{ "cells": [ { "cell_type": "code", "execution_count": 1, "id": "completed-phase", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Populating the interactive namespace from numpy and matplotlib\n" ] } ], "source": [ "%pylab inline" ] }, { "cell_type": "code", "execution_count": 2, "id": "digital-ticket", "metadata": {}, "outputs": [], "source": [ "import os\n", "os.environ['CUDA_VISIBLE_DEVICES'] = ''" ] }, { "cell_type": "code", "execution_count": 268, "id": "provincial-accounting", "metadata": {}, "outputs": [], "source": [ "import re\n", "import pandas as pd\n", "import numpy as np\n", "import librosa\n", "from scipy.io import wavfile\n", "import librosa.display\n", "from pydub import AudioSegment\n", "from IPython.display import display_html\n", "import torch\n", "\n", "\n", "def _convert_array(arr, force_int=False, force_float=False):\n", " if isinstance(arr, torch.Tensor):\n", " arr = arr.cpu().detach().numpy()\n", " assert(len(arr.shape) in (1, 2))\n", " if len(arr.shape) == 2:\n", " assert(arr.shape[0] == 1)\n", " arr = arr[0]\n", " if force_int and \"float\" in str(arr.dtype):\n", " if np.abs(arr).max() > 1:\n", " raise AssertionError(\"input seems too large\")\n", " arr = np.floor(arr.copy() * 32768).astype(np.int16)\n", " if force_float and \"int\" in str(arr.dtype):\n", " arr = arr.copy().astype(np.float32) / 32768\n", " return arr\n", "\n", "\n", "def plot_ft(arr, window_size=1024, hop_length=512, n_mels=128):\n", " arr = _convert_array(arr, force_float=True)\n", " window = np.hanning(window_size)\n", " stft= librosa.core.spectrum.stft(arr, n_fft=window_size, hop_length=hop_length, window=window)\n", " out = 2 * np.abs(stft) / np.sum(window)\n", " data = librosa.amplitude_to_db(out, ref=np.max)\n", "\n", " plt.figure(figsize=(16, 10))\n", " librosa.display.specshow(data, x_axis='time', y_axis='mel', sr=16000)\n", " plt.set_cmap('afmhot')\n", "\n", "\n", "def play_array(arr):\n", " arr = _convert_array(arr, force_int=True)\n", " display_html(AudioSegment(\n", " arr.tobytes(), \n", " frame_rate=16000,\n", " sample_width=2, \n", " channels=1\n", " ))" ] }, { "cell_type": "code", "execution_count": null, "id": "enhanced-family", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "incomplete-wedding", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "markdown", "id": "recovered-freight", "metadata": {}, "source": [ "## get transfer function from parallel recording" ] }, { "cell_type": "code", "execution_count": 4, "id": "sexual-luther", "metadata": {}, "outputs": [], "source": [ "# https://www.dsprelated.com/freebooks/filters/Time_Domain_Filter_Estimation.html" ] }, { "cell_type": "code", "execution_count": 257, "id": "divided-denmark", "metadata": {}, "outputs": [], "source": [ "import scipy.linalg\n", "from scipy import signal\n", "import warnings\n", "\n", "def learn_equalizer(signal_x, signal_y, detail_level):\n", " signal_x = _convert_array(signal_x, force_float=True)\n", " signal_y = _convert_array(signal_y, force_float=True)\n", " # Create Toeplitz Matrix\n", " padding = np.zeros(detail_level, signal_x.dtype)\n", " first_col = np.r_[signal_x, padding]\n", " first_row = np.r_[signal_x[0], padding]\n", " X = scipy.linalg.toeplitz(first_col, first_row)\n", " # Implement the formula for finding the optimal solution numerically\n", " X = X[0:len(X)-detail_level,]\n", " h = np.dot(X.T, X)\n", " h = scipy.linalg.inv(h)\n", " h = np.dot(h, X.T)\n", " equalizer = np.dot(h, signal_y)\n", " if np.abs(equalizer).max() > 5:\n", " warnings.warn(\"IR seems large\")\n", " return equalizer\n", "\n", "def apply_transfer(signal_x, h):\n", " signal_x = _convert_array(signal_x, force_float=True)\n", " # Use Toeplitz Matrix\n", " detail_level = len(h) - 1\n", " padding = np.zeros(detail_level, signal_x.dtype)\n", " first_col = np.r_[signal_x, padding]\n", " first_row = np.r_[signal_x[0], padding]\n", " X = scipy.linalg.toeplitz(first_col, first_row)\n", " # apply\n", " y = np.dot(X, h)\n", " # equivalently we can just convolve\n", "# y = signal.fftconvolve(signal_x, h, mode=\"full\")\n", " if np.abs(y).max() > 1:\n", " raise AssertionError(\"output seems too large\")\n", " return y[:signal_x.shape[0]]" ] }, { "cell_type": "code", "execution_count": 420, "id": "brave-stadium", "metadata": {}, "outputs": [], "source": [ "# _, noisy = wavfile.read('tmp/pairs/ring_noisy.wav')\n", "# _, clean = wavfile.read('tmp/pairs/ring_clean.wav')\n", "\n", "_, noisy = wavfile.read('tmp/pairs/company2_noisy.wav')\n", "_, clean = wavfile.read('tmp/pairs/company2_clean.wav')\n", "\n", "# _, noisy = wavfile.read('tmp/pairs/company3_noisy.wav')\n", "# _, clean = wavfile.read('tmp/pairs/company3_clean.wav')" ] }, { "cell_type": "code", "execution_count": 421, "id": "beautiful-sociology", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "play_array(clean)" ] }, { "cell_type": "code", "execution_count": 422, "id": "athletic-series", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "play_array(noisy)" ] }, { "cell_type": "code", "execution_count": 423, "id": "frequent-major", "metadata": {}, "outputs": [], "source": [ "# make clean -> noisy\n", "ir = learn_equalizer(clean, noisy, 5000)\n", "sim_noisy = apply_transfer(clean, ir)[:len(clean)]" ] }, { "cell_type": "code", "execution_count": 424, "id": "latest-lucas", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[]" ] }, "execution_count": 424, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.plot(ir)" ] }, { "cell_type": "code", "execution_count": 418, "id": "major-groove", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "play_array(sim_noisy)" ] }, { "cell_type": "code", "execution_count": 259, "id": "obvious-anchor", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/home/georg/venvs/py37/lib/python3.7/site-packages/ipykernel_launcher.py:20: UserWarning: IR seems large\n" ] } ], "source": [ "# do reverse\n", "hl = len(clean) // 2\n", "ir = learn_equalizer(noisy[:hl], clean[:hl], 5000)\n", "sim_clean = apply_transfer(noisy[hl:], ir)[:hl]" ] }, { "cell_type": "code", "execution_count": 261, "id": "stretch-translator", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[]" ] }, "execution_count": 261, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.plot(sim_noisy)" ] }, { "cell_type": "code", "execution_count": null, "id": "juvenile-academy", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": 260, "id": "fuzzy-commission", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[]" ] }, "execution_count": 260, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.plot(sim_clean)" ] }, { "cell_type": "code", "execution_count": null, "id": "acquired-trailer", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "electoral-mumbai", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "prescribed-hebrew", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "enormous-polymer", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "markdown", "id": "regulation-chess", "metadata": {}, "source": [ "#### try out of sample reverse" ] }, { "cell_type": "code", "execution_count": 402, "id": "smart-arrival", "metadata": {}, "outputs": [], "source": [ "_, noisy = wavfile.read('tmp/pairs/company2_noisy.wav')\n", "_, clean = wavfile.read('tmp/pairs/company2_clean.wav')\n", "noisy = _convert_array(noisy, force_float=True)\n", "clean = _convert_array(clean, force_float=True)" ] }, { "cell_type": "code", "execution_count": 403, "id": "lovely-techno", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "play_array(clean)" ] }, { "cell_type": "code", "execution_count": 404, "id": "considered-broadway", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "play_array(noisy)" ] }, { "cell_type": "code", "execution_count": 405, "id": "entertaining-reader", "metadata": {}, "outputs": [], "source": [ "# learn trafo on first half, apply on second\n", "ns = clean.shape[0]\n", "h = learn_equalizer(clean[:ns//2], noisy[:ns//2], 