SR-SpeechSignalProc.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "SR-SpeechSignalProc.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyO1+bYeqJjGllb3NSsixuYc",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/MSaber7/Machine-Learning/blob/master/SR-SpeechSignalProc.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C14GAZ_y8qiw",
        "colab_type": "text"
      },
      "source": [
        "# SR | Lab Two: Speech Signal Processing\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eXGYQRbO8ue7",
        "colab_type": "text"
      },
      "source": [
        "In this lab, you will write the core functions necessary to perform feature extraction on audio waveforms. Your program will convert an audio file to a sequence of log mel frequency filterbank (\"FBANK\") coefficients.\n",
        "The basic steps in features extraction are:\n",
        "Dithering & pre-emphasis of the waveform\n",
        "Dividing the signal into overlapping segments or frames\n",
        "For each frame of audio:\n",
        "Windowing the frame\n",
        "Computing the magnitude spectrum of the frame\n",
        "Applying the mel filterbank to the spectrum to create mel filterbank coefficient\n",
        "Applying a logarithm operation to the mel filterbank coefficient\n",
        "File speech_sigproc.py contains methods for performing these steps. Functions for dithering (1a), framing (2) and windowing (3a) are already implemented as well as method process_utterance which applies all needed methods one after another. Your task is to:\n",
        "Implement pre-emphasis of the signal as method pre_emphasize \n",
        "Implement computation of magnitude spectrum as method frames_to_magspec. (Note: the variable fft_size is useful there)\n",
        "Implement the application of mel filterbank and logarithm operation as method magspec_to_fbank (Note: the coefficients of the melt filterbank are computed and saved as mel_filterbank 2-D array. \n",
        "\n",
        "The lab also contain following files:\n",
        "1272-128104-0000.flac -- audio file from LibriSpeech library, which is a clean sample of english speech\n",
        "1272-128104-0000.feat -- file which contains data about expected result of feature extraction for mentioned audio file \n",
        "htk_featio.py -- additional file for extracting data from .feat file\n",
        "plotter.py -- script which consistently performs following steps:\n",
        "Reads the wav file and extracts features from it with method process_utterance from speech_sigproc.py \n",
        "Plot the original waveform, the mel filterbank, the expected and gained results of feature extraction\n",
        "Computes per-element absolute error of the feature extraction (Note: if you did everything correctly, the expected error is less than 1e-08).\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p8kWheJa8p5X",
        "colab_type": "text"
      },
      "source": [
        "# Library"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E1NY4TEu4P-R",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "7fc4837a-01e6-4c68-a7cb-6d2b765c4596"
      },
      "source": [
        "!pip install SoundFile"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting SoundFile\n",
            "  Downloading https://files.pythonhosted.org/packages/eb/f2/3cbbbf3b96fb9fa91582c438b574cff3f45b29c772f94c400e2c99ef5db9/SoundFile-0.10.3.post1-py2.py3-none-any.whl\n",
            "Requirement already satisfied: cffi>=1.0 in /usr/local/lib/python3.6/dist-packages (from SoundFile) (1.13.2)\n",
            "Requirement already satisfied: pycparser in /usr/local/lib/python3.6/dist-packages (from cffi>=1.0->SoundFile) (2.19)\n",
            "Installing collected packages: SoundFile\n",
            "Successfully installed SoundFile-0.10.3.post1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BQszCSzc-qMB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import os\n",
        "import soundfile as sf\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import htk_featio as htk\n",
        "import speech_sigproc as sp\n",
        "from IPython.display import Image, display\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6KA3BsAfDTKr",
        "colab_type": "text"
      },
      "source": [
        "### speech_sigproc"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E6VanYI1DIwU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class FrontEnd:\n",
        "\n",
        "    def __init__(self, samp_rate=16000, frame_duration=0.025, frame_shift=0.010, preemphasis=0.97,\n",
        "                 num_mel=40, lo_freq=0, hi_freq=None, mean_norm_feat=False, mean_norm_wav=True, compute_stats=False):\n",
        "        self.samp_rate = samp_rate\n",
        "        self.win_size = int(np.floor(frame_duration * samp_rate))\n",
        "        self.win_shift = int(np.floor(frame_shift * samp_rate))\n",
        "        self.lo_freq = lo_freq\n",
        "        if (hi_freq == None):\n",
