Sustained_Phonation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true,
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "# Sustained Phonation Feature Extraction Notebook\n",
    "In this iPython Notebook, we'll show you how to use\n",
    "our implementation of the features used in <paper citation>\n",
    "to extract them on your own dataset.\n",
    "\n",
    "To follow this tutorial, you'll need some audio files to run our\n",
    "feature extractors on. Ideally, these audio samples each\n",
    "contain a recording of someone speaking the vowel /a/ for a few seconds."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "## Step 1: Loading your audio samples\n",
    "\n",
    "In this part of the notebook, we load the audio samples,\n",
    "the goal is to have them all as Numpy arrays. We'll store them in\n",
    "a list of `AudioSample` objects which just stores the PCM array,\n",
    "its sampling rate, and and a string identifier for the sample.\n",
    "\n",
    "In this example, we suppose that all the audio samples are read from\n",
    "single WAV files in a folder."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "from features import AudioSample\n",
    "from pathlib import Path\n",
    "import pandas as pd\n",
    "from typing import List\n",
    "import librosa\n",
    "import numpy as np\n",
    "\n",
    "samples_folder = Path('example_sound/')\n",
    "audio_samples: List[AudioSample] = []\n",
    "for wav_file in samples_folder.iterdir():\n",
    "    if not wav_file.suffix == \".wav\":\n",
    "        continue\n",
    "    else:\n",
    "        audio_array, rate = librosa.core.load(str(wav_file), sr=44100)\n",
    "        audio_samples.append(AudioSample(data=audio_array,\n",
    "                                         rate=rate,\n",
    "                                         name=wav_file.name))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "## Step 2: Configure the feature extractors\n",
    "\n",
    "In this step, you can tweak, remove of add feature extractors.\n",
    "Here we've put all the feature extractors from our paper 2020 Interspeech paper ['Vocal markers from sustained phonation in Huntington’s Disease'](https://arxiv.org/abs/2006.05365).\n",
    "\n",
    "All of their actual implementations are kept in the `features` module contained\n",
    "in this repository."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The DFA and RPDE are not computed by default.\n"
     ]
    }
   ],
   "source": [
    "from typing import Dict\n",
    "\n",
    "print('The DFA and RPDE are not computed by default.')\n",
    "from features import *\n",
    "compute_dfa_rpde = False\n",
    "features_dict: Dict[str, SampleProcessor] = {\n",
    "    \"Maximum_phonation_time\": MPT(),\n",
    "    'Maximum_phonation_time_until_voice_break': MaximumPhonationUntilVoiceBreak(),\n",
    "    'Aperiodicity_Features': AperiodicityFeatures(),\n",
    "    'MFCCs_Mean_Of_Std': MFCCsMeanOfStd(),\n",
    "    \"hnr\": PraatHNR(),\n",
    "    \"tremor\": VocalTremorFeatures(),\n",
    "    \"signal_perturbation\": PraatVoiceFeatures(),\n",
    "    \"mps\": ModulationPowerSpectrumAnalysis(),\n",
    "    \"f0_stats\" : FundamentalFrequency(method='shennong'),\n",
    "}\n",
    "if compute_dfa_rpde:\n",
    "    features_dict[\"rpde\"]= RPDE(dim=4, tau=35, epsilon=0.12, tmax=5000)\n",
    "    features_dict[\"dfa\"]= DFA(scale_boundaries=(1.259, 1.559), scale_density=0.03)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "## Step 3: Build and run the extraction\n",
    "\n",
    "We'll define here a helper function that runs the extraction using the samples\n",
    "and the selected feature extractors (or processors). Then, we'll run\n",
    "that function on our audio samples dataset.\n",
    "\n",
    "Once the feature extraction is complete, all the features computed for all the\n",
    "samples should be stored in dictionary. We format the classical phonatory features in a final dataframe `classic_phonatory_features_df`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "from collections import defaultdict\n",
    "import tqdm\n",
    "\n",
    "def extract(samples_list: List[AudioSample],\n",
    "          features: Dict[str, SampleProcessor],\n",
    "          fail_on_error: bool = True):\n",
    "    \"\"\"This function takes care of running the extraction,\n",
    "    and stores the extracted features for each sample in a neat\n",
    "    little dictionnary\"\"\"\n",
    "    extracted_features = defaultdict(dict)\n",
    "    for feature_name, processor in features.items():\n",
    "        for audio_sample in tqdm.tqdm(samples_list):\n",
    "            try:\n",
    "                proc_output = processor(audio_sample)\n",
    "            except Exception as e:\n",
