>\n",
" <td>5</td>\n",
" <td>2</td>\n",
" <td>children_playing</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>100263-2-0-137.wav</td>\n",
" <td>100263</td>\n",
" <td>68.5</td>\n",
" <td>72.500000</td>\n",
" <td>1</td>\n",
" <td>5</td>\n",
" <td>2</td>\n",
" <td>children_playing</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" slice_file_name fsID start end salience fold classID \\\n",
"0 100032-3-0-0.wav 100032 0.0 0.317551 1 5 3 \n",
"1 100263-2-0-117.wav 100263 58.5 62.500000 1 5 2 \n",
"2 100263-2-0-121.wav 100263 60.5 64.500000 1 5 2 \n",
"3 100263-2-0-126.wav 100263 63.0 67.000000 1 5 2 \n",
"4 100263-2-0-137.wav 100263 68.5 72.500000 1 5 2 \n",
"\n",
" class \n",
"0 children_playing2 \n",
"1 children_playing \n",
"2 children_playing \n",
"3 children_playing \n",
"4 children_playing "
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Read Data\n",
"data = pd.read_csv('ASDmeta.csv')\n",
"data.head(5)"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(8732, 8)"
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"data.shape"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(7468, 4)"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Get data over 3 seconds long\n",
"valid_data = data[['slice_file_name', 'fold' ,'classID', 'class']][ data['end']-data['start'] >= 3 ]\n",
"valid_data.shape"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(128, 128)"
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Example of a children play spectrogram\n",
"y, sr = librosa.load('audio/fold5/100263-2-0-137.wav', duration=2.97)\n",
"ps = librosa.feature.melspectrogram(y=y, sr=sr)\n",
"ps.shape"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"<matplotlib.collections.QuadMesh at 0x2ae074fe280>"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"librosa.display.specshow(ps, y_axis='mel', x_axis='time')"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"###for one audio file \n",
"def features_extractor(file):\n",
" audio, sample_rate = librosa.load(file_name, res_type='kaiser_fast') \n",
" mfccs_features = librosa.feature.mfcc(y=audio, sr=sample_rate, n_mfcc=40)\n",
" mfccs_scaled_features = np.mean(mfccs_features.T,axis=0)\n",
" return mfccs_scaled_features"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"7468it [05:14, 23.76it/s]\n"
]
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"import os\n",
"from tqdm import tqdm #to see progress\n",
"#audio_dataset_path='D:/4thYear/Research/AudioClassifier(ETA)/UrbanSound8K/audio'\n",
"extracted_features=[]\n",
"for index_num,row in tqdm(valid_data.iterrows()):\n",
" #if row[\"class\"] == 'children_playing' or row[\"class\"] == 'children_playing2':\n",
" #y, sr = librosa.load('audio/fold'+ str(row[\"fold\"]) +'/'+ row[\"slice_file_name\"], duration=2.97)\n",
" y, sr = librosa.load('audio/fold'+ str(row[\"fold\"]) +'/'+ row[\"slice_file_name\"], duration=2.97)\n",
" #for newly created set\n",
" if os.path.exists('audio/fold' + str(row[\"fold\"]) + '/speed_107'+ row[\"slice_file_name\"]):\n",
" #os.makedirs(newpath)\n",
" ps = librosa.feature.melspectrogram(y=y, sr=sr)\n",
" if ps.shape != (128, 128): \n",
" continue\n",
" librosa.display.specshow(ps, y_axis='mel', x_axis='time')\n",
" ps.shape\n",
" #print(' ID:'+ str(row.classID))\n",
" extracted_features.append( (ps, row.classID) )\n",
" ps = librosa.feature.melspectrogram(y=y, sr=sr)\n",
" if ps.shape != (128, 128): \n",
" continue\n",
" librosa.display.specshow(ps, y_axis='mel', x_axis='time')\n",
" ps.shape\n",
" #print(' ID:'+ str(row.classID))\n",
" extracted_features.append( (ps, row.classID) )"
]
},
{
"cell_type": "code",
"execution_count": 41,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Number of samples: 1677\n"
]
}
],
"source": [
"print(\"Number of samples: \", len(extracted_features))\n"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [],
"source": [
"dataset = extracted_features\n",
"random.shuffle(dataset)\n",
"\n",
"train = dataset[:1500]\n",
"test = dataset[1500:]\n",
"\n",
"X_train, y_train = zip(*train)\n",
"X_test, y_test = zip(*test)\n",
"\n",
"# Reshape for CNN input\n",
"X_train = np.array([x.reshape( (128, 128, 1) ) for x in X_train])\n",
"X_test = np.array([x.reshape( (128, 128, 1) ) for x in X_test])\n",
"\n",
"# One-Hot encoding for classes\n",