5000)\n", "sim_noisy = apply_transfer(clean[:ns//2], h)" ] }, { "cell_type": "code", "execution_count": 407, "id": "square-scheduling", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[]" ] }, "execution_count": 407, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.plot(sim_noisy)" ] }, { "cell_type": "code", "execution_count": 406, "id": "joint-speaking", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "play_array(sim_noisy)" ] }, { "cell_type": "code", "execution_count": 408, "id": "united-watson", "metadata": {}, "outputs": [], "source": [ "h = learn_equalizer(sim_noisy, clean[:ns//2], 5000)\n", "sim_clean = apply_transfer(noisy[ns//2:], h)" ] }, { "cell_type": "code", "execution_count": 409, "id": "responsible-virginia", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[]" ] }, "execution_count": 409, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "plt.plot(sim_clean)" ] }, { "cell_type": "code", "execution_count": 410, "id": "small-niger", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "play_array(sim_clean)" ] }, { "cell_type": "code", "execution_count": null, "id": "binding-lightweight", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "crude-coordinate", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "identical-cholesterol", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "markdown", "id": "infectious-parent", "metadata": {}, "source": [ "#### apply to new audio" ] }, { "cell_type": "code", "execution_count": 14, "id": "formal-parcel", "metadata": {}, "outputs": [], "source": [ "_, a = wavfile.read('tmp/apple_short.wav')" ] }, { "cell_type": "code", "execution_count": 15, "id": "sound-chair", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "play_array(a)" ] }, { "cell_type": "code", "execution_count": 16, "id": "unable-convert", "metadata": {}, "outputs": [], "source": [ "b = apply_transfer(a, h)" ] }, { "cell_type": "code", "execution_count": 17, "id": "heard-manhattan", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "play_array(b)" ] }, { "cell_type": "code", "execution_count": null, "id": "duplicate-details", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "royal-africa", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "military-tamil", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "sustainable-tissue", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "caroline-forum", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "markdown", "id": "annoying-classroom", "metadata": {}, "source": [ "## DNS 2021" ] }, { "cell_type": "code", "execution_count": null, "id": "broad-romance", "metadata": {}, "outputs": [], "source": [ "import os" ] }, { "cell_type": "code", "execution_count": 185, "id": "vietnamese-baseline", "metadata": {}, "outputs": [], "source": [ "# main_dir = \"/data-ssd-3/georg/data/DNS-Challenge/datasets/impulse_responses/SLR28/\"\n", "main_dir = \"/data-ssd-3/georg/data/libri_slr28/\"" ] }, { "cell_type": "code", "execution_count": 200, "id": "emerging-glossary", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "325 rir metas found\n", "92 isotropic noise metas found\n" ] } ], "source": [ "ir_dir = main_dir + \"RIRS_NOISES/real_rirs_isotropic_noises/\"\n", "\n", "with open(ir_dir + \"rir_list\") as f:\n", " rir_list = [e.split() for e in f.read().strip().split('\\n')]\n", " rir_list = [(row[1], row[3], row[4]) for row in rir_list]\n", "rir_df = pd.DataFrame(rir_list, columns=[\"rir-id\", \"room-id\", \"path\"])\n", "rir_df['path'] = rir_df['path'].str.split('/').str[-1]\n", "print(rir_df.shape[0], 'rir metas found')\n", "\n", "with open(ir_dir + \"noise_list\") as f:\n", " noise_list = [e.split() for e in f.read().strip().split('\\n')]\n", " noise_list = [(row[1], row[3], row[5], row[6]) for row in noise_list]\n", "noise_df = pd.DataFrame(noise_list, columns=[\"noise-id\", \"noise-type\", \"room-linkage\", \"path\"])\n", "noise_df['path'] = noise_df['path'].str.split('/').str[-1]\n", "print(noise_df.shape[0], 'isotropic noise metas found')\n", "\n", "wav_files = [e for e in os.listdir(ir_dir) if \"wav\" in e]\n", "assert((set(rir_df['path']) | set(noise_df['path'])) == set(wav_files))\n", "# 325 rir metas found\n", "# 92 isotropic noise metas found" ] }, { "cell_type": "code", "execution_count": 187, "id": "preliminary-event", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
