{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Audio Decoder in DALI\n",
    "\n",
    "This tutorial presents, how to set up a simple pipeline, that loads and decodes audio data using DALI. We will use a simple example from Speech Commands Data Set. While this dataset consists of samples in .wav format, the following procedure can be used for most of the well-known digital audio coding formats as well.\n",
    "\n",
    "## Step-by-step guide\n",
    "1. Let's start by importing DALI and a handful of utils."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from nvidia.dali.pipeline import Pipeline\n",
    "import nvidia.dali.ops as ops            \n",
    "import nvidia.dali.types as types\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "batch_size = 1\n",
    "audio_files = \"../data/audio\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "used `batch_size` is `1`, to keep things simple.\n",
    "\n",
    "2. Next, let's implement the pipeline. Firstly, we need to load data from disk (or any other source). FileReader is able to load data, as well as it's labels. For more information, refer to FileReader docs. Furthermore, similarly to image data, you can use Reader operators that are specific for a given dataset or a dataset format (see [CaffeReader](https://docs.nvidia.com/deeplearning/sdk/dali-developer-guide/docs/supported_ops.html#nvidia.dali.ops.CaffeReader)). After loading the input data, the pipeline decodes the audio data. As stated above, the AudioDecoder operator is able to decode most of the well-known audio formats.\n",
    "   \n",
    "   Note: Please remember that you shall pass proper data type (argument `dtype`) to the operator. Supported data types can be found in the documentation. If you have 24-bit audio data and you set `dtype=INT16`, it will result in loosing some information from the samples. The default `dtype` for this operator is `INT16`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "class AudioDecoderExample(Pipeline):                   \n",
    "    def __init__(self, batch_size, num_threads, device_id):\n",
    "        super(AudioDecoderExample, self).__init__(batch_size, num_threads, device_id)\n",
    "        self.input = ops.FileReader(device=\"cpu\", file_root=audio_files)\n",
    "        self.decode = ops.AudioDecoder(device=\"cpu\", dtype=types.INT16)\n",
    "\n",
    "    def define_graph(self):                                                                \n",
    "        read, _ = self.input()\n",
    "        audio, rate = self.decode(read)\n",
    "        return audio, rate"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "3. Now let's just build and run the pipeline."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipecpu = AudioDecoderExample(batch_size=batch_size, num_threads=1, device_id=0)\n",
    "pipecpu.build()          \n",
    "cpu_output = pipecpu.run()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Outputs from `AudioDecoder` consist of a tensor with the decoded data, as well as some metadata (e.g. sampling rate). To access them just check another output. On top of that, AudioDecoder returns data in interleaved format, so we need to reshape the output tensor, to properly display it. Here's how to do that:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sampling rate: 16000 [Hz]\n",
      "Audio data: [[  -5]\n",
      " [ -95]\n",
      " [-156]\n",
      " ...\n",
      " [ 116]\n",
      " [ 102]\n",
      " [  82]]\n",
      "Audio data flattened: [  -5  -95 -156 ...  116  102   82]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "audio_data = cpu_output[0].at(0)\n",
    "sampling_rate = cpu_output[1].at(0)[0]\n",
    "print(\"Sampling rate:\", sampling_rate, \"[Hz]\")\n",
    "print(\"Audio data:\", audio_data)\n",
    "audio_data = audio_data.flatten()\n",
    "print(\"Audio data flattened:\", audio_data)\n",
    "plt.plot(audio_data)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Verification\n",
    "\n",
    "Let's verify, that the AudioDecoder actually works. The presented method can also come in handy for debugging DALI pipeline, in case something doesn't go as planned. \n",
    "\n",
    "We will use external tool to decode used data and compare the results against data decoded by DALI.\n",
    "\n",
    "### Important!\n",
    "\n",
    "Following snippet installs the external dependency (`simpleaudio`). In case you already have it, or don't want to install it, you might want to stop here and not run this one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "!{sys.executable} -m pip install simpleaudio"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Below is the side-by-side comparision of decoded data. If you have the `simpleaudio` module installed, you can run the snippet and see it for yourself."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "src: simpleaudio\n",
      "shape:  (16000,)\n",
      "data:  [  -5  -95 -156 ...  116  102   82]\n",
      "\n",
      "\n",
      "src: DALI\n",
      "shape:  (16000,)\n",
      "data:  [  -5  -95 -156 ...  116  102   82]\n",
      "\n",
      "Are the arrays equal? YES\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import simpleaudio as sa\n",
    "\n",
    "wav = sa.WaveObject.from_wave_file(\"../data/audio/wav/three.wav\")\n",
    "three_audio = np.frombuffer(wav.audio_data, dtype=np.int16)\n",
    "\n",
    "print(\"src: simpleaudio\")\n",
    "print(\"shape: \", three_audio.shape)\n",
    "print(\"data: \", three_audio)\n",
    "print(\"\\n\")\n",
    "print(\"src: DALI\")\n",
    "print(\"shape: \", audio_data.shape)\n",
    "print(\"data: \", audio_data)\n",
    "print(\"\\nAre the arrays equal?\", \"YES\" if np.all(audio_data == three_audio) else \"NO\")\n",
    "\n",
    "fig, ax = plt.subplots(1,2)\n",
    "ax[0].plot(three_audio)\n",
    "ax[1].plot(audio_data)\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}