    "                if fail_on_error:\n",
    "                    raise e\n",
    "                else:\n",
    "                    print(f\"Got error on sample '{audio_sample.name}':' \"\n",
    "                          f\"{type(e).__name__} :{str(e)}', skipping.\")\n",
    "            else:\n",
    "                # maybe we should use tuples for the dict keys?\n",
    "                extracted_features[audio_sample.name][feature_name] = proc_output\n",
    "\n",
    "    return extracted_features\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 5/5 [00:00<00:00, 13131.82it/s]\n",
      "100%|██████████| 5/5 [00:02<00:00,  1.67it/s]\n",
      "100%|██████████| 5/5 [00:02<00:00,  2.44it/s]\n",
      "100%|██████████| 5/5 [00:01<00:00,  4.31it/s]\n",
      "100%|██████████| 5/5 [00:01<00:00,  2.57it/s]\n",
      "100%|██████████| 5/5 [00:06<00:00,  1.32s/it]\n",
      "100%|██████████| 5/5 [00:01<00:00,  2.90it/s]\n",
      "100%|██████████| 5/5 [00:22<00:00,  4.53s/it]\n",
      "100%|██████████| 5/5 [00:01<00:00,  4.70it/s]\n"
     ]
    }
   ],
   "source": [
    "# then, we run the extraction.\n",
    "# WARNING: this may (and should) take some time if DFA and RPDE are included.\n",
    "extracted_features = extract(audio_samples, features_dict, fail_on_error=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Formatting the features\n",
    "for file_name in extracted_features.keys():\n",
    "    extracted_features[file_name]['Fraction of locally unvoiced frames'] = extracted_features[file_name]['Aperiodicity_Features']['Fraction of locally unvoiced frames']\n",
    "    extracted_features[file_name]['Number of voice breaks'] = extracted_features[file_name]['Aperiodicity_Features']['Number of voice breaks']\n",
    "    extracted_features[file_name]['Degree of voice breaks'] = extracted_features[file_name]['Aperiodicity_Features']['Degree of voice breaks']\n",
    "    extracted_features[file_name]['Mean_Of_Std_mfcc'] = extracted_features[file_name]['MFCCs_Mean_Of_Std']['mfcc']\n",
    "    extracted_features[file_name]['Mean_Of_Std_delta_mfcc'] = extracted_features[file_name]['MFCCs_Mean_Of_Std']['delta_mfcc']\n",
    "    extracted_features[file_name]['frequency tremor intensity index (FTrI)'] = extracted_features[file_name]['tremor']['frequency tremor intensity index (FTrI)']\n",
    "    extracted_features[file_name]['amplitude tremor intensity index (ATrI)'] = extracted_features[file_name]['tremor']['amplitude tremor intensity index (ATrI)']\n",
    "    extracted_features[file_name]['jitter'] = extracted_features[file_name]['signal_perturbation']['jitter']['local']\n",
    "    extracted_features[file_name]['shimmer'] = extracted_features[file_name]['signal_perturbation']['shimmer']['local']\n",
    "    extracted_features[file_name]['F0_Std'] = np.std(extracted_features[file_name]['f0_stats'])\n",
    "\n",
    "airflow_features_names = ['Maximum_phonation_time','Maximum_phonation_time_until_voice_break']\n",
    "aperiodicity_features_names = ['Number of voice breaks','Degree of voice breaks','Fraction of locally unvoiced frames']\n",
    "signal_perturbation_features_names = ['jitter','shimmer']\n",
    "if compute_dfa_rpde:\n",
    "    irregular_vibration_features_names = ['F0_Std','rpde']\n",
    "    increased_noise_features_names = ['hnr','dfa']\n",
    "else:\n",
    "    irregular_vibration_features_names = ['F0_Std']\n",
    "    increased_noise_features_names = ['hnr']\n",
    "vocal_tremor_features_names = ['frequency tremor intensity index (FTrI)','amplitude tremor intensity index (ATrI)']\n",
    "articulatory_features_names = ['Mean_Of_Std_mfcc','Mean_Of_Std_delta_mfcc']\n",
    "all_features_names = airflow_features_names+aperiodicity_features_names+signal_perturbation_features_names\n",
    "all_features_names += irregular_vibration_features_names+increased_noise_features_names\n",
    "all_features_names += vocal_tremor_features_names+articulatory_features_names\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "13"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classic_phonatory_features = {}\n",
    "for file_name in extracted_features.keys():\n",
    "    classic_phonatory_features[file_name] = {curr_feat: extracted_features[file_name][curr_feat] for curr_feat in all_features_names}\n",
    "\n",
    "classic_phonatory_features_df = pd.DataFrame.from_dict(classic_phonatory_features,orient='index')\n",
    "len(classic_phonatory_features_df.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "## Step 4 : Description and MPS visualisation \n",
    "Now, we can obtain some description of the dataset and visualise the Modulation Power Spectrum features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>Maximum_phonation_time</th>\n",
       "      <th>Maximum_phonation_time_until_voice_break</th>\n",
