analitics

Pages

Sunday, June 4, 2017

The SpeechRecognition python module - part 001.

First, you need to install the SpeechRecognition python module for Windows 10:
C:\Python27>cd Scripts
C:\Python27\Scripts>pip install --upgrade  --trusted-host  pypi.python.org  SpeechRecognition
Collecting SpeechRecognition
  Downloading SpeechRecognition-3.6.5-py2.py3-none-any.whl (31.8MB)
    100% |################################| 31.8MB 4.9MB/s
Installing collected packages: SpeechRecognition
  Found existing installation: SpeechRecognition 3.5.0
    Uninstalling SpeechRecognition-3.5.0:
      Successfully uninstalled SpeechRecognition-3.5.0
Successfully installed SpeechRecognition-3.6.5
The next step is the PyAudio python module:
C:\Python27\Scripts>pip install --upgrade  --trusted-host  pypi.python.org  PyAudio
Collecting PyAudio
  Downloading PyAudio-0.2.11-cp27-cp27m-win32.whl (49kB)
    100% |################################| 51kB 258kB/s
Installing collected packages: PyAudio
  Found existing installation: PyAudio 0.2.9
    Uninstalling PyAudio-0.2.9:
      Successfully uninstalled PyAudio-0.2.9
Successfully installed PyAudio-0.2.11
Also, this python module can be installed under python version 3.4.1:
C:\Python34\Scripts>pip install SpeechRecognition
Downloading/unpacking SpeechRecognition
Installing collected packages: SpeechRecognition
Successfully installed SpeechRecognition
Cleaning up...
The problem with Python 3.4.x version is PyAudio python module installation.
Anyway, I used the python 2.7.13 version to test this module with a simple python script:
import speech_recognition as sr
import os
print ("HELP: Set your microphone hardware on and try this script")
def active_listen():
    r = sr.Recognizer()
    with sr.Microphone() as src:
        audio = r.listen(src)
    msg = ''
    try:
        msg = r.recognize_google(audio)
        print (msg.lower())
    except sr.UnknownValueError:
        print("Google Speech Recognition could not understand audio")
    except sr.RequestError as e:
        print("Could not request results from Google STT; {0}".format(e))
    except:
        print("Unknown exception occurred!")
    finally:
        return msg.lower()
active_listen()
Just start your microphone hardware on and run the script.
Working well for me this test.