rir-idroom-idpath
000001RVB2014_largeroom1RVB2014_type1_rir_largeroom1_far_angla.wav
100002RVB2014_largeroom1RVB2014_type1_rir_largeroom1_far_anglb.wav
200003RVB2014_largeroom1RVB2014_type1_rir_largeroom1_near_angla.wav
\n", "
" ], "text/plain": [ " rir-id room-id path\n", "0 00001 RVB2014_largeroom1 RVB2014_type1_rir_largeroom1_far_angla.wav\n", "1 00002 RVB2014_largeroom1 RVB2014_type1_rir_largeroom1_far_anglb.wav\n", "2 00003 RVB2014_largeroom1 RVB2014_type1_rir_largeroom1_near_angla.wav" ] }, "execution_count": 187, "metadata": {}, "output_type": "execute_result" } ], "source": [ "rir_df.head(3)" ] }, { "cell_type": "code", "execution_count": 188, "id": "western-richards", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
noise-idnoise-typeroom-linkagepath
000001isotropicRVB2014_largeroom1RVB2014_type1_noise_largeroom1_1.wav
100002isotropicRVB2014_largeroom1RVB2014_type1_noise_largeroom1_10.wav
200003isotropicRVB2014_largeroom1RVB2014_type1_noise_largeroom1_2.wav
\n", "
" ], "text/plain": [ " noise-id noise-type room-linkage \\\n", "0 00001 isotropic RVB2014_largeroom1 \n", "1 00002 isotropic RVB2014_largeroom1 \n", "2 00003 isotropic RVB2014_largeroom1 \n", "\n", " path \n", "0 RVB2014_type1_noise_largeroom1_1.wav \n", "1 RVB2014_type1_noise_largeroom1_10.wav \n", "2 RVB2014_type1_noise_largeroom1_2.wav " ] }, "execution_count": 188, "metadata": {}, "output_type": "execute_result" } ], "source": [ "noise_df.head(3)" ] }, { "cell_type": "code", "execution_count": 207, "id": "chinese-pasta", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "843 pointsource noise metas found\n" ] } ], "source": [ "# point noises\n", "\n", "ir_dir = main_dir + \"RIRS_NOISES/pointsource_noises/\"\n", "\n", "with open(ir_dir + \"noise_list\") as f:\n", " noise_list_2 = [e.split() for e in f.read().strip().split('\\n')]\n", " noise_list_2 = [(row[1], row[3], row[5], row[6]) for row in noise_list_2]\n", "noise_2_df = pd.DataFrame(noise_list_2, columns=[\"noise-id\", \"noise-type\", \"bg-fg-type\", \"path\"])\n", "noise_2_df['path'] = noise_2_df['path'].str.split('/').str[-1]\n", "print(noise_2_df.shape[0], 'pointsource noise metas found')\n", "\n", "wav_files = [e for e in os.listdir(ir_dir) if \"wav\" in e]\n", "assert(len(wav_files) == noise_2_df.shape[0])\n", "# 843 noise metas found" ] }, { "cell_type": "code", "execution_count": 208, "id": "unauthorized-savage", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
noise-idnoise-typebg-fg-typepath
0noise-free-sound-0000point-sourceforegroundnoise-free-sound-0000.wav
1noise-free-sound-0001point-sourceforegroundnoise-free-sound-0001.wav
2noise-free-sound-0002point-sourceforegroundnoise-free-sound-0002.wav
3noise-free-sound-0003point-sourceforegroundnoise-free-sound-0003.wav
4noise-free-sound-0004point-sourceforegroundnoise-free-sound-0004.wav
\n", "
" ], "text/plain": [ " noise-id noise-type bg-fg-type path\n", "0 noise-free-sound-0000 point-source foreground noise-free-sound-0000.wav\n", "1 noise-free-sound-0001 point-source foreground noise-free-sound-0001.wav\n", "2 noise-free-sound-0002 point-source foreground noise-free-sound-0002.wav\n", "3 noise-free-sound-0003 point-source foreground noise-free-sound-0003.wav\n", "4 noise-free-sound-0004 point-source foreground noise-free-sound-0004.wav" ] }, "execution_count": 208, "metadata": {}, "output_type": "execute_result" } ], "source": [ "noise_2_df.head()" ] }, { "cell_type": "code", "execution_count": 209, "id": "impossible-certificate", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "foreground 