ibm_audio_transcriber.py

import subprocess
import argparse
import base64
import json
import time
import httplib2

## Going with IBM
from watson_developer_cloud import SpeechToTextV1


"""
You need the following
pip install --upgrade watson-developer-cloud
Credentials from IBM - Username and pass
"""

##IBM CREDS
speech_to_text = SpeechToTextV1(
    username='07802403-20f7-477d-bd84-9c849f3658a4',
    password='Q44hNwIZNEDW',
    x_watson_learning_opt_out=False
)

def extract_audio(video_file, speech_file):
    command = "ffmpeg -loglevel panic -i "+ str(video_file) + " -ab 160k -ac 1 -ar 16000 -vn " + str(speech_file)
    subprocess.call(command, shell=True)


def main(video_file, speech_file):
    """Transcribe the given audio file extracted from the video file.

    Args:
        video_file: the name of the video stream to extract audio from
        speech_file: the name of the audio file.
    """
    extract_audio(video_file, speech_file)

    with open(speech_file, 'rb') as speech:
        start = time.time()
        response = json.dumps(speech_to_text.recognize(
            speech, content_type='audio/wav', timestamps=True,
            word_confidence=True),
            indent=2)
        print ("####################")
        print ("####################")
        response_latency =  str(time.time() - start)

    print (response)
    print response_latency
    print ("####################")
    print ("####################")



if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('video_file', help='Full path of video file to get audio from')
    parser.add_argument('speech_file', help='Full path of audio file to be recognized')
    args = parser.parse_args()
    main(args.video_file, args.speech_file)