"y_train = np.array(keras.utils.np_utils.to_categorical(y_train, 10))\n",
"y_test = np.array(keras.utils.np_utils.to_categorical(y_test, 10))"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [],
"source": [
"model = Sequential()\n",
"input_shape=(128, 128, 1)\n",
"\n",
"model.add(Conv2D(24, (5, 5), strides=(1, 1), input_shape=input_shape))\n",
"model.add(MaxPooling2D((4, 2), strides=(4, 2)))\n",
"model.add(Activation('relu'))\n",
"\n",
"model.add(Conv2D(48, (5, 5), padding=\"valid\"))\n",
"model.add(MaxPooling2D((4, 2), strides=(4, 2)))\n",
"model.add(Activation('relu'))\n",
"\n",
"model.add(Conv2D(48, (5, 5), padding=\"valid\"))\n",
"model.add(Activation('relu'))\n",
"\n",
"model.add(Flatten())\n",
"model.add(Dropout(rate=0.5))\n",
"\n",
"model.add(Dense(64))\n",
"model.add(Activation('relu'))\n",
"model.add(Dropout(rate=0.5))\n",
"\n",
"model.add(Dense(10))\n",
"model.add(Activation('softmax'))"
]
},
{
"cell_type": "code",
"execution_count": 44,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Epoch 1/50\n",
"12/12 [==============================] - 19s 1s/step - loss: 0.5703 - accuracy: 0.7495 - val_loss: 0.4692 - val_accuracy: 0.8023\n",
"Epoch 2/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.6080 - accuracy: 0.7511 - val_loss: 0.4458 - val_accuracy: 0.8079\n",
"Epoch 3/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.6012 - accuracy: 0.7419 - val_loss: 0.4120 - val_accuracy: 0.8305\n",
"Epoch 4/50\n",
"12/12 [==============================] - 17s 1s/step - loss: 0.5109 - accuracy: 0.7659 - val_loss: 0.4143 - val_accuracy: 0.8249\n",
"Epoch 5/50\n",
"12/12 [==============================] - 18s 1s/step - loss: 0.4989 - accuracy: 0.7868 - val_loss: 0.4059 - val_accuracy: 0.8136\n",
"Epoch 6/50\n",
"12/12 [==============================] - 17s 1s/step - loss: 0.4659 - accuracy: 0.7853 - val_loss: 0.4052 - val_accuracy: 0.8362\n",
"Epoch 7/50\n",
"12/12 [==============================] - 17s 1s/step - loss: 0.4537 - accuracy: 0.8102 - val_loss: 0.4432 - val_accuracy: 0.8249\n",
"Epoch 8/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.4390 - accuracy: 0.7917 - val_loss: 0.4371 - val_accuracy: 0.8362\n",
"Epoch 9/50\n",
"12/12 [==============================] - 19s 2s/step - loss: 0.4265 - accuracy: 0.8203 - val_loss: 0.4093 - val_accuracy: 0.8192\n",
"Epoch 10/50\n",
"12/12 [==============================] - 17s 1s/step - loss: 0.4112 - accuracy: 0.8354 - val_loss: 0.3964 - val_accuracy: 0.8305\n",
"Epoch 11/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.4019 - accuracy: 0.8209 - val_loss: 0.4434 - val_accuracy: 0.7740\n",
"Epoch 12/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.3762 - accuracy: 0.8350 - val_loss: 0.4082 - val_accuracy: 0.8192\n",
"Epoch 13/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.3343 - accuracy: 0.8576 - val_loss: 0.4074 - val_accuracy: 0.8588\n",
"Epoch 14/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.3491 - accuracy: 0.8521 - val_loss: 0.4439 - val_accuracy: 0.8192\n",
"Epoch 15/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.3548 - accuracy: 0.8498 - val_loss: 0.5008 - val_accuracy: 0.8305\n",
"Epoch 16/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.4044 - accuracy: 0.8507 - val_loss: 0.6734 - val_accuracy: 0.8136\n",
"Epoch 17/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.5081 - accuracy: 0.8129 - val_loss: 0.4393 - val_accuracy: 0.7797\n",
"Epoch 18/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.4016 - accuracy: 0.8106 - val_loss: 0.4132 - val_accuracy: 0.8305\n",
"Epoch 19/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.3454 - accuracy: 0.8516 - val_loss: 0.4011 - val_accuracy: 0.8588\n",
"Epoch 20/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.3868 - accuracy: 0.8440 - val_loss: 0.4531 - val_accuracy: 0.8305\n",
"Epoch 21/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.3538 - accuracy: 0.8414 - val_loss: 0.3996 - val_accuracy: 0.8418\n",
"Epoch 22/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.3289 - accuracy: 0.8568 - val_loss: 0.4396 - val_accuracy: 0.8531\n",
"Epoch 23/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.2906 - accuracy: 0.8743 - val_loss: 0.4508 - val_accuracy: 0.8192\n",
"Epoch 24/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.2776 - accuracy: 0.8785 - val_loss: 0.4434 - val_accuracy: 0.8418\n",