806\n", "background 37\n", "Name: bg-fg-type, dtype: int64" ] }, "execution_count": 209, "metadata": {}, "output_type": "execute_result" } ], "source": [ "noise_2_df[\"bg-fg-type\"].value_counts()" ] }, { "cell_type": "code", "execution_count": null, "id": "infectious-bankruptcy", "metadata": {}, "outputs": [], "source": [ "# y = signal.fftconvolve(arr, rir, mode=\"full\")\n", "# y = y[:signal_x.shape[0]]" ] }, { "cell_type": "code", "execution_count": null, "id": "wrapped-camcorder", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "continued-summary", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "markdown", "id": "specified-option", "metadata": {}, "source": [ "## AEC 2021" ] }, { "cell_type": "code", "execution_count": 278, "id": "educational-sample", "metadata": {}, "outputs": [], "source": [ "main_dir = \"/data-ssd-3/georg/data/AEC-Challenge/datasets/\"" ] }, { "cell_type": "code", "execution_count": 288, "id": "accessible-martin", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
aliaschannelfrequencyRT60
0-2jLGNCgf0WDpKMY2iup7g1wideband2.050700
1-2jLGNCgf0WDpKMY2iup7g131.62282.244741
\n", "
" ], "text/plain": [ " alias channel frequency RT60\n", "0 -2jLGNCgf0WDpKMY2iup7g 1 wideband 2.050700\n", "1 -2jLGNCgf0WDpKMY2iup7g 1 31.6228 2.244741" ] }, "execution_count": 288, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# not sure what this is used for\n", "# \"latin1\", \"iso-8859-1\", \"cp1252\"\n", "df = pd.read_csv(main_dir + \"RIRs/AEC_challenge_ICASSP2021_RIRs.csv\", encoding=\"latin1\")\n", "df.head(2)" ] }, { "cell_type": "code", "execution_count": 336, "id": "prepared-neighborhood", "metadata": {}, "outputs": [], "source": [ "ir_dir = main_dir + \"real/\"\n", "\n", "wav_files = [e for e in os.listdir(ir_dir) if \"wav\" in e]\n", "data = [\n", " (\n", " re.search(r\"(.+)\\_(?:farend|nearend|doubletalk|sweep)\", e).group(1),\n", " re.search(r\"((?:farend|nearend|doubletalk|sweep).*?)\\.\", e).group(1),\n", " e\n", " ) for e in wav_files\n", "]\n", "df = pd.DataFrame(data, columns=[\"id\", \"type\", \"path\"])" ] }, { "cell_type": "code", "execution_count": 337, "id": "arranged-level", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "nearend_singletalk_mic 4182\n", "farend_singletalk_mic 3781\n", "farend_singletalk_lpb 3749\n", "farend_singletalk_with_movement_mic 3539\n", "farend_singletalk_with_movement_lpb 3537\n", "doubletalk_mic 3216\n", "doubletalk_lpb 3216\n", "doubletalk_with_movement_mic 3130\n", "doubletalk_with_movement_lpb 3128\n", "sweep_lpb 3055\n", "sweep_mic 3055\n", "Name: type, dtype: int64" ] }, "execution_count": 337, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df['type'].value_counts()" ] }, { "cell_type": "code", "execution_count": 383, "id": "intimate-victorian", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
idtypepath
10053nZS_z6nxz0az25RLUk_9ogdoubletalk_with_movement_micnZS_z6nxz0az25RLUk_9og_doubletalk_with_movemen...
\n", "
" ], "text/plain": [ " id type \\\n", "10053 nZS_z6nxz0az25RLUk_9og doubletalk_with_movement_mic \n", "\n", " path \n", "10053 nZS_z6nxz0az25RLUk_9og_doubletalk_with_movemen... " ] }, "execution_count": 383, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df.sample(1)" ] }, { "cell_type": "code", "execution_count": 384, "id": "dangerous-prophet", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
idtypepath
2681nZS_z6nxz0az25RLUk_9ogfarend_singletalk_lpbnZS_z6nxz0az25RLUk_9og_farend_singletalk_lpb.wav
5012nZS_z6nxz0az25RLUk_9ogfarend_singletalk_with_movement_lpbnZS_z6nxz0az25RLUk_9og_farend_singletalk_with_...
10053nZS_z6nxz0az25RLUk_9ogdoubletalk_with_movement_micnZS_z6nxz0az25RLUk_9og_doubletalk_with_movemen...
16666nZS_z6nxz0az25RLUk_9ogfarend_singletalk_with_movement_micnZS_z6nxz0az25RLUk_9og_farend_singletalk_with_...
21234nZS_z6nxz0az25RLUk_9ogdoubletalk_with_movement_lpbnZS_z6nxz0az25RLUk_9og_doubletalk_with_movemen...