        "            self.hi_freq = samp_rate//2\n",
        "        else:\n",
        "            self.hi_freq = hi_freq\n",
        "\n",
        "        self.preemphasis = preemphasis\n",
        "        self.num_mel = num_mel\n",
        "        self.fft_size = 2\n",
        "        while (self.fft_size<self.win_size):\n",
        "            self.fft_size *= 2\n",
        "\n",
        "        self.hamwin = np.hamming(self.win_size)\n",
        "\n",
        "        self.make_mel_filterbank()\n",
        "        self.zero_mean_wav = mean_norm_wav\n",
        "        self.global_mean = np.zeros([num_mel])\n",
        "        self.global_var = np.zeros([num_mel])\n",
        "        self.global_frames = 0\n",
        "        self.compute_global_stats = compute_stats\n",
        "\n",
        "    # linear-scale frequency (Hz) to mel-scale frequency\n",
        "    def lin2mel(self,freq):\n",
        "        return 2595*np.log10(1+freq/700)\n",
        "\n",
        "    # mel-scale frequency to linear-scale frequency\n",
        "    def mel2lin(self,mel):\n",
        "        return (10**(mel/2595)-1)*700\n",
        "\n",
        "    def make_mel_filterbank(self):\n",
        "\n",
        "        lo_mel = self.lin2mel(self.lo_freq)\n",
        "        hi_mel = self.lin2mel(self.hi_freq)\n",
        "\n",
        "        # uniform spacing on mel scale\n",
        "        mel_freqs = np.linspace(lo_mel, hi_mel,self.num_mel+2)\n",
        "\n",
        "        # convert mel freqs to hertz and then to fft bins\n",
        "        bin_width = self.samp_rate/self.fft_size # typically 31.25 Hz, bin[0]=0 Hz, bin[1]=31.25 Hz,..., bin[256]=8000 Hz\n",
        "        mel_bins = np.floor(self.mel2lin(mel_freqs)/bin_width)\n",
        "\n",
        "        num_bins = self.fft_size//2 + 1\n",
        "        self.mel_filterbank = np.zeros([self.num_mel,num_bins])\n",
        "        for i in range(0,self.num_mel):\n",
        "            left_bin = int(mel_bins[i])\n",
        "            center_bin = int(mel_bins[i+1])\n",
        "            right_bin = int(mel_bins[i+2])\n",
        "            up_slope = 1/(center_bin-left_bin)\n",
        "            for j in range(left_bin,center_bin):\n",
        "                self.mel_filterbank[i,j] = (j - left_bin)*up_slope\n",
        "            down_slope = -1/(right_bin-center_bin)\n",
        "            for j in range(center_bin,right_bin):\n",
        "                self.mel_filterbank[i,j] = (j-right_bin)*down_slope\n",
        "\n",
        "    def plot_mel_matrix(self):\n",
        "        for i in range(0, self.num_mel):\n",
        "            plt.plot(self.mel_filterbank[i,:])\n",
        "        plt.show()\n",
        "\n",
        "    def dither(self, wav):\n",
        "        n = 2*np.random.rand(wav.shape[0])-1\n",
        "        n *= 1/(2**15)\n",
        "        return wav + n\n",
        "\n",
        "    # replace this with proper pre-emphasis filtering, using the self.preemphasis coefficient\n",
        "    def pre_emphasize(self, wav):\n",
        "        # apply pre-emphasis filtering on waveform\n",
        "        preemph_wav = []\n",
        "        return preemph_wav\n",
        "\n",
        "    def wav_to_frames(self, wav):\n",
        "        # wav1 = np.array(wav)\n",
        "        # only process whole frames\n",
        "        num_frames = int(np.floor((wav.shape[0] - self.win_size) / self.win_shift) + 1)\n",
        "        frames = np.zeros([self.win_size, num_frames])\n",
        "        for t in range(0, num_frames):\n",
        "            frame = wav[t * self.win_shift:t * self.win_shift + self.win_size]\n",
        "            if (self.zero_mean_wav):\n",
        "                frame = frame - np.mean(frame)\n",
        "            frames[:, t] = self.hamwin * frame\n",
        "        return frames\n",
        "\n",
        "    # for each frame (column of 2D array 'frames'), compute the magnitude spectrum using the fft\n",
        "    def frames_to_magspec(self, frames):\n",
        "        # compute the fft\n",
        "        # compute magnitude\n",
        "        magspec = []\n",
        "        return magspec\n",
        "\n",
        "    # for each frame(column of 2D array 'magspec'), compute the log mel spectrum, by applying the mel filterbank to the magnitude spectrum\n",
        "    def magspec_to_fbank(self, magspec):\n",
        "        # apply the mel filterbank\n",
        "        fbank = []\n",
        "        return fbank\n",
        "\n",
        "    def process_utterance(self, utterance):\n",
        "   #     plt.plot(utterance)\n",
        "   #     plt.show()\n",
        "        wav     = self.dither(utterance)\n",
        "        wav     = self.pre_emphasize(wav)\n",
        "        frames  = self.wav_to_frames(wav)\n",
        "        magspec = self.frames_to_magspec(frames)\n",
        "        fbank   = self.magspec_to_fbank(magspec)\n",