"Epoch 25/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.3088 - accuracy: 0.8714 - val_loss: 0.4205 - val_accuracy: 0.8644\n",
"Epoch 26/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.2936 - accuracy: 0.8863 - val_loss: 0.3697 - val_accuracy: 0.8531\n",
"Epoch 27/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.2680 - accuracy: 0.8900 - val_loss: 0.3482 - val_accuracy: 0.8870\n",
"Epoch 28/50\n",
"12/12 [==============================] - 18s 1s/step - loss: 0.2504 - accuracy: 0.9007 - val_loss: 0.4004 - val_accuracy: 0.8870\n",
"Epoch 29/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.2617 - accuracy: 0.8909 - val_loss: 0.5071 - val_accuracy: 0.8644\n",
"Epoch 30/50\n",
"12/12 [==============================] - 19s 2s/step - loss: 0.2602 - accuracy: 0.8958 - val_loss: 0.5950 - val_accuracy: 0.8531\n",
"Epoch 31/50\n",
"12/12 [==============================] - 21s 2s/step - loss: 0.2396 - accuracy: 0.8976 - val_loss: 0.6384 - val_accuracy: 0.8644\n",
"Epoch 32/50\n",
"12/12 [==============================] - 18s 2s/step - loss: 0.2983 - accuracy: 0.8772 - val_loss: 0.6448 - val_accuracy: 0.8305\n",
"Epoch 33/50\n",
"12/12 [==============================] - 18s 2s/step - loss: 0.2954 - accuracy: 0.8778 - val_loss: 0.3525 - val_accuracy: 0.8531\n",
"Epoch 34/50\n",
"12/12 [==============================] - 17s 1s/step - loss: 0.2521 - accuracy: 0.9003 - val_loss: 0.3455 - val_accuracy: 0.8644\n",
"Epoch 35/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.2271 - accuracy: 0.8958 - val_loss: 0.4037 - val_accuracy: 0.8644\n",
"Epoch 36/50\n",
"12/12 [==============================] - 18s 1s/step - loss: 0.2035 - accuracy: 0.9187 - val_loss: 0.3996 - val_accuracy: 0.8983\n",
"Epoch 37/50\n",
"12/12 [==============================] - 20s 2s/step - loss: 0.2229 - accuracy: 0.9141 - val_loss: 0.5673 - val_accuracy: 0.8644\n",
"Epoch 38/50\n",
"12/12 [==============================] - 17s 1s/step - loss: 0.2107 - accuracy: 0.9098 - val_loss: 0.4846 - val_accuracy: 0.8475\n",
"Epoch 39/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.2309 - accuracy: 0.8998 - val_loss: 0.6008 - val_accuracy: 0.8588\n",
"Epoch 40/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.2562 - accuracy: 0.9124 - val_loss: 0.4937 - val_accuracy: 0.8588\n",
"Epoch 41/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.2103 - accuracy: 0.9180 - val_loss: 0.5335 - val_accuracy: 0.8644\n",
"Epoch 42/50\n",
"12/12 [==============================] - 18s 1s/step - loss: 0.1739 - accuracy: 0.9195 - val_loss: 0.5026 - val_accuracy: 0.8701\n",
"Epoch 43/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.2074 - accuracy: 0.9024 - val_loss: 0.8531 - val_accuracy: 0.8362\n",
"Epoch 44/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.2637 - accuracy: 0.9094 - val_loss: 0.4599 - val_accuracy: 0.8588\n",
"Epoch 45/50\n",
"12/12 [==============================] - 16s 1s/step - loss: 0.2997 - accuracy: 0.8787 - val_loss: 0.4988 - val_accuracy: 0.8588\n",
"Epoch 46/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.2285 - accuracy: 0.9101 - val_loss: 0.5542 - val_accuracy: 0.8588\n",
"Epoch 47/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.2146 - accuracy: 0.9145 - val_loss: 0.7282 - val_accuracy: 0.7853\n",
"Epoch 48/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.2460 - accuracy: 0.9063 - val_loss: 0.9213 - val_accuracy: 0.7288\n",
"Epoch 49/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.4337 - accuracy: 0.8456 - val_loss: 0.4394 - val_accuracy: 0.8531\n",
"Epoch 50/50\n",
"12/12 [==============================] - 15s 1s/step - loss: 0.2752 - accuracy: 0.8978 - val_loss: 0.4272 - val_accuracy: 0.8531\n",
"6/6 [==============================] - 1s 87ms/step - loss: 0.4272 - accuracy: 0.8531\n",
"Test loss: 0.427224338054657\n",
"Test accuracy: 0.8531073331832886\n"
]
}
],
"source": [
"model.compile(optimizer=\"Adam\",loss=\"categorical_crossentropy\",metrics=['accuracy'])\n",
"\n",
"model.fit(x=X_train, y=y_train,epochs=50,batch_size=128,validation_data= (X_test, y_test))\n",
"\n",
"score = model.evaluate(x=X_test,y=y_test)\n",
"\n",
"print('Test loss:', score[0])\n",
"print('Test accuracy:', score[1])"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [],