23987nZS_z6nxz0az25RLUk_9ogdoubletalk_micnZS_z6nxz0az25RLUk_9og_doubletalk_mic.wav
25650nZS_z6nxz0az25RLUk_9ogdoubletalk_lpbnZS_z6nxz0az25RLUk_9og_doubletalk_lpb.wav
27019nZS_z6nxz0az25RLUk_9ognearend_singletalk_micnZS_z6nxz0az25RLUk_9og_nearend_singletalk_mic.wav
32993nZS_z6nxz0az25RLUk_9ogfarend_singletalk_micnZS_z6nxz0az25RLUk_9og_farend_singletalk_mic.wav
36474nZS_z6nxz0az25RLUk_9ogsweep_micnZS_z6nxz0az25RLUk_9og_sweep_mic.wav
37270nZS_z6nxz0az25RLUk_9ogsweep_lpbnZS_z6nxz0az25RLUk_9og_sweep_lpb.wav
\n", "
" ], "text/plain": [ " id type \\\n", "2681 nZS_z6nxz0az25RLUk_9og farend_singletalk_lpb \n", "5012 nZS_z6nxz0az25RLUk_9og farend_singletalk_with_movement_lpb \n", "10053 nZS_z6nxz0az25RLUk_9og doubletalk_with_movement_mic \n", "16666 nZS_z6nxz0az25RLUk_9og farend_singletalk_with_movement_mic \n", "21234 nZS_z6nxz0az25RLUk_9og doubletalk_with_movement_lpb \n", "23987 nZS_z6nxz0az25RLUk_9og doubletalk_mic \n", "25650 nZS_z6nxz0az25RLUk_9og doubletalk_lpb \n", "27019 nZS_z6nxz0az25RLUk_9og nearend_singletalk_mic \n", "32993 nZS_z6nxz0az25RLUk_9og farend_singletalk_mic \n", "36474 nZS_z6nxz0az25RLUk_9og sweep_mic \n", "37270 nZS_z6nxz0az25RLUk_9og sweep_lpb \n", "\n", " path \n", "2681 nZS_z6nxz0az25RLUk_9og_farend_singletalk_lpb.wav \n", "5012 nZS_z6nxz0az25RLUk_9og_farend_singletalk_with_... \n", "10053 nZS_z6nxz0az25RLUk_9og_doubletalk_with_movemen... \n", "16666 nZS_z6nxz0az25RLUk_9og_farend_singletalk_with_... \n", "21234 nZS_z6nxz0az25RLUk_9og_doubletalk_with_movemen... \n", "23987 nZS_z6nxz0az25RLUk_9og_doubletalk_mic.wav \n", "25650 nZS_z6nxz0az25RLUk_9og_doubletalk_lpb.wav \n", "27019 nZS_z6nxz0az25RLUk_9og_nearend_singletalk_mic.wav \n", "32993 nZS_z6nxz0az25RLUk_9og_farend_singletalk_mic.wav \n", "36474 nZS_z6nxz0az25RLUk_9og_sweep_mic.wav \n", "37270 nZS_z6nxz0az25RLUk_9og_sweep_lpb.wav " ] }, "execution_count": 384, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sub_df = df[df['id'] == \"nZS_z6nxz0az25RLUk_9og\"]\n", "sub_df" ] }, { "cell_type": "code", "execution_count": 385, "id": "encouraging-clarity", "metadata": {}, "outputs": [], "source": [ "# not all are correctly recorded, need sanity checks\n", "# unclear what's noisy and whats clean\n", "\n", "# ~ 10 hours of PAIRS:\n", "\n", "# farend_singletalk_lpb\n", "# farend_singletalk_mic\n", "\n", "# farend_singletalk_with_movement_lpb\n", "# farend_singletalk_with_movement_mic\n", "\n", "# sweep_lpb\n", "# sweep_mic" ] }, { "cell_type": "code", "execution_count": 386, "id": "mexican-greeting", "metadata": {}, "outputs": [], "source": [ "# TODO: do we know noisy vs clean?" ] }, { "cell_type": "code", "execution_count": 389, "id": "stock-truck", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ], "text/plain": [ "" ] }, "execution_count": 389, "metadata": {}, "output_type": "execute_result" } ], "source": [ "n = \"farend_singletalk_lpb\"\n", "# n = \"farend_singletalk_with_movement_lpb\"\n", "\n", "AudioSegment.from_wav(ir_dir + sub_df[sub_df['type'] == n].iloc[0]['path'])" ] }, { "cell_type": "code", "execution_count": 390, "id": "heard-mirror", "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " " ], "text/plain": [ "" ] }, "execution_count": 390, "metadata": {}, "output_type": "execute_result" } ], "source": [ "n = \"farend_singletalk_mic\"\n", "# n = \"farend_singletalk_with_movement_mic\"\n", "\n", "AudioSegment.from_wav(ir_dir + sub_df[sub_df['type'] == n].iloc[0]['path'])" ] }, { "cell_type": "code", "execution_count": null, "id": "broken-spotlight", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "female-matter", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "genuine-missouri", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "marked-twins", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "italian-military", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "medical-alignment", "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "id": "sudden-target", "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": "3.7.12" } }, "nbformat": 4, "nbformat_minor": 5 }