        "        return fbank"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kgPyD8gWDXg4",
        "colab_type": "text"
      },
      "source": [
        "### htk_featio "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Trp1nBs-DNSC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def write_htk_user_feat(x, name='filename'):\n",
        "    default_period = 100000 # assumes 0.010 ms frame shift\n",
        "    num_dim = x.shape[0]\n",
        "    num_frames = x.shape[1]\n",
        "    hdr = struct.pack(\n",
        "        '>iihh',  # the beginning '>' says write big-endian\n",
        "        num_frames,  # nSamples\n",
        "        default_period,  # samplePeriod\n",
        "        4*num_dim,  # 2 floats per feature\n",
        "        9)  # user features\n",
        "\n",
        "    out_file = open(name, 'wb')\n",
        "    out_file.write(hdr)\n",
        "\n",
        "    for t in range(0, num_frames):\n",
        "        frame = np.array(x[:,t],'f')\n",
        "        if sys.byteorder == 'little':\n",
        "            frame.byteswap(True)\n",
        "        frame.tofile(out_file)\n",
        "\n",
        "    out_file.close()\n",
        "\n",
        "def read_htk_user_feat(name='filename'):\n",
        "    f = open(name,'rb')\n",
        "    hdr = f.read(12)\n",
        "    num_samples, samp_period, samp_size, parm_kind = struct.unpack(\">IIHH\", hdr)\n",
        "    if parm_kind != 9:\n",
        "        raise RuntimeError('feature reading code only validated for USER feature type for this lab. There is other publicly available code for general purpose HTK feature file I/O\\n')\n",
        "\n",
        "    num_dim = samp_size//4\n",
        "\n",
        "    feat = np.zeros([num_samples, num_dim],dtype=float)\n",
        "    for t in range(num_samples):\n",
        "        feat[t,:] = np.array(struct.unpack('>' + ('f' * num_dim), f.read(samp_size)),dtype=float)\n",
        "\n",
        "    return feat\n",
        "\n",
        "\n",
        "def write_ascii_stats(x,name='filename'):\n",
        "    out_file = open(name,'w')\n",
        "    for t in range(0, x.shape[0]):\n",
        "        out_file.write(\"{0}\\n\".format(x[t]))\n",
        "    out_file.close()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uSB7Sngy-vWL",
        "colab_type": "text"
      },
      "source": [
        "# Plotter"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ezy8E1LiAjv8",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "71e2a6c9-67ca-4f64-ae96-67a59f0b1d45"
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qwIPNA-G-6bg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "4c6922ee-d54a-4f5f-de82-ad0c2ca4d5a2"
      },
      "source": [
        "%cd/content/drive/My Drive/Colab Notebooks/SR/Lab2"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/drive/My Drive/Colab Notebooks/SR/Lab2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ah2hBiqNABqT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m3jB0j6oABnB",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "70db86b1-1e63-47bc-b05d-d9a69f6c870e"
      },
      "source": [
        "wav_file='1272-128104-0000.flac'\n",
        "feat_file=('1272-128104-0000.feat')\n",
        "\n",
        "if not os.path.isfile(wav_file):\n",
        "    raise RuntimeError('input wav file is missing.')\n",
        "\n",
        "samp_rate = 16000\n",
        "\n",
        "x, s = sf.read(wav_file)\n",
        "if (s != samp_rate):\n",
        "    raise RuntimeError(\"WAV file is 16000 Hz, found {0}\".format(s))\n",
        "\n",
        "fe = sp.FrontEnd(samp_rate=samp_rate,mean_norm_feat=True)\n",
        "\n",
        "\n",
        "feat = fe.process_utterance(x)\n",
        "\n",
        "if not os.path.exists('fig'):\n",
        "    os.mkdir('fig')\n",
        "\n",
        "# plot waveform\n",
        "plt.plot(x)\n",
        "plt.title('waveform')\n",
        "plt.savefig('fig/waveform.png', bbox_inches='tight')\n",
        "plt.close()\n",
        "\n",
        "# plot mel filterbank\n",
        "for i in range(0, fe.num_mel):\n",
        "    plt.plot(fe.mel_filterbank[i, :])\n",
        "plt.title('mel filterbank')\n",
        "plt.savefig('fig/mel_filterbank.png', bbox_inches='tight')\n",
        "plt.close()\n",
        "\n",
        "# plot log mel spectrum (fbank)\n",
        "plt.imshow(feat, origin='lower', aspect=4) # flip the image so that vertical frequency axis goes from low to high\n",
        "plt.title('log mel filterbank features (fbank)')\n",
        "plt.savefig('fig/fbank.png', bbox_inches='tight')\n",
        "plt.close()\n",
        "\n",
        "htk.write_htk_user_feat(feat, feat_file)\n",
        "print(\"Wrote {0} frames to {1}\".format(feat.shape[1], feat_file))\n",
        "\n",
        "# if you want to verify, that the file was written correctly:\n",
        "feat2 = htk.read_htk_user_feat(name=feat_file).transpose()\n",