"source": [
"#data augmentation\n",
"#time variation\n",
"import soundfile as sf\n",
"\n",
"y, sr = librosa.load('audio/fold1/14113-4-0-1.wav', duration=2.97)\n",
"y_changed = librosa.effects.time_stretch(y, rate=0.81)\n",
"sf.write('augmented/fold1/speed_81/14113-4-0-1.wav' ,y_changed, sr)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"#newpath = r'C:\\Program Files\\arbitrary' \n",
"if not os.path.exists('audio/fold' + str(row[\"fold\"]) + '/speed_' + str(int(rate*100))):\n",
" os.makedirs(newpath)"
]
},
{
"cell_type": "code",
"execution_count": 34,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"1it [00:00, 4.35it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"new path created : audio/fold5/speed_107\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"9it [00:02, 4.66it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"new path created : audio/fold2/speed_107\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"13it [00:02, 4.65it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"new path created : audio/fold10/speed_107\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"48it [00:03, 6.61it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"new path created : audio/fold6/speed_107\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"57it [00:05, 5.26it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"new path created : audio/fold1/speed_107\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"95it [00:06, 8.00it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"new path created : audio/fold4/speed_107\n",
"new path created : audio/fold3/speed_107\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"400it [00:12, 13.17it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"new path created : audio/fold9/speed_107\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"580it [00:21, 10.03it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"new path created : audio/fold8/speed_107\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"591it [00:22, 10.09it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"new path created : audio/fold7/speed_107\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"7468it [06:22, 19.50it/s]\n"
]
}
],
"source": [
"rate = 1.07 # replace with 0.81 and execute again\n",
"\n",
"for index_num,row in tqdm(valid_data.iterrows()):\n",
" if row[\"class\"] == 'children_playing' or row[\"class\"] == 'children_playing2':\n",
" y, sr = librosa.load('audio/fold'+ str(row[\"fold\"]) +'/'+ row[\"slice_file_name\"]) \n",
" y_changed = librosa.effects.time_stretch(y, rate=rate)\n",
" if not os.path.exists('audio/fold' + str(row[\"fold\"]) + '/speed_' + str(int(rate*100))):\n",
" os.makedirs('audio/fold' + str(row[\"fold\"]) + '/speed_' + str(int(rate*100)))\n",
" print('new path created : ' + 'audio/fold' + str(row[\"fold\"]) + '/speed_' + str(int(rate*100)))\n",
" sf.write('audio/fold' + str(row[\"fold\"]) + '/speed_' + str(int(rate*100)) + '/' + row[\"slice_file_name\"] ,y_changed, sr)"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1677"
]
},
"execution_count": 37,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"len(extracted_features)"
]
},
{
"cell_type": "code",
"execution_count": 45,
"metadata": {},
"outputs": [],
"source": [
"model.save_weights('model_weights_acc_new.h5')"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"#for varying pitch\n",
"n_steps = 2 #-1, -2, 2, 1\n",
"\n",
"for row in valid_data.itertuples():\n",
" y, sr = librosa.load('audio/' + row.path) \n",
" y_changed = librosa.effects.pitch_shift(y, sr, n_steps=n_steps)\n",
" librosa.output.write_wav('augmented/fold' + str(row.fold) + '/ps1_' + str(int(n_steps)) + '/' + row.slice_file_name ,y_changed, sr)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"n_steps = 2.5 #-2.5, -3.5, 2.5, 3.5\n",
"\n",
"for row in valid_data.itertuples():\n",
" y, sr = librosa.load('audio/' + row.path) \n",
" y_changed = librosa.effects.pitch_shift(y, sr, n_steps=n_steps)\n",
" librosa.output.write_wav('code/augmented/fold' + str(row.fold) + '/ps2_m' + str(int(n_steps*10)) + '/' + row.slice_file_name ,y_changed, sr)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.5"
}
},
"nbformat": 4,
"nbformat_minor": 4
}
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment