Simple, fast, reasonably accurate speech-to-text processing for audio recordings of speech.

Lightweight Speech to Text with PocketSphinx.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Lightweight Speech-to-Text (STT) with PocketSphinx\n",
    "\n",
    "More info:\n",
    "https://github.com/cmusphinx/pocketsphinx\n",
    "\n",
    "Adapted from:\n",
    "https://github.com/Uberi/speech_recognition/blob/master/examples/audio_transcribe.py\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "## Dependencies:\n",
    "\n",
    "## pip install pocketsphinx\n",
    "## pip install speechrecognition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import speech_recognition as sr\n",
    "\n",
    "# Downloading WAV file to home directory\n",
    "\n",
    "os.chdir(os.path.expanduser('~/Desktop/'))\n",
    "\n",
    "#!wget http://www.stephenmclaughlin.net/hipstas/wgbh_temp/wav/cpb-aacip-15-p26pz51t8d__CBSNS730426X_.h264.wav\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "stores over the floodwaters was over guns constitutes a leopold because it is really a coleman the same way that go goals famous work good souls was titled by google in russian flamboyantly on the title page that's also a pall of russia in that sense i may regard myself as a poet i do not regard myself the poet is a bullet in the sense that shall we regarded himself was a poet when he wrote the stove and home to the skylark i do not regard myself as being close to john keats as opposed i don't even regardless of who's been close that alexander popov support than either famously this miserable throb revolt and robert frost was my idea of a kind of poetry or was interested him however i did regard sock produces decided what that i'm a goner is as a possibility that socrates and the tinge line on most of the evidence would can gianni was a surprise to provence bought this organism that confronted the rug on\n"
     ]
    }
   ],
   "source": [
    "# use the audio file above as STT audio source\n",
    "\n",
    "wav_pathname=\"PUA_Kaldi_tests_161106/Antin-David_Studio111-Q_A_UPenn_3-16-04_1min.wav\"\n",
    "\n",
    "r = sr.Recognizer()\n",
    "with sr.AudioFile(wav_pathname) as source:\n",
    "    audio = r.record(source)    # read the entire audio file\n",
    "\n",
    "# recognize speech using pocketsphinx\n",
    "\n",
    "try:\n",
    "    print(r.recognize_sphinx(audio))\n",
    "except sr.UnknownValueError:\n",
    "    print(\"Sphinx could not understand audio\")\n",
    "except sr.RequestError as e:\n",
    "    print(\"Sphinx error; {0}\".format(e))\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}