        "print(\"Read {0} frames rom {1}\".format(feat2.shape[1], feat_file))\n",
        "print(\"Per-element absolute error is {0}\".format(np.linalg.norm(feat-feat2)/(feat2.shape[0]*feat2.shape[1])))\n",
        "\n",
        "plt.imshow(feat2, origin='lower', aspect=4)  # flip the image so that vertical frequency axis goes from low to high\n",
        "plt.title('Expected log mel filterbank features')\n",
        "plt.savefig('fig/fbank_exp.png', bbox_inches='tight')\n",
        "plt.close()\n"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "400\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "shape of frame: (257,)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "(257,)\n",
            "(40, 257)\n",
            "Wrote 40 frames to 1272-128104-0000.feat\n",
            "Read 40 frames rom 1272-128104-0000.feat\n",
            "Per-element absolute error is 2.6203704962273145e-10\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4wf_j3cPDo8S",
        "colab_type": "text"
      },
      "source": [
        "# Visualization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eNKBHppMABvp",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "de0d4818-f00d-4654-fcc6-92af3619eb24"
      },
      "source": [
        "display(Image('fig/fbank.png'))"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAN4AAAEICAYAAAAwd0oxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCIjTPwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x6Jl6XeMABti",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "e03a2969-0954-4538-a3b9-cedd8cdafb13"
      },
      "source": [
        "Image('fig/fbank_exp.png')"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOoAAAEICAYAAABRb2hzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nVEUZBP5/KjJMXEGyLGAAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qZiyDb0UBESK",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "016c2623-d775-4759-d96b-1b1767528fbd"
      },
      "source": [
        "Image('fig/mel_filterbank.png')"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n79ChQ4cdgo7QO3To0GGH4P8Dbfo4l8THxXIAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dlpq8Fw_Bv8y",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "7ce7fcd0-cd5d-4c17-b80e-0583274bafa3"
      },
      "source": [
        "Image('fig/waveform.png')"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAEICAYAAABcVE8dAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nO2deVyU1f7HP6ziBioCIqAIA6SI6xBq\nGSoShoVXQyPT7LZgZTe73TS7v1LzunDrtl7bzEpaxMq6oZmYuKem4ZpgSQrKJiGLIso65/cHzjjD\n7POsM/N9v168XswzZ57znWfO833O+Z7v4sIYYyAIgiAcHlepBSAIgiDEgRQ+QRCEk0AKnyAIwkkg\nhU8QBOEkkMInCIJwEkjhEwRBOAmk8AmCA/v370dERAS6deuG7777TmpxCMIkLuSHTxC2k5CQgJSU\nFMyfP19qUQjCLDTDJwgOnD9/HtHR0TZ9trW1lWdpCMI0pPAJu+aTTz7BPffco3kdERGB6dOna16H\nhITg+PHjmD9/PkJCQuDt7Y2RI0di3759AIDy8nJ07twZNTU1ms8cO3YMvXv3RktLCwDg448/xsCB\nA9GzZ08kJSXh/PnzAIDw8HCcO3cO99xzD7p164ampiaUl5cjJSUFvXr1gkKhwIcffqg579KlS5Ga\nmopZs2bB29sb69atw9KlSzF9+nTMmjUL3bt3R0xMDM6cOYNVq1bB398fISEh+PHHHwW9hoTzQAqf\nsGvi4+Oxb98+qFQqlJeXo7m5GQcPHgQAnDt3DlevXsWQIUMQGxuL48ePo6amBjNnzsT06dPR2NiI\nvn37YvTo0fjmm28051y/fj1SU1Ph4eGB7OxsrFy5Et9++y2qqqowduxY3H///QCAs2fPol+/fti8\neTOuXr2KTp06IS0tDcHBwSgvL8fGjRvxz3/+Ezt37tScOzs7G6mpqairq8MDDzwAANi8eTNmz56N\n2tpaDB8+HElJSVCpVCgrK8PixYsxd+5cEa8o4dAwgrBzgoOD2ZEjR1hWVhZ77LHHWGxsLDt9+jT7\n+OOP2T333GPwMz169GDHjx9njDH24YcfsvHjxzPGGFOpVCw4OJjt2bOHMcbYpEmT2Nq1azWfa2tr\nY507d2bFxcWMMcb69+/Ptm/fzhhj7MKFC8zV1ZVduXJF037RokVszpw5jDHGlixZwsaOHasjx5Il\nS9jEiRM1rzdt2sS6du3KWltbGWOMXblyhQFgtbW1Nl8fglBDM3zC7omPj8fu3buxd+9exMfHY9y4\ncdizZw/27NmD+Ph4AMB//vMfDBw4ED4+PujRowcuX76MS5cuAQDuvfdeHDx4EBUVFdi7dy9cXV0x\nduxYAO02+vnz56NHjx7o0aMHevXqBcYYysrK9OQoLy9Hr1690L17d82x/v3767QNCQnR+1xAQIDm\n/86dO6N3795wc3PTvAaAq1evcr1MBEEmHcL+USv8ffv2IT4+HvHx8ToKf9++fXjllVfw1Vdfoba2\nFnV1dfDx8QG74aDWs2dP3Hnnnfjyyy+xfv16pKWlwcXFBUC7gv7ggw9QV1en+bt+/TrGjBmjJ0ff\nvn1RU1OD+vp6zbELFy4gKChI81p9XoKQAlL4hN0THx+PXbt24fr16wgODsbYsWORk5OD6upqDB8+\nHPX19XB3d4efnx9aW1uxbNkyXLlyReccM2fOxKeffoqNGzdi5syZmuOPP/44Vq1ahfz8fADA5cuX\n8fXXXxuUIyQkBGPGjMELL7yAxsZGnDx5Eh999BFmzZol3JcnCCsghU/YPZGRkejWrZvGDOPt7Y2w\nsDDcdtttcHNzQ1JSEiZNmoTIyEj0798fXl5eeqaVlJQUFBYWok+fPhg6dKjm+NSpU/H8888jLS0N\n3t7eGDx4MLZu3WpUlqysLBQXF6Nv376YOnUqXn75ZUycOFGYL04QVkKBVwRBEE4CzfAJgiCcBFL4\nBEEQTgIpfIIgCCeBFD5BEIST4C61AMbo3bs3QkNDpRaDIAjCriguLtYEFXZEtgo/NDQUeXl5UotB\nEARhVyiVSqPvkUmHIAjCSSCFTxAE4STwovBzcnIQFRUFhUKBjIwMo+2++eYbuLi4kKmGIAhCAjgr\n/La2NsybNw9bt25FQUEBsrKyUFBQoNeuvr4eb731FuLi4rh2SRAEQdgAZ4V/+PBhKBQKhIWFwdPT\nE2lpacjOztZr99JLL+H555+Hl5cX1y4JgiAIG+Cs8MvKynQSUQUHB+vlCj969ChKSkowefJkk+da\ns2YNlEollEolqqqquIpGEARBaCH4pq1KpcKzzz6L1157zWzb9PR05OXlIS8vD35+fkKLRhAE4VRw\nVvhBQUEoKSnRvC4tLdUp+FBfX49Tp05h3LhxCA0Nxc8//4yUlBTauCVky9ELtSgov2K+IUHYGZwV\nfmxsLAoLC1FUVITm5mZs2LABKSkpmvd9fHxw6dIlFBcXo7i4GKNGjcKmTZtMBgcQQPGlBhRfapBa\nDKdk2rsHkPz2PqnFIAje4azw3d3dsXr1aiQlJWHgwIGYMWMGoqOjsXjxYmzatIkPGZ2Scf/ZjXH/\n2S21GARBOBC8pFZITk5GcnKyzrFly5YZbLt7924+uiQIgiCshCJtCYIgnARS+ARBEE4CKXwn44df\nK1Df2CK1GARBSAApfDugsaWNl/MUVtbjyS+OYuHGk7ycjyAI+4IUvgy5eLlR8/+2/Iu45aUc7P79\nT87nbWhuf3CU113nfC6CIOwPUvgypPLKTYU/97MjAICfCg1XsCEIgrAUUvgEQRBOAil8J4IxJrUI\nBEFICCl8Z8TFRWoJCIKQAFL4BEEQTgIpfBniSBPwHacr8f6es1KLQRAEeMqlQwhPZX2T1CLYxCOZ\n7WmwH48Pl1gSgiBohi9DDO2tbj5Rzv28nM9AEPYFYwyf7C/C5esUXQ6QwndKHMhiRBAmOVRUg5c3\nF+DF705JLYosIIVPEITD0tSqAgDUXWuWWBJ5QApfhgi1aatOqVBaS6kVCOdiX+ElHDpXLbUYkkMK\n34n45kgpAODSVfvcACYILvx35x9SiyA5pPCdCNq0JZwZRncAPwo/JycHUVFRUCgUyMjI0Hv//fff\nR0xMDIYNG4bbb78dBQUFfHRLEILz3u6zZAogHAbOCr+trQ3z5s3D1q1bUVBQgKysLD2FPnPmTPz6\n6684fvw4Fi5ciGeffZZrt4RMqbzS6FA5e/6d8xvuW/Oz1GIQPOBAw9JmOCv8w4cPQ6FQICwsDJ6e\nnkhLS0N2drZOG29vb83/DQ0NcHGkUFJCw5nKesSt3IFP9hdLLQpB6EEKnweFX1ZWhpCQEM3r4OBg\nlJWV6bV75513EB4ejoULF+Ltt982eK41a9ZAqVRCqVSiqqqKq2hEB4Qe8EWXGgAABx3EBEKlIAlH\nQ7RN23nz5uHs2bP497//jeXLlxtsk56ejry8POTl5cHPz08s0eyGP7UKo8id3y/W40xlvdRicIJK\nQcqbyiuNuHzN8ocybdrykEsnKCgIJSUlmtelpaUICgoy2j4tLQ1PPPEE126dEnUQiVx5d3d7krTm\nVhWS3twrsTTcuWhHD1hnJG7lDnT1dEP+sklSi2I3cJ7hx8bGorCwEEVFRWhubsaGDRuQkpKi06aw\nsFDz/5YtWxAREcG1W0KGnCipAwA0NLVKLAk/HLtQx+v5VCqGr/NK0NIm7we3PaGu02yMA3/cLA1K\nNnweZvju7u5YvXo1kpKS0NbWhocffhjR0dFYvHgxlEolUlJSsHr1auTm5sLDwwM9e/ZEZmYmH7IT\nhF2x6UQ5Fmw8ifK6RsyfSJMeoWlpU+GDveekFkNW8JIeOTk5GcnJyTrHli1bpvn/rbfe4qMbp8GF\n0ps5JOqMjRTpLA7Lv9d1D6cJPkXaEgLgiF63fGyYq6+LVJuHJTXXnMrz6MDZDt5ipPFJ4RP844hx\nFvPWH+V8DvV1kcqWPPaVXbj3vQPSdC4B6mSBxE1I4csQYzNAe9l0cjx1DzQ0md4ctAT1dVFJ+Due\nqbwqXeci03FDl9wySeE7BKGLtuDVbb9JLYZDkF9+WbBzu2pWPtIqnqZW7g8vwj4hhS9DbNm0fWeX\n+ULhYqkZe7bozP3siGDnVl8XlcRemVEv5uCrX0rMN7SRDYcvCPrgtBV7WSELCSl8O6exRX6zNfIy\nMoyrxJu22nyw1/wEwVYWffsrJr/9k2DntxXpr7r0kMK3I4qrG/SOrfrhtASS6HOh+prUIsge9YNQ\nShu+Gj72JORIbUMzSmtpLBqDFL4d8eDHh/WOFctE0U55R34zOtmhnuHLQOE7atqIUat24PZ/7zL4\nniOl7bYVUvhOhJCGllqtJFauDjiq+FAV6k1bOZh0HBW555uSGge8Ne0foTY9Rdu0dUAb/umKK5zP\nobkqpO8lQX3ZW9pUyD5e5pQzflL4MuCjn4pw93/3aV5LNQ6LLzUg6/AFaTp3AjReOk6oaOSA+rK/\nt/ss5m84ju9PVkgrkATwkkuH4Ma/vue/xu/CjSfQ3KrCm2nDLf7M1Hf3o/ZaC+5ThsDV1fZZuj27\nZQqJs12XP/6sh8K/u9Ri6PFnffv+Rd21ZoklER+a4csQPgqHfJVXiu+Ol1uVO0Vth3c2xUQIw59X\n5JUkruO66q0df0gih5SQwpchP2nl8DaHOeNAY8vNTSxntFlKTUubCltOVjjNta/WygQq92986WqT\nJoOps0AKX4a4WWFO2XtGv/Zvm8iO3lc7FDxxxORptvLfnX9g3vqj+LGgUmpRRGHmh4c0/1Mxe/lB\nCl+GuHOwnwPA3sKbD4HMA8U6sy4hyHUQZXbkfA1Ka41nWPz9ovWmtouX28/nLPbiwj9vXqPc044x\nLhwJUvgyxNIZ/rVmw6UE29puzvBX7/oDI5fnAgAu1IgTpGWP8/sL1ddw73sHTbZJenMvWu20PKEh\nt9IDf1xyqkRqV6ww36zeWYgp7+zH37KOyTJ9ia2Qwpchls7wvzlaZvB4mxF78XkLo3I/3HcOJRwe\nDhwXKJJQWmfZ9+XDWCaFbduQrXrm2kN4eTP/HmJCc6WxBTtsWD0UXdJPTWKM//x4BidK6rD5RDnm\nGIhwt1d4Ufg5OTmIioqCQqFARkaG3vuvv/46Bg0ahCFDhiAhIQHnz5/no1uHxc2CUNUj52vx0nen\ndI6pNwaNzfwtZeUPv3Ea5HLfrJMKOQakrT/EX9xFa5tKlDxBT60/hkcy84Tv6AaHimpE60toOCv8\ntrY2zJs3D1u3bkVBQQGysrJQUKA7axg+fDjy8vJw8uRJpKamYuHChVy7dUjUCjsyoJvRNlduuFla\nM1uxhQYODw35qTX+sNXZxhmcdK6JZPootmDsq33tCV04K/zDhw9DoVAgLCwMnp6eSEtLQ3Z2tk6b\n8ePHo0uXLgCAUaNGobS0lGu3DonaD96veyejbZgJEzIXpcJn6Tvy0rmJ9qx+1+9/CtLHC9/+itBF\nW0y2OdixvqtIXLwsjeKtbbDMXl/b4Byb6Wo4K/yysjKEhIRoXgcHB6OszLBtGQA++ugj3HXXXQbf\nW7NmDZRKJZRKJaqq9N0NHR0p1eSR87US9n6Ta82t+OLQeYf0W88+Xg6A/99ZnQ7j0cxfjLZ5a0ch\nz71aRuIbe3g/J5/zieoGyzzY3tn1B1rsdMNeG1E3bT///HPk5eVhwYIFBt9PT09HXl4e8vLy4Ofn\nJ6ZossCSgCt1pkVDY/6ut/ahqbUNf//yhN57mQeKOUpnmIuXG1HAQ2IxNcu3nMb//e8U9hiIL7CV\nNhVD1ItbeckTpL7+15utM1+8s1v4qM7c07atIPhIDGeM+kZu+0m2YmlGUkvnFa9u+x2fHbT/vUfO\nCj8oKAglJTfLpZWWliIoKEivXW5uLlasWIFNmzahUyfjJgtnhqvb5O+V9dhp5KZfsinfqnNZeiOM\nWrUDa/aeM9tOOzhr84lyFJQbVjI1V9uX2NYqVFM0tbahqVWFZTx5pBw6V42Bi3PwU6HlEdElNTf9\n++W2drnG47W2N6z5La47gHsmZ4UfGxuLwsJCFBUVobm5GRs2bEBKSopOm2PHjmHu3LnYtGkT/P39\nuXbp9Jytuop/fK0/iweAJ744Kli/xy7UInTRFlRcNh6cBBhefaijf5tbVfhb1jEkv73PQCv+2XKy\nAnErdwDg74Y9fMNr4+dz5u3ilwQOenNGaIfIdjgrfHd3d6xevRpJSUkYOHAgZsyYgejoaCxevBib\nNm0CACxYsABXr17F9OnTMWzYML0HAqGLORvlC9/+Ko4gHfjs5/Yl7f4/TCs6U/Kv1CrJaEvkqrW8\nlH2KV7OCtVsLO34TZqOWby5fb0Hcylz8mH9RalHMwmeVN2t+z2MX5LHPxQVe0iMnJycjOTlZ59iy\nZcs0/+fm5vLRDWEnmLqJ8ssva/5PenMvjrw4EQUVVzD7o8PY9NRtIkjHHbmZZMxxxYKMqWlrfkbl\nlSakf3YExRmTRZDKcjadKMdvFVewcNItFn/G2Bj8745CaO+9WuMcYOseiZygfPgyxNQYXC9SgZI/\n6203RRic4Rv5Tuq0DwCwk8fZcEnNNZTXmTY92YL2d7PVW+SAyC6S5lID8OWSu0Ggsfl01jEAsErh\nG+O17WdwSx/55egXC1L4dsar235HbP9eUothBumtrGNfaS9k3aurJ6/nZQyc3fOqODxMbUGsCN/c\nAvuYAbdqhQPb22qNK5RLhxAA/duovqkFKpHTNgvFf3e2u1jKJXbBHFLnNvpb1jFeivpYg6lVsrYZ\nJ9+It5ijQgpfRAor6y3KcyPbQFUO+vr2f++yKviHr0eDkJfywNlq0WsP2ILUIm4+UY6n1gvnPWaI\nZhOrMO3LsWbvWavOa8l+iJwhhS8SLW0qJL6xF49/zm3gixmAamvYubHUClt+NV80mu+HXbXAofML\nN54U9PxcyS+/jEdMROCKxZnKq6L2Z2nyv0oryzBOFsmdWChI4QtAbUOznvlCPRM05bttqceApVGE\nlvJr6WWDx/ef7RBYJPDKgzHuD7TL11rwdV6J+YY88c1RafNCGcuR06ZieGDtz5j89k84aeT35Ys9\nZ6oQumgLztcIm9DPGIZSP5ssXchhjGkH0NkjpPB55s/6Rgz/13bj5gsZWgAe//yIKP24wPINRFuf\nLX//6jgWCDjrNvRAEju4asS/tiN00RaoVMxo31X1TWbjJfjiu2PtubOsnS3zxfNW/t4yvAVFgxQ+\nz6izA3ZU+NaYKv6z7YzJ9/n2ujAWgWprP6ZWKsdL62w6p6VIkRZXuVzcOJOaG2aqc5fENZPIFWvT\nMje32n8SNFshhc8zap9hADhXpX9DNrep8OQXhmfUatu3uWRkfJt0miy9YSzs1liAiouL8DebAybZ\nNApjxn8S2W78c8CYO6y1mVU93U2rvQs8RvLKDVL4PKMd9v3lL4ZtyT/8Kn34+vaCSs0+Q4OR5FnG\nlIZQukSqFL72jDFl54D6Hq9vN73ytRRzD4gTAq9CpYQUvkjIrbzdY5/mIfNgsck2fEtszTV44ouj\nmiRlllJ3rdmsX7WjxAIAwOaTFcZXNCIONyG70t6jeG/3WYNF5JtarFs18pmLx94ghS8gKq27UY5L\n7Aoz1YikkLlVdfPmnfHBQas++6QFmULPG0lBbenDaF+hfArzvO0EK6K6a7putesM1HU4XOw4NWeF\nhhS+gHy4r0jzfxOPtutfisWK8DSsBG2dI1vyAOlo/7cm0MWS3DnaIuz+/U/ELN1mVdH3T/YXm3x/\n4cYTmPtZnmgVu17sUMheDd8rymvNrWhqlT4fvNBxFY4OKXyBUW80fWXEnm9XyHCVoo21tXRf3fY7\n6htbca6qweKN8INmcuB/lVeKbfmVOGtgw14ItAvLaMP36mzQ4m1IfH2vwfeKqoXzv+/43Hxvt3WR\nsYQupPAFRu21I0fLsblKVXpKQ45fQouiS+YVj1hfQajyp6fKTAdRqVcWXPS9sWpjxiqyHbvguJuc\njgYpfIHZekp6jxxbMaY0bFUm1s7AAf7dLHecruT3hEbg23VWzTYzBUr+dyMIypZrrWbg4hyjKwex\nkePelz1DCl8kaNzK4xos33Ja75gUvvvahWCswVyt341HSnHxciOUy7fbdH41tuZRIuQNKXyR6DhT\nMbSpJ9ZGn6U8+5XhurmOgpCzR3M/5eYT5hPJGWLtT0Um329VMWw9VcE5Q6bMhiLBE7wo/JycHERF\nRUGhUCAjI0Pv/b1792LEiBFwd3fHxo0b+ejS7uioW7KPl0siR0dMFePge1lvi4KV20NQ9tDlEpxC\nkXP78wlnhd/W1oZ58+Zh69atKCgoQFZWFgoKCnTa9OvXD+vWrcPMmTO5dme3dLwPDSW9+l3kFLIA\nUFprfRCKkG6ZHdkggndTzbVmXOWh0Ln2w+niFfFz+gDtewd8LFyE2oOwHjkYAnXJOmy/HnecFf7h\nw4ehUCgQFhYGT09PpKWlITs7W6dNaGgohgwZAldXsiCZYvMJ8Wf9ddfkXdChRkBb8qmy9qjcOR8f\nRvpn3DOGagcF1fPwALEFxrht2Kqx1I8/55R1pqnZHx1C9vEyi9v/IkFQlSNvFHPWwGVlZQgJCdG8\nDg4ORlmZ5T+oI2Gq+pFcx9B7e6z3a7bZS8eGTxoz6dTLrPLQ7xfrRS/jZww+FJal57C2oM++wkuY\nv+G4xe1f+PZXq85PmEZWU+41a9ZAqVRCqVSiqko+IeyWkvjGHqlFsJorpgpFyABD+n7nb5WIWfqj\n1bl2hKT2WrPOUt9c6Tyh0jiLaYj5/OfzIvYmHua2jex5BcBZ4QcFBaGk5OZALy0tRVBQkE3nSk9P\nR15eHvLy8uDn58dVNNE5V2U88KfjMpuPZTcf/HZRvFnpr2aChgxh6N5TV3k6dkE+RcQ7zkTV5iJj\nCJUmurzuuiirSZWKGU3rwJVfimtELypjDfbsR8BZ4cfGxqKwsBBFRUVobm7Ghg0bkJKSwodsDoVM\n9LvdoTJxdzW2qJD+aR5Ka6/hT4k2SdUYivLdc8b4KrVSIHkrLjeKMtiE0HktbSpUXL6O6e8fxNR3\n9wvQA+HO+QTu7li9ejWSkpLQ1taGhx9+GNHR0Vi8eDGUSiVSUlLwyy+/YOrUqaitrcXmzZuxZMkS\n5Ofn8yG/3eAI+l6KiY0hfa9eHW0/fRGnyq7Aw90VLTKsYrT/j0tG37MlmZ6pPSJt+BhrUkxQlm7K\nxxeHLgAQt3Zs8aUGhPbuKlp/UsJZ4QNAcnIykpOTdY4tW7ZM839sbCxKS6Ut9iw1x0t0zRkXBEw4\nJTSXrjbh0LlqxIX5SiqH+mGw5WQFxkb0llQWa7HFLFDdYJmZQwxlfdpMVTZznK9uQKuKIdyvm+bY\nzt8MV0oTmo4PX3PXz55X67LatHVkvjmq+8DLPGh/G17qcb5q62+4b83PovRpKiJZ+8Y7cl4+9nw1\n1riUVl9twju7/jAZaFZaa9msl4/UyOrC5Ma4+78/WX1O7e8W/+puJLym6+Rgrj6DUFgbc+DUNnyC\nEBJTVozqq81a7eR3F248YnxV23GW+Pw3J/Hqtt/1ah0cOHtJ44v+x5+WBebxMQP9z4/8lBPUxtIH\nltSYG0qFf8rD/dYWSOHzhC0Rq4YoMZKCVo5YUzjEVgzNvtQ2fO0ZYaOVZe7khjqNRatKhc9/Po/i\nG5vAMz88hOnvt1f+crNQk8vV4uDqKk/JWlqtmyzsKzS+NyN3eLHhE8CHZnLLW8pRGbkamuN+Ecw6\nn/98Acv/EqNzzBHy65wsNeyiqlLdrGLVrZPu7ZlrYWrnFpnW7ZWpvseXeRcQExxjvqEDQDN8nqjn\nKdHYu7vkW9Gnoxo5YURpCY1UaQuERG13b9N6mHVMXmdpbQWhfPwBbu6krgZWKJdlkNqjYxH0MgtK\nZdorpPB54tuj/KST+F0m4flyhmtCNal99k2h4mF2LuQKiEsGVUMKPyPnN6Pt+bgWlnC+WteMaqn7\nqz1CCp9wOqa+e0BqEYzSakTZWKPEhbR4cTm3IZNO1uELmPDabjS26Bd2saaAPRcOS5CgTSpI4RMW\nI1MTrNXIecneaqQYrjV7O8KmNrb93IZm+EB7ShJD38+W4DTCNKTwJabuGpWSI27yxBeGs09eM1Pa\nUBs+cvsDwCf79atrcbF2rPhBv7ykmpkfHtI7Jpe6uo4EKXwRMJXK9wkr08uKzQGt9ACOa9mUHj6j\nN2t4mkS8vLlA79j2AtuLwJuKSzDEpDf32twXYRhS+CJgrPA0YwznZZ5iYeZa/ZkXIT6WBl0B7a6s\nQvHqtt8FO3dHWtqkmWK89qN431FsSOFLyPcnK3BZ5vnoCXlgaLYtBofOVUvSr1QwxjgXgJczpPAl\npKq+CQ1W2GadnabWNmzLt8wX3d6QayxZgYEkaXxFlcsRR1b2AEXaikKbkbt52ffSzNrslX98dQLf\nn7SuhirBjVoDCeDufINs640tbfDycJNaDKuhGb4I/C5iVSkhkdotc8dpadLnOjNv7/xD75g1HkOO\nSsfst/YCKXwR+GAPP3l2pCLnVPusWurV7nUDwTmOwkEns5XLFUsD3DpG59oLpPBFoNlIMI298LjE\nrqOZB4rx8U/6PuGEOAiZm8deWcNTskSxIRu+CMixOIc9sWSTY5bDvHS1Cb27dZJaDLPssDBLpyMg\n9SpWaGiGzwMtdj6DtxQ5FhmxlYytxpN2iYW97O3YQ4qDBorKtQheFH5OTg6ioqKgUCiQkZGh935T\nUxPuu+8+KBQKxMXFobi4mI9uZcOKLcZDxh2J7OPlUovAG+/vkT4N9QNrD+F6c5ssUgSb4pkvj0st\nglmil2xD6KItNn9eXaPXmjnNpweL0djSxqkQEGMMf9aLl73VhXHMpdrW1obIyEhs374dwcHBiI2N\nRVZWFgYNGqRp8+677+LkyZN4//33sWHDBvzvf//Dl19+afK8SqUSeXl5VstTd60Zw5Ztt+oz0X29\n4eHmiuMldejk7oqpw4PQ3csdl642w8vDDVmHL2Dq8CDsPVOFbl7uGB/lj3UHitG9kztvefAJgrB/\novt6IzkmELmnK9G3R2eE+nbB6Yp63DqgF97YfgaMWb6nV5wx2SYZTOlOzjb8w4cPQ6FQICwsDACQ\nlpaG7OxsHYWfnZ2NpUuXAryrnYsAAB/4SURBVABSU1Px1FNPgTGmKVXHJ7YUcM4vvxlc0tSqMphv\n/X83ijpXNzRj3YFiAPwVPSEIwjHIL7+i0SfHLtRpju/8TR4uxZxNOmVlZQgJCdG8Dg4ORllZmdE2\n7u7u8PHxQXW1vhvamjVroFQqoVQqUVVVZZM8Pl08bPocQRCEoyMrL5309HSkp6cDaF+W2EpxxmQ0\nNLXC1cUFnT3dwBgDY8IVUeZiO7QXDr4wAaNX7ZRaDIfjjfuG4lpzG/7vf6ekFsWpeWZiBJ4cp0DS\nm3tRdMm6hIbeXu7Y+swdCOrR2ep+m1tViHxxKwBg8d2DMF0ZjMYWFfy6C+O9xVnhBwUFoaTkpgmk\ntLQUQUFBBtsEBwejtbUVly9fhq+vL9euTdJVqwC0i4sLr+lnO1K0KhkDXvjB6PuRAd1wptLybIdy\nJNDH+sFMmObbJ8dgRL+eACBrhR/m1xXnquSd1bU4YzKnidczEyMBALueG2fxefJfTtLRM7bg6e6q\nZ6vv7sXplCbhbNKJjY1FYWEhioqK0NzcjA0bNiAlJUWnTUpKCjIzMwEAGzduxIQJEwSx30uFue9y\n74hgkSQhLGX/oglSi4DhIT2kFsEi7hrcR2oRLMLWTU5b4arspYCzwnd3d8fq1auRlJSEgQMHYsaM\nGYiOjsbixYuxadMmAMAjjzyC6upqKBQKvP766wZdNx2Z5JhAqUXgxFAZKKbxUX68ns+W5Tff2Muk\n5x+JUVKLQPAEL4+o5ORkJCcn6xxbtmyZ5n8vLy98/fXXfHRllwi1dyAWGx4bJWn/a2aPxJ3RfZxi\nr0SO2Pv4JW5CkbYiYM/3S4R/N3T2lDYN7JDg9hXGew+MkFQOIRkb0VtqESwm0EdAI7OdkBxjH2au\njpDCFwF7yJdiDDkkU2A3pAj37yaxJMIhBxOTpWx8YozUIkjO6zOGSS2CTZDCFwEPN/u9zBwDsXnB\nHjfHrEVhRw+z3t08pRZBcjq52+c9bZ9SOwivTR8qtQhmkV7dA95eFEwnJ9zsZLNZSOxlw70jpPB5\nwsvD+ksZ5tdVAEkcF087XimZw54ykbqJsCn17ZNjkCWRs0BsaE9J+hUDx72DRMaWWY9d3OIyEjK0\nd1e8mjqE83mGBPvwIA2/2FPxbDFmtyP69cTocGGDM43haqezd0sghc8Ttt4EPp3lba6Qmx6argwx\n38gMbTLUrt1kuk/Rr1cXqUUQHQfW96Tw+cKWzc1O7q7Y9NRtAkjDH9rfS44zY1uQo8JPiw3B85Nu\nkVoMPT6aY3tOK1v5+400BwAwOMhb9P5tybhrL5DCl5Dovj7o72s/dny5r0YsRY7mcnc3VzwxLhwf\nzB5p8P07IvmNNLaUHl3E98iZPzFC8/+YcOHjE8Z1iOKmGT5hFhnqEN7p7yuP5X2vrtyUUJscNf4N\nkqINB/RI5R4rVNZGSxFD93a02ZPCJ8wiYx3CiT5aUZUvTh6ECbf4a15nTIuRQiSN8rNV8atkaNIx\nxzNas16CXzo6HXVylzayXEhI4fNEaG/7Mc1Yw7sP3DQxeHm44eOHYjWv027tJ4VIeOnuQejq6Waz\niUnOM3xjjOzfS2oRRGFEP91EfWLst3R0uDA3wX/uzkgzLeQLKXye+OyRW6UWQRAMzaJDJTbtTBsR\njPxlk+DhZtvaW46bttqcXHonp88LmZfn1MtJgp3bEGI8nK1NX26vQVcAKXze4DNfTlosd9dDIdn0\nt9uxb+F4qz/32NgBVn9m/aNxRt+z1V9a7hN8rpHFg/ry49kyvJ9+Wmyh3UcDvHUTs4nxW0X16a7z\n2o71uVlI4csQuQ84by8PhNjgn93dBkU2RmF8tmrrTGvdX2PNN+IBU5vc8QJ63fDlVvj13NG8nMca\nMu7VDayTYjUW3de0+7G9FK4xBCl8WSJzjW8j6XeE8Xo+W69SREB38414QGnC7s7V08gUvbry4z7r\nbiKVxX08BMB15LGxA/T2ZVolUPjm9oZMTULkDil8GSL3Gb6teHnI1/th1ij+N6B9TWSVtMXN8sHR\n/S1qZ8tKyloi+/D70PRwczEYeMZ3pTNLcNT7DyCFL0sceLzxCl83poebC9xd+b8VZihtr2Xsq7UC\n+MuwvgDETRMdYSZds6350/6RGImvH9c3FfXr1cXgiuJOI3EJQnKfzPfQuMBplNfU1CAxMRERERFI\nTExEbW2twXaTJk1Cjx49cPfdd3Ppzq6xdKbiyBkh5crAQG+BApuMa0VzvWnPdqcMC7K4x+//djsv\nEwZ1lTE17z0wAi9OHqh5bWsff0uIQGyovqlrQG/51APo4inPvEZ8wEm7ZGRkICEhAYWFhUhISDBa\nnHzBggX47LPPuHRl93RcZnc2YN64Z2hf/L58ktlz3T3Evoui88UMHu3IYphBrEF79WLNBi9f3mKs\nwyPprphAPDqW3z0Ybd64T7raEM6U35+Tws/OzsacOXMAAHPmzMF3331nsF1CQgK6dxdno0yudAzM\n6pi/A2ifNbm4uDi0DZFPtG3aXK4ZY+KnjTC3oFC7nE4bHqQpIh43QLzgq0iBN7a/eDROpx6ElA/c\nfjJJGSIGnBR+ZWUlAgPbZ5t9+vRBZWUlJ2HWrFkDpVIJpVKJqqoqTueSO4YUlPqYmNn6Arztt96u\n2i0zJsiHU/UwN1cX3oNpvnliNKeHkPqz2oVRxkX5G2nNP+lmZvNcr9dtit7Y8rexANpXtoQ4mDVW\nTZw4ERcvXtQ7vmLFCp3X7TNTboMgPT0d6enpAAClUvy0rGJiSKlLMbFfck80nvziqAQ9G2ZQoDcK\nKq5Y3P7Hv9+BQB8vdPfywLNfnbCpz07urrxc+/dnjcTjnx8BAPT37YrL11uMtjVnw1fP8G3ZWeDj\n2eVqZlfWlgpvHens6YbijMmcz0NYjlmFn5uba/S9gIAAVFRUIDAwEBUVFfD3F28GYm/camCjSg6M\n6Cevcm6jwnytUvh8mR74UJLhWiYKcyYbc5vEN2f41snQ0fYuFPeOCMbz3/xq1WeOvZQokDSEpXB6\nTKekpCAzMxMAkJmZiSlTpvAilCNyuxX5TcwpH3vO5WGOfybfInot05485Xzv6DYZaqLWgbm0EOr3\nbal1K4ZJ0FRQljF6ChhsRlgGJ4W/aNEibN++HREREcjNzcWiRYsAAHl5eXj00Uc17caOHYvp06dj\nx44dCA4OxrZt27hJ7QAMDNSfmTqyIreEe0cEw93N1WAOFyFQ9/PQbaGcZ/g7/xGPvj06azbjfTp7\nmCz2Hdyzs8nzaUw6triLOvcwEpRPRErLIRScHE59fX2xY8cOveNKpRJr167VvN63bx+XbhwGLw9X\nNLaoAMBkLhpnvV9H9G9XwGJF5Krd8dxcXTjPioNuKPB1f+Una6r6WaFSWfc5uSeGs3fGi7hxLgQU\n5SMiv/3rLs3/tmZ6FJOhEiaJ0raH24q5eqgJAwMAAIE+XqK7wprTy2p5xLLJ24KYbqIEP5DC55En\nx4Vb3HbSYP2QcbXOEdO04+kuzyGQ/dTt2P73Ozidw9ys/fH4MBx7KRHBPbn7YfNtN78j0g9jI3pj\n0V0DzTfmmXwLc96bMlnZC+ZSSDga8rzb7ZTBQabTqmrjIXEKhROL78Rnj9xqOmujAPaBkF7Gbdfa\n3XXr5M45q6Wh5+bAQG+t912s3kicayTjp7XPaHOXtounOz57JA4DrKykxsDdJChmzh6xGR3mK7UI\nkkIKn0c4z3csPAEf8yqfLh4YG2E6ZF++xgTLMHSduObM6e5lWBla+5vYgUXPLHL5Dsr+trsW2/sY\ntxZS+DzC9QYQM8JWjvB98wX6mPaE0YarGc3az1tj/rMGYZLAGcbWmsJ8481BjldSh5hv5ECQwucR\nS276j+Yo8enDpj055DJzEkJ3mHyo8dzhK9Mtv5m5XnJrPy9k7hhts5WQrJwaI0o/5uDykFOQDZ+w\nFUtu+oSBAbjDTPZDubjWCeEhIubDzNvLQ6/2rrFra6lcxtxp5fKQDurRGYODfESJau1hRcDabQrb\nbedCXltr7rUUB8j5QwqfR7ibBXgSxI7Qdp0U4jln6XLfUnOasZteLM+qQB8vk++r5ZBLVCsfKUW0\nS2MuuWeQybYPjQnFY2MHcO7TEIYy3NobpPB5hC+zgFwUvxArDe0CL2tmj9RRoL5d+c/c6d7BdZDr\nqkXqaGi+0kDwRRfP9iC512cYzlaqiSfgcNkX3Bml+T91pOkqYktTovEPrfYd4fL7S1Bel3dI4fOI\noSp599/Kf7k0sXSOEAp/7ZybWVBH9u+Jx8aGIffZO/DeAyOQHMN/ObuOLoZcTTpSYy4lg1SE+Rm2\nhasfCFxWHO5uruh243c09PN1PObl4YYv0y3Lx+RtxOvKEGJuiAsFKXwe6WgW2LNgHFZNs94LwJG9\ndfr7dtXx/XdxcYHCvzvuigmUdPYcGWD95l24X1e8nBKNt9KGCSCRYV6627RJQxsu7oqWcv+t7cXf\nPdwM/3Yj+/fE8r8Mxqpp3DZ43d3UuYWA3/5lviqcqYBC7cIr1ow5Rwg0I4XPJ1zdMmU2niyZz4RZ\nGBikXZ3qv/cPx5hwX6s2/SzBkiClpxMiDB5X+NsW5DVnTKhVNWe50smGPPR3GYjq5ov/Sx6I3/41\nyWAtZt+unpg1qj9mjeoPb45eSV+mj8ZT4xXw9nLnlGuJMdi0WTQ3PswhCrWQwueRjvlxrF0Bqmf2\nYit+Y0rbkiXs1mfGWtRHUI+bpojbFL2x/rFRvM+Y1mllMpw2wrAS5vOmldqebw61eLNG9TfdkAOu\nri5GFfDHD8Xy9lCP6tMdzyVFGbzm7oZsqSaw5Lbc9dw4zf9z7wjDC3cNlDw6ng/s/xvICHnf/sbp\n7Gn7jKmTuziZLS1Bexb5+gzxzCxy5eWUwRgd5ouR/Xvi6QkKQfvie7VmDSunDtY7ZkqpWzKR0faG\n4qswvBwghc8jo8N1fY2tnQD2793u423qY+lGcrkIwZs82qb7mHEntCdiQ9tt40ODbcsmymXjteP+\njqmMoIP6eiMrfRS8PNx4SRBnCr/unfDT87oxD2ItgPy99ceWKZ0eK9Pqc2JACp9HuC75JkVbZmvl\n+z4ydmPe0seyiM0XJ5vP6NhJplk5bSH+RuBcHx/zMz+hfbe/eWKMRe26dBJ+JRbcswv2LBgneD8A\nMOEW2/LSM1hm1pO5tc5mHOculAkP3yZM0IeQcPUKCuuQu96Y/Vxo5HiTPjQmlNfzdSwebqlJTSzP\nr/5aZR2F7NPQJrEuxqf4lqRT6OTupvFy6maF66bcIYXPMwuSbgZ92DrgTSkuU/bHe0eYDkqxpT9L\niOjg4SJ1jpWOvtWfPxInkSS6/F8y99z2QubgsSeGhFieirwjfXtYZlJb/9go/GtKNGYo+Y+lkQpO\nCr+mpgaJiYmIiIhAYmIiamtr9docP34co0ePRnR0NIYMGYIvv/ySS5eyp7Onm802Wq5eH68ZiXYU\nmo75ZQx7bQg/w1TPduM65Dw3lYNfTIz5qouBFKsfIfu8Lby3TZ8zvzLQauvuitmjQx3C/14NJ4Wf\nkZGBhIQEFBYWIiEhARkZGXptunTpgk8//RT5+fnIycnBM888g7q6Oi7dyp658e2pb3t3t8xzIdRX\nV2GaUvxydwWUks6ebtj+9zvwdtpwneNyC2Tj6yf07667hzBvvDApl20lwoZgNr4wthCWalIkFzgp\n/OzsbMyZMwcAMGfOHHz33Xd6bSIjIxER0R7s0rdvX/j7+6OqqopLt7Jn9qj+KM6YjC6eltn+Rt+Y\nrRgrrtGRvydG2iybIcRQhwHe4ri2RQR013Mz5fsZyTXCnq8I/YmDAnReL0i6hZ8T84SQLrsxVlSX\n0ybAgEePM8FJ4VdWViIwMBAA0KdPH1RWVppsf/jwYTQ3NyM83PBMZM2aNVAqlVAqlQ7/UNDm5ZRo\n7HpunEX+vowx9PftinMrkw2+/yzPDwO+GN5P+DB/Y6iX5OYUvzXLfcD6lYOUqzN5rXG44+pAZhYx\nMTulnDhxIi5evKh3fMWKFTqvXVxcTA7oiooKzJ49G5mZmXA1EhmXnp6O9PR0AIBSqTTYxhHxdHe1\nunapsQH/dEIEXt9+hg+xHIa+PTpjxdTBSBwYYLLdhFv8kZOvP9bVjAm3Pqe7tQ8RSxgd5ouD56p1\nopfNYSyPP+FcmFX4ubm5Rt8LCAhARUUFAgMDUVFRAX9/w76xV65cweTJk7FixQqMGmVZFjtnRvS5\nixPsCzwQxz29gC0ulh2D8fhg/WNx2HSiHJNjAi3+zGAbTSD2iv3ntRQGTtOPlJQUZGZmAgAyMzMx\nZcoUvTbNzc2YOnUqHnzwQaSmpnLpjoB41bBu6WNbMjFCFxcXF4znOfjKxcUFU4YFwf3G6uH9WSM4\nZ+xUexDJNf0ywQ+cFP6iRYuwfft2REREIDc3F4sWLQIA5OXl4dFHHwUAfPXVV9i7dy/WrVuHYcOG\nYdiwYTh+/Dh3yR0Zx59wEzwyaXAg54ydDpDq3SommjHvOSqcQsh8fX2xY8cOveNKpRJr164FAMya\nNQuzZs3i0g0hMEnRAThRUqfz2lgaYWMUZ0xG6KItfIsmKk5g2TJKYA8vRPh3x7zxwiZZEwtzwVUf\nPjhSJEnkBUXaOiDL/6KfPdAUT8Trek19MFuJ6L7OZfO1BGNVnSwlwLsTArw7YbEVRUzEws3FBR8/\nFIuRPBVNGRIs7fgxt6FtzsnEUXGcJBEOhCl3vzkWbBwaG8cDendF0aUGA+2db+DbgiU5WEzh6e6K\nQ/+cyJM08iX32XjR4i4I66AZvp3Bxb2uY0SvkMwZLVzRDaGw9Ll3x41smeNtzNgoR16eYt2q0BQK\n/26U80emkMKXIVwn3CEC5z53doaG9EBxxmTezB9SM7xfD03KZ0di6T3yM51JDSl8B+QOB7x5CcJa\nHuqQqlzI2r72Ail8GeIIFnV73BdIHCSMq566/J+cykESzgkpfDvmi0fjLE64ZildONS3tXemDg/G\n1vk3i7LztVJ6eUo0Xrp7kE2pGQj+sMM5CO+Qwpchlg7M2xS9kftsPAD9VLkd+WruaE1pt64CKfXo\nvt5W++/LjYGBN8s6WlLs2hK8vTzwyO0D7HLVY+9ouyg7W3CZIcgt085R65COM/OPH1Li4XV5mte3\nDugFzxt1ZU25F3K5KbY8PdZ8I4IQEZ/O5C2kDc3wZchfraiL69/dC4vuugWfPqxbxm/CLQE4s/wu\nvkWzmAfi+knWN0GoidfKY0QzfFL4ssSSvPjaPB4fjn4GfOzVM3opiAiw/+RrjqYgnDEhnreXB957\nYITUYsgGUvgOzgNx/dC9kzSWu1/+z76jSpmDJdmdbYfBcHwQcyPNw7QR3BLMOQJkw3dwVkyNwYqp\nMZL07WdmI5kgxCC4ZxcUZ0yWWgxZQDN8QgcpC0/LDW+t9ADD+/WQUBJ+kFsxd0J8SOHbEX+sEH4T\ndt1fbxW8D7lz/63tG84Z04ZojrlTDVXCASCTjh3hLkB91I707EJubKumxWDVNGnMYAQhJDTDJwgT\njArrBQB4aIzlrrIEIVdohk8Iykt3D8LWXyukFsNmNqSPlloE3tAO9O3r44Xyy43SCUNIAqcZfk1N\nDRITExEREYHExETU1tbqtTl//jxGjBiBYcOGITo6Gu+//z6XLp2WeePDzTeSIY/cPgAbnxgjtRiE\nA9KrqycmDnScmgRiwEnhZ2RkICEhAYWFhUhISEBGRoZem8DAQBw8eBDHjx/HoUOHkJGRgfLyci7d\nOiUUIk5wxVjg1dtpw0WWhB+OvpSItXNipRbDruCk8LOzszFnzhwAwJw5c/Ddd9/ptfH09ESnTu3+\n2E1NTVCpVFy6dFocLeqTEJ/h/fQLtoyP8uNURY2wLzgp/MrKSgQGBgIA+vTpg8rKSoPtSkpKMGTI\nEISEhOD5559H3759DbZbs2YNlEollEolqqqquIhGEIQJaP7gnJjdtJ04cSIuXryod3zFihU6r01V\ngQ8JCcHJkydRXl6Ov/zlL0hNTUVAgH6xifT0dKSnpwMAlEqlRV/AWegkYV4cgiAcA7MKPzc31+h7\nAQEBqKioQGBgICoqKuDvb3oDpW/fvhg8eDD27duH1NRU66V1YmbGcc+Doo4dMlR5ycvDFY0tKjBG\nhSIcmXnjw/HOrrNSi0FIBKdpY0pKCjIzMwEAmZmZmDJlil6b0tJSXL9+HQBQW1uLn376CVFRUVy6\ndUr4yHwZE+SDpxMi8Pb9+pt02fNux8JJUXCliFKHJtCns9QiEBLCSYssWrQI27dvR0REBHJzc7Fo\n0SIAQF5eHh599FEAwOnTpxEXF4ehQ4ciPj4ezz33HGJiKIpRClxcXPBsYiT6+HjpvRfVpzueHKeQ\nQCqCIMSCU+CVr68vduzYoXdcqVRi7dq1AIDExEScPHmSSzcEQfAMeX05JxRpK3O+/9vtuHy9RWox\nCAeB9mecG1L4MmdwkI/UIhAE4SCQrx9BEISTQDN8mfLJQ7EY0Lur1GIQDsqA3l1x8UojJtxCuWic\nCVL4MmU83YiEAHi4ti/qIwO64f1ZI+HdmVSAM0G/NkE4EX8ZHoSzl67iqfEKdPeihHzOBil8gnAi\nPN1d8cJdA6UWg5AI2rQlCIJwEkjhEwRBOAmk8AmCIJwEUvgEQRBOAil8giAIJ4EUPkEQhJNACp8g\nCMJJIIVPEAThJLgwJs/M2L1790ZoaKjNn6+qqoKfnx9/AtkhdA3oGgB0DQDnugbFxcW4dOmSwfdk\nq/C5olQqkZeXJ7UYkkLXgK4BQNcAoGughkw6BEEQTgIpfIIgCCfBbenSpUulFkIoRo4cKbUIkkPX\ngK4BQNcAoGsAOLANnyAIgtCFTDoEQRBOAil8giAIJ8HhFH5OTg6ioqKgUCiQkZEhtTicKSkpwfjx\n4zFo0CBER0fjrbfeAgDU1NQgMTERERERSExMRG1tLQCAMYann34aCoUCQ4YMwdGjRzXnyszMRERE\nBCIiIpCZmak5fuTIEcTExEChUODpp5+GHK18bW1tGD58OO6++24AQFFREeLi4qBQKHDfffehubkZ\nANDU1IT77rsPCoUCcXFxKC4u1pxj1apVUCgUiIqKwrZt2zTH7WXM1NXVITU1FbfccgsGDhyIgwcP\nOt04eOONNxAdHY3Bgwfj/vvvR2Njo1OOBZthDkRraysLCwtjZ8+eZU1NTWzIkCEsPz9farE4UV5e\nzo4cOcIYY+zKlSssIiKC5efnswULFrBVq1YxxhhbtWoVW7hwIWOMsS1btrBJkyYxlUrFDh48yG69\n9VbGGGPV1dVswIABrLq6mtXU1LABAwawmpoaxhhjsbGx7ODBg0ylUrFJkyaxH374QYJvaprXXnuN\n3X///Wzy5MmMMcamT5/OsrKyGGOMzZ07l7377ruMMcbeeecdNnfuXMYYY1lZWWzGjBmMMcby8/PZ\nkCFDWGNjIzt37hwLCwtjra2tdjVmHnzwQfbhhx8yxhhrampitbW1TjUOSktLWWhoKLt27RpjrH0M\nfPLJJ045FmzFoRT+gQMH2J133ql5vXLlSrZy5UoJJeKflJQU9uOPP7LIyEhWXl7OGGt/KERGRjLG\nGEtPT2fr16/XtFe3W79+PUtPT9ccV7crLy9nUVFRmuMd28mBkpISNmHCBLZjxw42efJkplKpmK+v\nL2tpaWGM6f7ud955Jztw4ABjjLGWlhbm6+vLVCqV3lhQt7OXMVNXV8dCQ0OZSqXSOe5M46C0tJQF\nBwez6upq1tLSwiZPnsxycnKcbixwwaFMOmVlZQgJCdG8Dg4ORllZmYQS8UtxcTGOHTuGuLg4VFZW\nIjAwEADQp08fVFZWAjB+DUwdDw4O1jsuJ5555hm88sorcHVtH67V1dXo0aMH3N3bSzJry6z9Pd3d\n3eHj44Pq6mqrr4vcKCoqgp+fH/76179i+PDhePTRR9HQ0OBU4yAoKAjPPfcc+vXrh8DAQPj4+GDk\nyJFONxa44FAK35G5evUq7r33Xrz55pvw9vbWec/FxQUuLi4SSSYs33//Pfz9/Z3eh7q1tRVHjx7F\nE088gWPHjqFr1656NmZHHgcAUFtbi+zsbBQVFaG8vBwNDQ3IycmRWiy7wqEUflBQEEpKSjSvS0tL\nERQUJKFE/NDS0oJ7770XDzzwAKZNmwYACAgIQEVFBQCgoqIC/v7+AIxfA1PHS0tL9Y7Lhf3792PT\npk0IDQ1FWloadu7cifnz56Ourg6tra0AdGXW/p6tra24fPkyfH19rb4uciM4OBjBwcGIi4sDAKSm\npuLo0aNOMw4AIDc3FwMGDICfnx88PDwwbdo07N+/3+nGAiektinxSUtLCxswYAA7d+6cZtPl1KlT\nUovFCZVKxWbPns3mz5+vc/y5557T2axbsGABY4yx77//XmezLjY2ljHWvlkXGhrKampqWE1NDQsN\nDWXV1dWMMf3Nui1btoj4DS1n165dmk3b1NRUnY26d955hzHG2OrVq3U26qZPn84YY+zUqVM6G3UD\nBgxgra2tdjVmbr/9dvbbb78xxhhbsmQJe+6555xqHPz8889s0KBBrKGhgalUKvbggw+yt99+2ynH\ngq04lMJnrN07ISIigoWFhbHly5dLLQ5n9u3bxwCwmJgYNnToUDZ06FC2ZcsWdunSJTZhwgSmUChY\nQkKC5qZVqVTsySefZGFhYWzw4MHsl19+0Zzro48+YuHh4Sw8PJx9/PHHmuO//PILi46OZmFhYWze\nvHl6G4NyQVvhnz17lsXGxrLw8HCWmprKGhsbGWOMXb9+naWmprLw8HAWGxvLzp49q/n88uXLWVhY\nGIuMjNTxQLGXMXPs2DE2cuRIFhMTw6ZMmcJqamqcbhwsXryYRUVFsejoaDZr1izW2NjolGPBVii1\nAkEQhJPgUDZ8giAIwjik8AmCIJwEUvgEQRBOAil8giAIJ4EUPkEQhJNACp8gCMJJIIVPEAThJPw/\nMiYqU1IkEk4AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CQxJpaJrB0ii",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}