       "      <th>Number of voice breaks</th>\n",
       "      <th>Degree of voice breaks</th>\n",
       "      <th>Fraction of locally unvoiced frames</th>\n",
       "      <th>jitter</th>\n",
       "      <th>shimmer</th>\n",
       "      <th>F0_Std</th>\n",
       "      <th>hnr</th>\n",
       "      <th>frequency tremor intensity index (FTrI)</th>\n",
       "      <th>amplitude tremor intensity index (ATrI)</th>\n",
       "      <th>Mean_Of_Std_mfcc</th>\n",
       "      <th>Mean_Of_Std_delta_mfcc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>5.00000</td>\n",
       "      <td>5.000000</td>\n",
       "      <td>5.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>12.029061</td>\n",
       "      <td>5.673669</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>3.223400</td>\n",
       "      <td>8.963200</td>\n",
       "      <td>0.006281</td>\n",
       "      <td>0.070812</td>\n",
       "      <td>21.604178</td>\n",
       "      <td>15.761006</td>\n",
       "      <td>6.428500</td>\n",
       "      <td>25.54940</td>\n",
       "      <td>7.251858</td>\n",
       "      <td>1.238376</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>4.303302</td>\n",
       "      <td>5.649126</td>\n",
       "      <td>4.636809</td>\n",
       "      <td>5.975897</td>\n",
       "      <td>4.297089</td>\n",
       "      <td>0.002058</td>\n",
       "      <td>0.029557</td>\n",
       "      <td>28.267693</td>\n",
       "      <td>5.024811</td>\n",
       "      <td>5.323764</td>\n",
       "      <td>3.89228</td>\n",
       "      <td>1.067679</td>\n",
       "      <td>0.088678</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>6.652517</td>\n",
       "      <td>0.300000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4.209000</td>\n",
       "      <td>0.003603</td>\n",
       "      <td>0.034545</td>\n",
       "      <td>3.090616</td>\n",
       "      <td>7.238912</td>\n",
       "      <td>2.694000</td>\n",
       "      <td>21.57700</td>\n",
       "      <td>6.417114</td>\n",
       "      <td>1.124982</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>9.260408</td>\n",
       "      <td>0.300000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>6.798000</td>\n",
       "      <td>0.005233</td>\n",
       "      <td>0.059730</td>\n",
       "      <td>3.672771</td>\n",
       "      <td>16.570320</td>\n",
       "      <td>3.894750</td>\n",
       "      <td>21.66300</td>\n",
       "      <td>6.613118</td>\n",
       "      <td>1.179904</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>12.553696</td>\n",
       "      <td>6.652517</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.425000</td>\n",
       "      <td>7.151000</td>\n",
       "      <td>0.006411</td>\n",
       "      <td>0.067890</td>\n",
       "      <td>10.215859</td>\n",
       "      <td>16.761564</td>\n",
       "      <td>4.346000</td>\n",
       "      <td>26.47800</td>\n",
       "      <td>6.754651</td>\n",
       "      <td>1.238235</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>13.815828</td>\n",
       "      <td>7.300000</td>\n",
       "      <td>3.000000</td>\n",
       "      <td>1.867000</td>\n",
       "      <td>11.722000</td>\n",
       "      <td>0.007037</td>\n",
       "      <td>0.076147</td>\n",
       "      <td>20.438389</td>\n",
       "      <td>17.670435</td>\n",
       "      <td>6.879750</td>\n",
       "      <td>27.45900</td>\n",
       "      <td>7.441609</td>\n",
       "      <td>1.310514</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>17.862857</td>\n",
       "      <td>13.815828</td>\n",
       "      <td>11.000000</td>\n",
       "      <td>13.825000</td>\n",
       "      <td>14.936000</td>\n",
       "      <td>0.009123</td>\n",
       "      <td>0.115750</td>\n",
       "      <td>70.603256</td>\n",
       "      <td>20.563798</td>\n",
       "      <td>14.328000</td>\n",
       "      <td>30.57000</td>\n",
       "      <td>9.032798</td>\n",
       "      <td>1.338245</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Maximum_phonation_time  Maximum_phonation_time_until_voice_break  \\\n",
       "count                5.000000                                  5.000000   \n",
       "mean                12.029061                                  5.673669   \n",
       "std                  4.303302                                  5.649126   \n",
       "min                  6.652517                                  0.300000   \n",
       "25%                  9.260408                                  0.300000   \n",
       "50%                 12.553696                                  6.652517   \n",
       "75%                 13.815828                                  7.300000   \n",
       "max                 17.862857                                 13.815828   \n",
       "\n",
       "       Number of voice breaks  Degree of voice breaks  \\\n",
       "count                5.000000                5.000000   \n",
       "mean                 3.000000                3.223400   \n",
       "std                  4.636809                5.975897   \n",
       "min                  0.000000                0.000000   \n",
       "25%                  0.000000                0.000000   \n",
       "50%                  1.000000                0.425000   \n",
       "75%                  3.000000                1.867000   \n",
       "max                 11.000000               13.825000   \n",
       "\n",
       "       Fraction of locally unvoiced frames    jitter   shimmer     F0_Std  \\\n",
       "count                             5.000000  5.000000  5.000000   5.000000   \n",
       "mean                              8.963200  0.006281  0.070812  21.604178   \n",
       "std                               4.297089  0.002058  0.029557  28.267693   \n",
       "min                               4.209000  0.003603  0.034545   3.090616   \n",
       "25%                               6.798000  0.005233  0.059730   3.672771   \n",
       "50%                               7.151000  0.006411  0.067890  10.215859   \n",
       "75%                              11.722000  0.007037  0.076147  20.438389   \n",
       "max                              14.936000  0.009123  0.115750  70.603256   \n",
       "\n",
       "             hnr  frequency tremor intensity index (FTrI)  \\\n",
       "count   5.000000                                 4.000000   \n",
       "mean   15.761006                                 6.428500   \n",
       "std     5.024811                                 5.323764   \n",
       "min     7.238912                                 2.694000   \n",
       "25%    16.570320                                 3.894750   \n",
       "50%    16.761564                                 4.346000   \n",
       "75%    17.670435                                 6.879750   \n",
       "max    20.563798                                14.328000   \n",
       "\n",
       "       amplitude tremor intensity index (ATrI)  Mean_Of_Std_mfcc  \\\n",
       "count                                  5.00000          5.000000   \n",
       "mean                                  25.54940          7.251858   \n",
       "std                                    3.89228          1.067679   \n",
       "min                                   21.57700          6.417114   \n",
       "25%                                   21.66300          6.613118   \n",
       "50%                                   26.47800          6.754651   \n",
       "75%                                   27.45900          7.441609   \n",
       "max                                   30.57000          9.032798   \n",
       "\n",
       "       Mean_Of_Std_delta_mfcc  \n",
       "count                5.000000  \n",
       "mean                 1.238376  \n",
       "std                  0.088678  \n",
       "min                  1.124982  \n",
       "25%                  1.179904  \n",
       "50%                  1.238235  \n",
       "75%                  1.310514  \n",
       "max                  1.338245  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classic_phonatory_features_df.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def plot_mps(mps,wt,wf, DBNOISE=50):\n",
    "    plt.figure()\n",
    "    plt.clf()\n",
    "    cmap = plt.get_cmap('jet')\n",
    "    ex = (wt.min(), wt.max(), wf.min() * 1e3,\n",
    "          wf.max() * 1e3)\n",
    "    logMPS = 10.0 * np.log10(mps)\n",
    "    maxMPS = logMPS.max()\n",
    "    minMPS = maxMPS - DBNOISE\n",
    "    logMPS[logMPS < minMPS] = minMPS\n",
    "    plt.imshow(\n",
    "        logMPS,\n",
    "        interpolation='nearest',\n",
    "        aspect='auto',\n",
    "        origin='lower',\n",
    "        cmap=cmap,\n",
    "        extent=ex)\n",
    "    plt.ylabel('Spectral Frequency (Cycles/KHz)')\n",
    "    plt.xlabel('Temporal Frequency (Hz)')\n",
    "    plt.colorbar()\n",
    "    plt.ylim((0, wf.max() * 1e3))\n",
    "    plt.title('Modulation Power Spectrum')\n",
    "    plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "file_name = 'aaa_maureen.wav'\n",
    "plot_mps(extracted_features[file_name]['mps']['mps'], extracted_features[file_name]['mps']['mps_wt'],extracted_features[file_name]['mps']['mps_wf'])\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "file_name = 'ah_mathieu.wav'\n",
    "plot_mps(extracted_features[file_name]['mps']['mps'], extracted_features[file_name]['mps']['mps_wt'],extracted_features[file_name]['mps']['mps_wf'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}